concept: gan-vocoder last_updated: '2026-07-30' paper_count: 60 papers: - id: '1609.03499' published_date: "2016-09-12" entry_date: '2026-07-30' year: 2016 venue: arXiv task: - TTS architecture: - autoregressive-LM relevance: medium evidence_role: - architecture_variant - acceleration_evidence - control_evidence current_role: foundational method_family: - autoregressive_waveform_and_token_decoders claims: - claim_id: direct_generation_of_raw_audio_waveforms_without_intermediate_vocoder_parameters role: supports claim: Direct generation of raw audio waveforms, without intermediate vocoder parameters, produces substantially higher naturalness than parametric or concatenative synthesis pipelines as judged by human listeners. source: §3.2, Table 1 evidence: Direct generation of raw audio waveforms, without intermediate vocoder parameters, produces substantially higher naturalness than parametric or concatenative synthesis pipelines as judged by human listeners. confidence: high relevance: high - claim_id: dilated_causal_convolutions_enable_autoregressive_audio_models_to_achieve_receptive role: supports claim: Dilated causal convolutions enable autoregressive audio models to achieve receptive fields orders of magnitude larger than standard causal convolutions with comparable computational cost. source: §2.1, Figure 3 evidence: Dilated causal convolutions enable autoregressive audio models to achieve receptive fields orders of magnitude larger than standard causal convolutions with comparable computational cost. confidence: high relevance: medium - claim_id: a_single_autoregressive_model_conditioned_on_speaker_identity_can_represent role: supports claim: A single autoregressive model conditioned on speaker identity can represent many voices with shared internal structure, and multi-speaker training improves per-speaker quality relative to single-speaker training. source: §3.1 evidence: A single autoregressive model conditioned on speaker identity can represent many voices with shared internal structure, and multi-speaker training improves per-speaker quality relative to single-speaker training. confidence: high relevance: medium - claim_id: receptive_field_size_is_a_binding_constraint_for_prosodic_naturalness role: supports claim: 'Receptive field size is a binding constraint for prosodic naturalness: when the receptive field is insufficient to cover phrase-level F0 contours, prosody degrades even when segmental quality remains high.' source: §3.2 evidence: 'Receptive field size is a binding constraint for prosodic naturalness: when the receptive field is insufficient to cover phrase-level F0 contours, prosody degrades even when segmental quality remains high.' confidence: high relevance: medium - claim_id: autoregressive_raw_waveform_generation_achieves_high_naturalness_at_the_cost role: complicates claim: Autoregressive raw-waveform generation achieves high naturalness at the cost of sequential sample-level inference, creating a fundamental speed-quality trade-off that constrains deployment in real-time applications. source: §4, §3.2 evidence: Autoregressive raw-waveform generation achieves high naturalness at the cost of sequential sample-level inference, creating a fundamental speed-quality trade-off that constrains deployment in real-time applications. confidence: high relevance: medium limitations: - Inference is strictly sequential at the sample level, requiring approximately one computation step per generated sample. At the reported generation rates (roughly 1.5× real-time compute), WaveNet is not suitable for real-time TTS deployment without hardware-specific optimisation or a parallel decoding approximation. - Evaluation is conducted on proprietary Google TTS databases, making direct replication by external researchers impossible. The MOS comparison is fair internally (same data, same test sentences for all systems) but cannot be directly compared to numbers from other published evaluations. - The receptive field of 240 milliseconds covers 3-4 phonemes. Long-range prosodic structure above the syllable and phrase level requires either an external F0 model or a substantially larger receptive field than dilated convolutions alone provide efficiently. - WaveNet as presented requires high-quality linguistic features derived from a text analysis front-end. It is not end-to-end trainable from text to waveform, deferring the alignment and duration prediction problems to external modules. caveats: [] - id: '1712.05884' published_date: "2017-12-16" entry_date: '2026-07-30' year: 2017 venue: arXiv task: - TTS architecture: - transformer-enc-dec relevance: medium evidence_role: - architecture_variant current_role: foundational method_family: - transformer_conditioned_vocoder_systems claims: - claim_id: a_compact_low_level_acoustic_intermediate_representation_can_bridge_text role: supports claim: A compact low-level acoustic intermediate representation can bridge text encoding and neural vocoder synthesis without requiring hand-crafted linguistic features, enabling fully end-to-end neural TTS at human-quality levels. source: §2.1, Table 1 evidence: A compact low-level acoustic intermediate representation can bridge text encoding and neural vocoder synthesis without requiring hand-crafted linguistic features, enabling fully end-to-end neural TTS at human-quality levels. confidence: high relevance: high - claim_id: location_sensitive_attention_which_incorporates_cumulative_attention_weights_as_a role: supports claim: Location-sensitive attention, which incorporates cumulative attention weights as a conditioning signal, reduces failure modes such as repetition and omission compared to standard additive attention in autoregressive TTS decoders. source: §2.2 evidence: Location-sensitive attention, which incorporates cumulative attention weights as a conditioning signal, reduces failure modes such as repetition and omission compared to standard additive attention in autoregressive TTS decoders. confidence: high relevance: low - claim_id: a_pre_net_information_bottleneck_in_the_autoregressive_decoder_is role: supports claim: A pre-net information bottleneck in the autoregressive decoder is essential for stable attention alignment during training. source: §2.2 evidence: A pre-net information bottleneck in the autoregressive decoder is essential for stable attention alignment during training. confidence: high relevance: medium - claim_id: training_the_vocoder_on_predicted_rather_than_ground_truth_intermediate role: supports claim: Training the vocoder on predicted rather than ground-truth intermediate features is necessary because predicted features are systematically oversmoothed; vocoders trained on clean features degrade when exposed to predicted inputs. source: §3.3.1, Table 2 evidence: Training the vocoder on predicted rather than ground-truth intermediate features is necessary because predicted features are systematically oversmoothed; vocoders trained on clean features degrade when exposed to predicted inputs. confidence: high relevance: high - claim_id: neural_vocoders_conditioned_on_compact_mel_spectrograms_can_operate_with role: supports claim: Neural vocoders conditioned on compact mel spectrograms can operate with substantially smaller receptive fields than those conditioned on linguistic features, enabling significant architecture simplification without quality loss. source: §3.3.4, Table 4 evidence: Neural vocoders conditioned on compact mel spectrograms can operate with substantially smaller receptive fields than those conditioned on linguistic features, enabling significant architecture simplification without quality loss. confidence: high relevance: medium limitations: - All experiments use a single proprietary speaker and an internal dataset. The paper provides no evidence of generalisation to multiple speakers, different languages, or recording conditions, and no training or evaluation data is publicly released, limiting independent reproducibility. - 'Prosody remains imperfect: 23 of 100 test sentences are rated as containing unnatural prosody (wrong emphasis, unnatural pitch), and occasional mispronunciations are identified as the main gap between the system and human speech. The paper''s authors explicitly identify prosody modeling as the primary remaining challenge.' - 'The two-stage training setup, while producing high quality, introduces a train/inference mismatch that the ablations expose: the vocoder trained on predicted spectrograms is specifically tuned to the oversmoothed distribution of the encoder-decoder output. Any change to the first stage (retraining, fine-tuning) requires retraining the vocoder independently. Joint end-to-end training is left as future work.' - Real-time synthesis is not addressed; WaveNet inference was known to be slow at the time of publication. caveats: [] - id: '2006.04558' published_date: "2020-06-08" entry_date: '2026-07-30' year: 2020 venue: arXiv task: - TTS architecture: - transformer-enc-dec relevance: medium evidence_role: - architecture_variant current_role: influential method_family: - transformer_conditioned_vocoder_systems claims: - claim_id: explicit_variance_conditioning_on_pitch_energy_and_duration_in_non role: supports claim: Explicit variance conditioning on pitch, energy, and duration in non-autoregressive TTS reduces the information gap between text and speech, enabling quality matching or surpassing autoregressive models without autoregressive inference. source: §2.2, §3.2.1, Table 1 evidence: Explicit variance conditioning on pitch, energy, and duration in non-autoregressive TTS reduces the information gap between text and speech, enabling quality matching or surpassing autoregressive models without autoregressive inference. confidence: high relevance: low - claim_id: phoneme_durations_derived_from_forced_alignment_are_substantially_more_accurate role: supports claim: Phoneme durations derived from forced alignment are substantially more accurate than those extracted from autoregressive teacher attention maps, and this accuracy difference directly improves output voice quality. source: §3.2.2, Table 5 evidence: Phoneme durations derived from forced alignment are substantially more accurate than those extracted from autoregressive teacher attention maps, and this accuracy difference directly improves output voice quality. confidence: high relevance: medium - claim_id: modeling_pitch_in_the_frequency_domain_via_continuous_wavelet_transform role: supports claim: Modeling pitch in the frequency domain via continuous wavelet transform produces more natural pitch distributions in synthesized speech than direct time-domain F0 regression. source: §2.3, §3.2.3, Table 3 evidence: Modeling pitch in the frequency domain via continuous wavelet transform produces more natural pitch distributions in synthesized speech than direct time-domain F0 regression. confidence: high relevance: medium - claim_id: removing_teacher_student_distillation_from_non_autoregressive_tts_training_can role: supports claim: Removing teacher-student distillation from non-autoregressive TTS training can simultaneously simplify the pipeline, reduce training time, and improve output quality relative to the distillation approach. source: §1, §3.2.1, Table 2 evidence: Removing teacher-student distillation from non-autoregressive TTS training can simultaneously simplify the pipeline, reduce training time, and improve output quality relative to the distillation approach. confidence: high relevance: low - claim_id: end_to_end_parallel_text_to_waveform_generation_is_technically role: supports claim: End-to-end parallel text-to-waveform generation is technically feasible but more sensitive to variance conditioning than mel-spectrogram-based acoustic models, particularly in the pitch dimension. source: §2.4, §3.2.3, Table 6 evidence: End-to-end parallel text-to-waveform generation is technically feasible but more sensitive to variance conditioning than mel-spectrogram-based acoustic models, particularly in the pitch dimension. confidence: high relevance: medium limitations: - All experiments use a single studio-quality English speaker (LJSpeech). Generalisation to multi-speaker, emotional, noisy, or cross-lingual conditions is untested. The paper does not include speaker similarity metrics or speaker-conditioned variants. - The variance adaptor relies on an external forced alignment tool (MFA) and pitch extraction (PyWorldVocoder) at training time; these are not learned end-to-end, which limits the system from being fully self-contained. The authors acknowledge this and note future work toward eliminating external alignment tools. - FastSpeech 2s requires the mel-spectrogram decoder during training even though it is discarded at inference. This coupling adds training complexity and means the waveform decoder benefits from a pre-trained mel-spectrogram branch, creating a two-stage dependency that partially reintroduces the pipeline complexity the paper otherwise avoids. - The CMOS evaluation protocol is used for ablation comparisons rather than the more interpretable absolute MOS, which makes it harder to assess the magnitude of individual component contributions in absolute terms. caveats: [] - id: '2010.05646' published_date: "2020-10-12" entry_date: '2026-07-30' year: 2020 venue: arXiv task: - TTS architecture: - GAN relevance: high evidence_role: - core_evidence - acceleration_evidence - control_evidence current_role: foundational method_family: - adversarial_waveform_vocoders claims: - claim_id: explicitly_modeling_periodic_structure_in_speech_at_multiple_timescales_is role: supports claim: Explicitly modeling periodic structure in speech at multiple timescales is necessary for GAN-based vocoders to match the perceptual quality of autoregressive models. source: §2.3, Table 2 evidence: Explicitly modeling periodic structure in speech at multiple timescales is necessary for GAN-based vocoders to match the perceptual quality of autoregressive models. confidence: high relevance: high - claim_id: a_compact_gan_vocoder_can_achieve_cpu_real_time_synthesis role: supports claim: A compact GAN vocoder can achieve CPU real-time synthesis with quality comparable to autoregressive models when discriminator design, rather than generator depth, is the primary quality bottleneck. source: §4.1, Table 1 evidence: A compact GAN vocoder can achieve CPU real-time synthesis with quality comparable to autoregressive models when discriminator design, rather than generator depth, is the primary quality bottleneck. confidence: high relevance: high - claim_id: vocoders_trained_on_single_speaker_data_generalise_to_unseen_speakers role: supports claim: Vocoders trained on single-speaker data generalise to unseen speakers when the generator is conditioned only on mel-spectrograms, with quality exceeding flow-based and autoregressive alternatives. source: §4.3, Table 3 evidence: Vocoders trained on single-speaker data generalise to unseen speakers when the generator is conditioned only on mel-spectrograms, with quality exceeding flow-based and autoregressive alternatives. confidence: high relevance: high - claim_id: fine_tuning_a_mel_spectrogram_vocoder_on_predicted_rather_than role: supports claim: Fine-tuning a mel-spectrogram vocoder on predicted (rather than ground-truth) spectrograms substantially improves end-to-end TTS quality, while flow-based vocoders do not benefit from the same adaptation. source: §4.4, Table 4 evidence: Fine-tuning a mel-spectrogram vocoder on predicted (rather than ground-truth) spectrograms substantially improves end-to-end TTS quality, while flow-based vocoders do not benefit from the same adaptation. confidence: high relevance: high - claim_id: discriminator_architecture_choices_have_a_larger_impact_on_gan_vocoder role: supports claim: Discriminator architecture choices have a larger impact on GAN vocoder quality than generator architecture choices. source: §4.2, Table 2 evidence: Discriminator architecture choices have a larger impact on GAN vocoder quality than generator architecture choices. confidence: high relevance: high limitations: - Evaluation is limited to LJSpeech (single speaker, high-quality studio recordings) and VCTK (multi-speaker English). Performance on noisy, spontaneous, or cross-lingual speech is not assessed. The end-to-end experiment is conducted with a single acoustic model (Tacotron2), so generalisation of the fine-tuning recipe to other front-ends is not established. The paper does not report streaming or chunk-wise inference latency, which matters for real-time interactive applications. - The ablation study uses only V3 (smallest variant) trained to 500k steps, rather than the full V1 model trained to convergence; it is possible that at higher capacity the relative importance of MPD versus MRF would differ. caveats: [] - id: '2104.00355' published_date: "2021-04-01" entry_date: '2026-07-30' year: 2021 venue: arXiv task: - TTS - VC architecture: - GAN - VAE relevance: medium evidence_role: - architecture_variant - control_evidence - evaluation_caution current_role: influential method_family: - adversarial_waveform_vocoders - vae_codec_vocoder_decoders claims: - claim_id: ssl_content_representations_that_are_well_disentangled_from_speaker_identity role: supports claim: SSL content representations that are well-disentangled from speaker identity also exhibit stronger voice conversion performance, while representations that entangle speaker information perform worse at conversion but better at pitch reconstruction. source: §4, Table 1, Table 2 evidence: SSL content representations that are well-disentangled from speaker identity also exhibit stronger voice conversion performance, while representations that entangle speaker information perform worse at conversion but better at pitch reconstruction. confidence: high relevance: low - claim_id: discrete_speech_units_learned_by_ssl_models_can_form_the role: supports claim: Discrete speech units learned by SSL models can form the basis of an ultra-low-bitrate speech codec that outperforms classical parametric codecs in subjective quality. source: §4, Figure 2 evidence: Discrete speech units learned by SSL models can form the basis of an ultra-low-bitrate speech codec that outperforms classical parametric codecs in subjective quality. confidence: high relevance: low - claim_id: among_self_supervised_content_encoders_hubert_units_carry_less_speaker role: supports claim: Among self-supervised content encoders, HuBERT units carry less speaker and pitch information than VQ-VAE units, making them better suited for downstream controllable synthesis. source: §4, Table 2 evidence: Among self-supervised content encoders, HuBERT units carry less speaker and pitch information than VQ-VAE units, making them better suited for downstream controllable synthesis. confidence: high relevance: medium - claim_id: pitch_and_speaker_identity_can_be_independently_conditioned_in_a role: supports claim: Pitch and speaker identity can be independently conditioned in a GAN vocoder through separate discrete token streams, enabling controllable F0 manipulation without retraining. source: §3, §4 evidence: Pitch and speaker identity can be independently conditioned in a GAN vocoder through separate discrete token streams, enabling controllable F0 manipulation without retraining. confidence: high relevance: high limitations: - The codec evaluation uses only 20 utterances from 5 VCTK speakers, all unseen during training but from the same corpus. Generalization to out-of-domain speech (conversational, noisy, or non-English) is untested. - The resynthesis MOS scores remain well below ground truth on both LJSpeech (3.66 vs. 4.33) and VCTK (3.41 vs. 4.08), indicating a quality gap the system does not close. Disentanglement is evaluated indirectly through proxy metrics (EER, VDE, FFE) rather than a direct information-theoretic measure. The speaker encoder requires speaker embeddings from training-set speakers for the lookup-table variant; the d-vector approach generalizes but relies on a separately trained verification model. No ablation isolates the contribution of the F0 conditioning stream to final MOS. The MUSHRA scores in Figure 2 are visual only, making exact numerical comparison to baselines difficult to reproduce from the paper text alone. caveats: [] - id: '2106.15561' published_date: "2021-06-29" entry_date: '2026-07-30' year: 2021 venue: arXiv task: - TTS architecture: - autoregressive-LM - flow-matching - diffusion - GAN - VAE - transformer-enc-dec relevance: medium evidence_role: - architecture_variant - control_evidence - infrastructure current_role: influential method_family: - autoregressive_waveform_and_token_decoders - flow_matching_waveform_generation - diffusion_vocoder_and_restoration_systems - adversarial_waveform_vocoders - vae_codec_vocoder_decoders - transformer_conditioned_vocoder_systems claims: - claim_id: the_transition_from_attention_based_alignment_to_explicit_duration_prediction role: supports claim: The transition from attention-based alignment to explicit duration prediction in non-autoregressive acoustic models eliminates word skipping and repeating artefacts while enabling parallel mel-spectrogram generation. source: §3.4, Table 11 evidence: The transition from attention-based alignment to explicit duration prediction in non-autoregressive acoustic models eliminates word skipping and repeating artefacts while enabling parallel mel-spectrogram generation. confidence: high relevance: medium - claim_id: modelling_variation_information_prosody_style_speaker_at_multiple_granularities_from role: supports claim: Modelling variation information (prosody, style, speaker) at multiple granularities from frame to utterance level improves expressiveness while fine-grained control remains difficult when relying on implicit latent representations. source: §3.5.1, §3.5.3 evidence: Modelling variation information (prosody, style, speaker) at multiple granularities from frame to utterance level improves expressiveness while fine-grained control remains difficult when relying on implicit latent representations. confidence: high relevance: medium - claim_id: diffusion_based_vocoders_achieve_high_perceptual_quality_but_suffer_from role: supports claim: Diffusion-based vocoders achieve high perceptual quality but suffer from slow inference due to iterative reverse diffusion, motivating step-reduction methods. source: §2.4, Table 6 evidence: Diffusion-based vocoders achieve high perceptual quality but suffer from slow inference due to iterative reverse diffusion, motivating step-reduction methods. confidence: high relevance: low - claim_id: zero_shot_speaker_adaptation_that_relies_solely_on_a_speaker role: supports claim: Zero-shot speaker adaptation that relies solely on a speaker encoder without any fine-tuning produces noticeably lower voice quality than few-shot adaptation, particularly for speakers distant from the training distribution. source: §3.6.2 evidence: Zero-shot speaker adaptation that relies solely on a speaker encoder without any fine-tuning produces noticeably lower voice quality than few-shot adaptation, particularly for speakers distant from the training distribution. confidence: high relevance: medium - claim_id: low_resource_tts_benefits_from_cross_lingual_transfer_because_human role: supports claim: Low-resource TTS benefits from cross-lingual transfer because human languages share vocal organs and phonetic structures, though phoneme set mismatches require explicit bridging. source: §3.3 evidence: Low-resource TTS benefits from cross-lingual transfer because human languages share vocal organs and phonetic structures, though phoneme set mismatches require explicit bridging. confidence: high relevance: low limitations: - 'The coverage horizon is mid-2021, which means the survey predates several paradigm shifts: neural codec-based discrete speech tokenisation, large language model-based TTS (VALL-E, NaturalSpeech 2), flow-matching acoustic models, and the instruction-conditioned and zero-shot capabilities that came with LLM-scale training. The survey''s taxonomy (acoustic model + vocoder pipeline) is also less applicable to fully codec-native systems where this separation dissolves.' - The survey focuses on English and Mandarin and, while noting multilingual coverage, does not deeply address the orthographic diversity or tonal complexity of the majority of the world's languages. Evaluation methodology is discussed mainly by cataloguing metrics rather than analysing their reliability or mutual disagreement, which limits the survey's utility for assessing evaluation practices. - 'Open questions identified include: better variation information modeling that reconciles controllability and expressiveness, casual and spontaneous speech synthesis, parameter-efficient adaptation at scale, and energy-efficient training and inference.' caveats: [] - id: '2206.04658' published_date: "2022-06-09" entry_date: '2026-07-30' year: 2022 venue: arXiv task: - TTS architecture: - GAN relevance: high evidence_role: - core_evidence - control_evidence - evaluation_caution current_role: influential method_family: - adversarial_waveform_vocoders claims: - claim_id: periodic_activation_functions_with_learnable_frequency_parameters_improve_gan_vocoder role: supports claim: Periodic activation functions with learnable frequency parameters improve GAN vocoder robustness to out-of-distribution audio, including unseen recording environments, languages, and non-speech signals. source: §3.2, §4.5, §4.6 evidence: Periodic activation functions with learnable frequency parameters improve GAN vocoder robustness to out-of-distribution audio, including unseen recording environments, languages, and non-speech signals. confidence: high relevance: high - claim_id: anti_aliasing_via_low_pass_filtering_around_nonlinear_activations_reduces role: supports claim: Anti-aliasing via low-pass filtering around nonlinear activations reduces high-frequency artifacts in neural vocoders and is complementary to periodic inductive bias. source: §3.3, §4.7, Table 4 evidence: Anti-aliasing via low-pass filtering around nonlinear activations reduces high-frequency artifacts in neural vocoders and is complementary to periodic inductive bias. confidence: high relevance: medium - claim_id: gan_vocoders_benefit_significantly_from_scaling_model_capacity_when_paired role: supports claim: GAN vocoders benefit significantly from scaling model capacity when paired with appropriate architectural improvements and training stabilisation techniques, whereas naive scaling without those changes leads to early collapse. source: §3.4, §4.7 evidence: GAN vocoders benefit significantly from scaling model capacity when paired with appropriate architectural improvements and training stabilisation techniques, whereas naive scaling without those changes leads to early collapse. confidence: high relevance: high - claim_id: similarity_mos_smos_which_requires_direct_comparison_with_ground_truth role: supports claim: Similarity MOS (SMOS), which requires direct comparison with ground truth, is a more sensitive evaluation metric than standard MOS for distinguishing speaker fidelity under diverse and out-of-distribution conditions. source: §4.3 evidence: Similarity MOS (SMOS), which requires direct comparison with ground truth, is a more sensitive evaluation metric than standard MOS for distinguishing speaker fidelity under diverse and out-of-distribution conditions. confidence: high relevance: low - claim_id: training_on_diverse_multi_condition_speech_data_including_noisy_recordings role: supports claim: Training on diverse, multi-condition speech data (including noisy recordings) is a necessary complement to architectural improvements for achieving universal vocoding performance. source: §4.7, Table 5 evidence: Training on diverse, multi-condition speech data (including noisy recordings) is a necessary complement to architectural improvements for achieving universal vocoding performance. confidence: high relevance: medium limitations: - BigVGAN is trained and evaluated on 24 kHz audio only. Generalisation to 44.1 kHz or 48 kHz audio (common in music and broadcast) is not tested and the model architecture is not designed for it. - The 112M model synthesises at 44.72x real-time, which is slower than HiFi-GAN (93.75x) and BigVGAN-base (70.18x). For latency-sensitive applications, the architectural overhead of filtered Snake activations is a concrete trade-off. The paper acknowledges that spectral normalisation, anti-aliased upsampling layers, and larger discriminators all degraded quality; the practical lessons are reported informally but without a principled explanation of why regularisation harms audio GANs when it helps image GANs. MOS evaluations are conducted with native English-speaking Mechanical Turk workers even for non-English and OOD conditions, which may introduce evaluator bias. The model is trained exclusively on speech despite being evaluated on music and instruments, and the extrapolation limits of this approach for broader audio generation are not characterised. caveats: [] - id: '2210.13438' published_date: "2022-10-24" entry_date: '2026-07-30' year: 2022 venue: arXiv task: - codec architecture: - GAN relevance: medium evidence_role: - architecture_variant - acceleration_evidence - control_evidence current_role: foundational method_family: - adversarial_waveform_vocoders claims: - claim_id: a_multi_scale_stft_discriminator_alone_is_sufficient_for_perceptual role: supports claim: A multi-scale STFT discriminator alone is sufficient for perceptual quality in neural audio codecs, removing the need for waveform-domain discriminators. source: §4.5.1, Table 2 evidence: A multi-scale STFT discriminator alone is sufficient for perceptual quality in neural audio codecs, removing the need for waveform-domain discriminators. confidence: high relevance: high - claim_id: gradient_balancers_that_normalise_loss_contributions_by_expected_gradient_magnitude role: supports claim: Gradient balancers that normalise loss contributions by expected gradient magnitude substantially stabilise training when combining reconstruction, adversarial, and commitment losses with widely varying natural scales. source: §3.4, Table A.4 evidence: Gradient balancers that normalise loss contributions by expected gradient magnitude substantially stabilise training when combining reconstruction, adversarial, and commitment losses with widely varying natural scales. confidence: high relevance: medium - claim_id: residual_vector_quantization_supports_variable_bitrate_operation_from_a_single role: supports claim: Residual vector quantization supports variable-bitrate operation from a single model by varying the number of active codebooks at inference, with each additional codebook yielding diminishing quality returns. source: §3.2, Table 1 evidence: Residual vector quantization supports variable-bitrate operation from a single model by varying the number of active codebooks at inference, with each additional codebook yielding diminishing quality returns. confidence: high relevance: medium - claim_id: auxiliary_transformer_language_models_over_rvq_codes_can_reduce_effective role: complicates claim: Auxiliary Transformer language models over RVQ codes can reduce effective bitrate by 25-40% through entropy coding without perceptual quality degradation, at the cost of increased latency. source: §3.3, §4.5 evidence: Auxiliary Transformer language models over RVQ codes can reduce effective bitrate by 25-40% through entropy coding without perceptual quality degradation, at the cost of increased latency. confidence: high relevance: low - claim_id: neural_audio_codecs_outperform_traditional_dsp_codecs_at_low_bitrates role: supports claim: Neural audio codecs outperform traditional DSP codecs at low bitrates across both speech and music domains, with the quality gap widening as bitrate decreases. source: §4.5, Table 1, Figure 3 evidence: Neural audio codecs outperform traditional DSP codecs at low bitrates across both speech and music domains, with the quality gap widening as bitrate decreases. confidence: high relevance: medium limitations: - The 48 kHz model in non-streamable configuration operates slower than real time on CPU, limiting deployment without GPU acceleration or hardware-specific optimisation. Arithmetic coding precision issues (floating-point non-determinism across architectures) required a probability rounding workaround that the authors note may be insufficient for practical deployment. - The Transformer language model is small (5 layers, 200 channels) and neglects mutual information between codebooks at the same time step, leaving compression gains on the table. Music compression at 1.5 kbps still shows significant perceptual degradation. Training datasets mix multiple licenses, which may complicate commercial use. No evaluation of robustness to codec chaining (encode-decode-re-encode) or of downstream task performance degradation from quantization artefacts is provided. caveats: [] - id: '2305.02765' published_date: "2023-05-04" entry_date: '2026-07-30' year: 2023 venue: arXiv task: - codec architecture: - GAN - VAE relevance: medium evidence_role: - architecture_variant - evaluation_caution current_role: influential method_family: - adversarial_waveform_vocoders - vae_codec_vocoder_decoders claims: - claim_id: grouping_residual_vector_quantization_into_parallel_chains_rather_than_a role: supports claim: Grouping residual vector quantization into parallel chains rather than a single sequential chain improves reconstruction quality per codebook, enabling competitive fidelity with fewer total quantizers. source: §3.3, Table 1 evidence: Grouping residual vector quantization into parallel chains rather than a single sequential chain improves reconstruction quality per codebook, enabling competitive fidelity with fewer total quantizers. confidence: high relevance: medium - claim_id: the_burden_that_codec_codebook_count_imposes_on_downstream_generation role: supports claim: The burden that codec codebook count imposes on downstream generation models is a practical constraint that drives codec architecture choices independently of raw reconstruction quality. source: §1 Introduction, §3.3 evidence: The burden that codec codebook count imposes on downstream generation models is a practical constraint that drives codec architecture choices independently of raw reconstruction quality. confidence: high relevance: low - claim_id: objective_speech_quality_metrics_such_as_pesq_and_stoi_are role: supports claim: Objective speech quality metrics such as PESQ and STOI are insufficient alone to characterise codec reconstruction quality, and subjective evaluation is necessary but often omitted in codec research. source: §6 Limitations evidence: Objective speech quality metrics such as PESQ and STOI are insufficient alone to characterise codec reconstruction quality, and subjective evaluation is necessary but often omitted in codec research. confidence: high relevance: low - claim_id: publicly_available_training_code_and_pre_trained_baselines_for_neural role: supports claim: Publicly available training code and pre-trained baselines for neural audio codecs are necessary for reproducible research, as previously these were unavailable for EnCodec and SoundStream. source: §5 Conclusion, §6 Limitations evidence: Publicly available training code and pre-trained baselines for neural audio codecs are necessary for reproducible research, as previously these were unavailable for EnCodec and SoundStream. confidence: high relevance: medium limitations: - 'No subjective evaluation is included. The paper''s own §6 acknowledges this as a limitation: "Subjective evaluation is always the best choice, but this part is missed in this study." All quality comparisons rest solely on PESQ and STOI, which the authors themselves note may not accurately reflect perceptual quality.' - Beyond the missing subjective evaluation, the paper does not validate GRVQ on downstream generation tasks. The claim that 4 codebooks reduce burden on generation models is plausible and consistent with the motivation, but no TTS or audio LM experiments are included. The training data is limited to English and Chinese speech from public TTS corpora; generalisation to music, environmental sound, or noisy/spontaneous speech is not tested. Finally, the paper trains at 16kHz and 24kHz only; higher sample rates (44.1kHz, 48kHz) used in music and high-fidelity audio applications are not addressed. caveats: [] - id: '2306.00814' published_date: "2023-06-01" entry_date: '2026-07-30' year: 2023 venue: arXiv task: - TTS architecture: - GAN relevance: high evidence_role: - core_evidence current_role: influential method_family: - adversarial_waveform_vocoders claims: - claim_id: maintaining_constant_temporal_resolution_throughout_a_gan_vocoder_with_istft role: supports claim: Maintaining constant temporal resolution throughout a GAN vocoder, with ISTFT as the sole upsampling step, eliminates aliasing artefacts and dramatically reduces inference cost without sacrificing perceptual quality. source: §3.1, §4.3, Table 6 evidence: Maintaining constant temporal resolution throughout a GAN vocoder, with ISTFT as the sole upsampling step, eliminates aliasing artefacts and dramatically reduces inference cost without sacrificing perceptual quality. confidence: high relevance: high - claim_id: implicit_phase_wrapping_via_a_unit_circle_activation_is_essential role: supports claim: Implicit phase wrapping via a unit-circle activation is essential for stable GAN training of complex-valued spectrogram generators; alternatives that clamp or clip phase angles substantially degrade output quality. source: §3.2, §4.1.1, Table 1 evidence: Implicit phase wrapping via a unit-circle activation is essential for stable GAN training of complex-valued spectrogram generators; alternatives that clamp or clip phase angles substantially degrade output quality. confidence: high relevance: high - claim_id: fourier_domain_vocoders_reduce_periodicity_errors_more_effectively_than_time role: supports claim: Fourier-domain vocoders reduce periodicity errors more effectively than time-domain GANs, suggesting that modelling harmonics in the frequency domain provides a stronger inductive bias for voiced speech. source: §4.1.1, Table 1 evidence: Fourier-domain vocoders reduce periodicity errors more effectively than time-domain GANs, suggesting that modelling harmonics in the frequency domain provides a stronger inductive bias for voiced speech. confidence: high relevance: medium - claim_id: convnext_blocks_with_isotropic_architecture_outperform_dilated_resblocks_in_the role: supports claim: ConvNeXt blocks with isotropic architecture outperform dilated ResBlocks in the Fourier-domain vocoder setting, even though dilated convolutions were motivated by the need to expand receptive fields in time-domain models. source: §4.1.1, Table 1 evidence: ConvNeXt blocks with isotropic architecture outperform dilated ResBlocks in the Fourier-domain vocoder setting, even though dilated convolutions were motivated by the need to expand receptive fields in time-domain models. confidence: high relevance: high - claim_id: a_fourier_domain_gan_vocoder_trained_as_a_neural_codec role: supports claim: A Fourier-domain GAN vocoder trained as a neural codec decoder can substantially improve perceptual quality over the original codec decoder across all bitrates without architectural changes to the upstream codec. source: §4.2, Table 5 evidence: A Fourier-domain GAN vocoder trained as a neural codec decoder can substantially improve perceptual quality over the original codec decoder across all bitrates without architectural changes to the upstream codec. confidence: high relevance: high limitations: - Vocos's mel-spectrogram MOS scores are reported on LibriTTS using crowd-sourced listeners; the ground-truth MOS (3.81) is noticeably below what might be expected for studio speech, suggesting the evaluation pool or headphone compliance filtering may limit the discriminative power of the subjective test. Statistical equivalence with BigVGAN is shown, but the test may be underpowered for detecting small differences. - The speed benchmarks are conducted without hardware-specific optimisations and use batch size 16, which is unlikely to reflect latency-critical single-sample streaming scenarios. The CPU advantage (169x real-time) is particularly striking but is not validated in a streaming or low-latency deployment setting. - The MDCT variant explored in Appendix A performs worse than the ISTFT variant across all metrics, suggesting the overcomplete STFT representation provides a beneficial inductive bias. Whether MDCT-based generation could become competitive with improved training strategies remains open. - The model is trained exclusively on 24 kHz speech and music; performance at higher sample rates (44.1 kHz or 48 kHz), which matter for high-fidelity TTS applications, is not reported. caveats: [] - id: '2409.05377' published_date: "2024-09-09" entry_date: '2026-07-30' year: 2024 venue: arXiv task: - codec architecture: - GAN - VAE relevance: medium evidence_role: - architecture_variant - control_evidence current_role: influential method_family: - adversarial_waveform_vocoders - vae_codec_vocoder_decoders claims: - claim_id: at_very_low_bitrates_around_1_kbps_model_capacity_is role: supports claim: At very low bitrates (around 1 kbps), model capacity is a decisive factor in reconstruction quality, with larger models substantially outperforming architecturally sophisticated but smaller codecs. source: §III-C, §IV-B, Table I evidence: At very low bitrates (around 1 kbps), model capacity is a decisive factor in reconstruction quality, with larger models substantially outperforming architecturally sophisticated but smaller codecs. confidence: high relevance: medium - claim_id: low_dimensional_vector_quantisation_before_codebook_lookup_substantially_improves_codebook role: supports claim: Low-dimensional vector quantisation before codebook lookup substantially improves codebook utilisation in single-codebook, single-quantisation-step codec designs. source: §III-A, §IV-B evidence: Low-dimensional vector quantisation before codebook lookup substantially improves codebook utilisation in single-codebook, single-quantisation-step codec designs. confidence: high relevance: low - claim_id: adding_sequential_lstm_modelling_to_a_convolutional_codec_encoder_improves role: supports claim: Adding sequential (LSTM) modelling to a convolutional codec encoder improves both perceptual quality and speaker similarity at low bitrates, independently of the parameter count effect. source: §III-A, §IV-D, Table III evidence: Adding sequential (LSTM) modelling to a convolutional codec encoder improves both perceptual quality and speaker similarity at low bitrates, independently of the parameter count effect. confidence: high relevance: low - claim_id: scaling_codec_model_size_beyond_a_saturation_point_approximately_159m role: supports claim: Scaling codec model size beyond a saturation point (approximately 159M parameters in this setting) yields no further reconstruction benefit, analogous to scale-up behaviour observed in neural vocoders. source: §IV-D, Table III evidence: Scaling codec model size beyond a saturation point (approximately 159M parameters in this setting) yields no further reconstruction benefit, analogous to scale-up behaviour observed in neural vocoders. confidence: high relevance: low - claim_id: increasing_training_data_volume_from_960_hours_to_60k_hours role: supports claim: Increasing training data volume from 960 hours to 60k hours does not improve codec reconstruction quality, suggesting that model capacity rather than data quantity is the binding constraint at this bitrate. source: §IV-D, Table III evidence: Increasing training data volume from 960 hours to 60k hours does not improve codec reconstruction quality, suggesting that model capacity rather than data quantity is the binding constraint at this bitrate. confidence: high relevance: low limitations: - BigCodec is trained exclusively on clean English speech (LibriSpeech 960h), while competing codecs such as EnCodec and DAC train on diverse multilingual datasets including music and environmental sounds. The multilingual generalisation result is encouraging, but the clean-speech-only training domain limits applicability to noisy or music-heavy audio without fine-tuning. - The RTF of 1.1x on a high-end desktop CPU means BigCodec barely achieves real-time decoding; edge-device or streaming deployments would require hardware acceleration. The model's 159M parameter footprint is an order of magnitude larger than EnCodec (14M), imposing memory costs for downstream TTS systems that embed a codec. The paper does not report streaming latency or chunked-inference performance, which is relevant for spoken conversational agent use cases. - Whether the subjective superiority over ground truth reflects a genuine perceptual enhancement or a test artefact (e.g., listener anchoring, signal processing smoothing) remains unaddressed and warrants scepticism. caveats: [] - id: '2411.01156' published_date: "2024-11-02" entry_date: '2026-07-30' year: 2024 venue: arXiv task: - TTS architecture: - autoregressive-LM - GAN relevance: medium evidence_role: - architecture_variant - acceleration_evidence current_role: active_evidence method_family: - autoregressive_waveform_and_token_decoders - adversarial_waveform_vocoders claims: - claim_id: eliminating_grapheme_to_phoneme_conversion_by_directly_feeding_raw_text role: supports claim: Eliminating grapheme-to-phoneme conversion by directly feeding raw text to an LLM backbone is viable for multilingual TTS and can improve handling of context-dependent polyphonic words. source: §1, §3 evidence: Eliminating grapheme-to-phoneme conversion by directly feeding raw text to an LLM backbone is viable for multilingual TTS and can improve handling of context-dependent polyphonic words. confidence: high relevance: low - claim_id: hierarchical_decomposition_of_autoregressive_token_generation_into_semantic_level_and role: supports claim: Hierarchical decomposition of autoregressive token generation into semantic-level and acoustic-level stages improves codebook stability in grouped scalar quantization. source: §3.1, §3.1.1 evidence: Hierarchical decomposition of autoregressive token generation into semantic-level and acoustic-level stages improves codebook stability in grouped scalar quantization. confidence: high relevance: medium - claim_id: grouped_finite_scalar_vector_quantization_achieves_higher_codebook_utilisation_than role: supports claim: Grouped Finite Scalar Vector Quantization achieves higher codebook utilisation than residual vector quantization alternatives, mitigating dead-code collapse. source: §3.2.2, §3.2.3 evidence: Grouped Finite Scalar Vector Quantization achieves higher codebook utilisation than residual vector quantization alternatives, mitigating dead-code collapse. confidence: high relevance: medium - claim_id: real_time_tts_inference_with_low_first_packet_latency_is role: supports claim: Real-time TTS inference with low first-packet latency is achievable on consumer GPU hardware through standard inference optimisations without architectural compromise. source: §4.2 evidence: Real-time TTS inference with low first-packet latency is achievable on consumer GPU hardware through standard inference optimisations without architectural compromise. confidence: high relevance: low limitations: - The entire experimental evaluation is conducted on a proprietary test set with undisclosed composition and size. No public benchmark is used, making it impossible to independently verify the claimed superiority over CosyVoice and F5-TTS or to compare against the broader literature. - The MOS evaluation uses "naive listeners" rather than trained raters or crowdsourced panels following standard listening test protocols (e.g. ITU-T P.800), which may inflate scores relative to conventional evaluations. The paper does not report model size, training compute, or inference memory requirements in full, limiting reproducibility. DPO training details are omitted from the main training description. The paper does not evaluate cross-lingual transfer or accent preservation, which are claimed motivations for the non-G2P design. It is also unclear how the system handles low-resource languages beyond the eight listed in the training data. caveats: [] - id: 2025.chipsal-1.18 published_date: "2025-01-19" entry_date: '2026-07-30' year: 2025 venue: CHiPSAL 2025 task: - TTS - evaluation architecture: - GAN - transformer-enc-dec relevance: high evidence_role: - core_evidence - acceleration_evidence - control_evidence current_role: minor method_family: - adversarial_waveform_vocoders - transformer_conditioned_vocoder_systems claims: - claim_id: gan_based_vocoders_offer_substantially_lower_inference_latency_than_autoregressive role: supports claim: GAN-based vocoders offer substantially lower inference latency than autoregressive vocoders when generating speech from mel-spectrograms. source: §5.1, Table 4 evidence: Tacotron2 + MelGAN synthesises speech in 0.142 s versus 1,320 s for Tacotron2 + WaveNet on an NVIDIA GTX 960M GPU, owing to MelGAN's parallel waveform generation. confidence: high relevance: high - claim_id: gan_based_vocoders_achieve_comparable_or_superior_perceptual_quality_to role: supports claim: GAN-based vocoders achieve comparable or superior perceptual quality to autoregressive vocoders in low-resource language TTS. source: §5.2, Tables 5–7; §3.4.2, Figure 2 evidence: Tacotron2 + MelGAN outperforms Tacotron2 + WaveNet in MOS on both Nepali test sets (4.245 vs. 3.65 on OpenSLR; 2.885 vs. 2.31 on male voice data), confirmed by lower MCD on the OpenSLR set. confidence: high relevance: high - claim_id: naturalness_scores_for_nepali_tts_systems_trained_on_multi_speaker role: complicates claim: Naturalness scores for Nepali TTS systems trained on multi-speaker male voice data are substantially lower than those achieved on curated single-accent female corpora. source: §5.2, Tables 5–7 evidence: Both Tacotron2 + MelGAN and Tacotron2 + WaveNet score more than 1 MOS point lower on the male voice dataset relative to OpenSLR, attributable to speaker variability and limited data volume, with the gap persisting across both objective (MCD) and subjective (MOS) metrics. confidence: high relevance: low - claim_id: performance_of_well_studied_vocoders_on_low_resource_languages_lags role: complicates claim: Performance of well-studied vocoders on low-resource languages lags behind their English-language benchmarks, even when the acoustic model and training procedure are unchanged. source: §5.3, Table 8 evidence: Tacotron2 + WaveNet achieves 3.65 MOS on Nepali OpenSLR versus 4.53 MOS reported for WaveNet on US English data; Tacotron2 + MelGAN achieves 2.885 MOS on the male voice set, below MelGAN's 3.49-3.61 MOS on English (LJSpeech, VCTK). confidence: high relevance: medium limitations: - The MOS evaluation is based on only 40 participants rating four samples per model. This is too small a listener pool for statistically robust conclusions, and no statistical significance tests are reported. The MOS advantage of MelGAN should be interpreted with caution. - The male voice dataset is self-collected from Nepal Television (approximately 2,100 segments) and is not publicly released, limiting reproducibility. Both models struggle with numbers and symbols in Nepali text, an acknowledged limitation with no proposed fix. The study does not explore fine-tuning pre-trained vocoders from English to Nepali rather than training from scratch, which could be a more data-efficient baseline for future work. caveats: [] - id: '2502.05512' published_date: "2025-02-08" entry_date: '2026-07-30' year: 2025 venue: arXiv task: - TTS architecture: - autoregressive-LM - GAN relevance: medium evidence_role: - architecture_variant - acceleration_evidence - control_evidence current_role: active_evidence method_family: - autoregressive_waveform_and_token_decoders - adversarial_waveform_vocoders claims: - claim_id: speaker_conditioning_via_a_multi_reference_conformer_perceiver_improves_zero role: supports claim: Speaker conditioning via a multi-reference Conformer Perceiver improves zero-shot voice cloning stability and timbre consistency over single-vector speaker embeddings. source: §2.3, Table 4 evidence: Speaker conditioning via a multi-reference Conformer Perceiver improves zero-shot voice cloning stability and timbre consistency over single-vector speaker embeddings. confidence: high relevance: medium - claim_id: direct_waveform_decoding_from_lm_hidden_states_via_a_gan role: supports claim: Direct waveform decoding from LM hidden states via a GAN vocoder achieves competitive audio quality with faster inference than diffusion-based intermediate representation decoding. source: §2.4, Table 5 evidence: Direct waveform decoding from LM hidden states via a GAN vocoder achieves competitive audio quality with faster inference than diffusion-based intermediate representation decoding. confidence: high relevance: high - claim_id: fsq_reaches_near_100_codebook_utilisation_with_less_training_data role: supports claim: FSQ reaches near-100% codebook utilisation with less training data than VQ, though VQ converges to similar utilisation with sufficient data scale. source: §3.3.2, Figure 2 evidence: FSQ reaches near-100% codebook utilisation with less training data than VQ, though VQ converges to similar utilisation with sufficient data scale. confidence: high relevance: medium - claim_id: hybrid_character_pinyin_tokenisation_enables_reliable_correction_of_chinese_polyphonic role: supports claim: Hybrid character-pinyin tokenisation enables reliable correction of Chinese polyphonic character mispronunciations at inference time without requiring a separate grapheme-to-phoneme module. source: §3.3.1, Table 2 evidence: Hybrid character-pinyin tokenisation enables reliable correction of Chinese polyphonic character mispronunciations at inference time without requiring a separate grapheme-to-phoneme module. confidence: high relevance: medium limitations: - Model size is not reported, and the training data pipeline uses proprietary internet-sourced audio with pseudo-labels from commercial ASR; neither the data nor the code is released, limiting reproducibility. - The system is limited to Chinese and English, with acknowledged weak emotional expression replication. Instruction-based voice generation is explicitly unsupported. The MOS evaluation relies on 100 samples from an unspecified test set distribution, and the SPK-SIM metric uses ERes2Net rather than a standardised model, making direct comparison with published baselines difficult. The paper does not report streaming latency or real-time factor, despite positioning the hybrid architecture as streaming-capable. caveats: [] - id: iclr-2025-tQ1PmLfPBL published_date: "2025-04-24" entry_date: '2026-07-30' year: 2025 venue: ICLR task: - TTS architecture: - flow-matching relevance: medium evidence_role: - architecture_variant - acceleration_evidence current_role: influential method_family: - flow_matching_waveform_generation claims: - claim_id: flow_matching_enables_higher_quality_waveform_generation_than_diffusion_with role: supports claim: Flow matching enables higher-quality waveform generation than diffusion with fewer inference steps. source: §4.6, Table 8 evidence: PeriodWave with CFM and 6 steps achieves UTMOS 3.628 versus PeriodWave with DDPM at 50 steps achieving UTMOS 3.377 on LibriTTS; the CFM model also reaches comparable quality to 50-step DDPM at only 16 steps. confidence: high relevance: low - claim_id: explicit_multi_period_decomposition_in_the_generator_architecture_improves_pitch role: supports claim: Explicit multi-period decomposition in the generator architecture improves pitch accuracy and periodicity over GAN and diffusion vocoders. source: §4.2, §4.6, Table 1, Table 7 evidence: PeriodWave achieves pitch error of 15.04 cents and periodicity 0.0744 on LJSpeech, substantially below BigVGAN (19.02 cents, 0.0782) and all diffusion baselines; ablation shows monotonic improvement as more distinct prime-number periods are added. confidence: high relevance: high - claim_id: single_step_gan_vocoders_achieve_significantly_faster_inference_than_iterative role: complicates claim: Single-step GAN vocoders achieve significantly faster inference than iterative flow-matching vocoders. source: §E, Table 15, Table 16 evidence: PeriodWave at 16 steps runs at 7.48× real-time; HiFi-GAN runs at 166.70× real-time; even PeriodWave at 2 steps (56.36×) is slower than HiFi-GAN, though it already outperforms HiFi-GAN on all quality metrics. confidence: high relevance: high - claim_id: iterative_waveform_generation_reduces_train_inference_mismatch_artefacts_in_two role: supports claim: Iterative waveform generation reduces train-inference mismatch artefacts in two-stage TTS relative to one-step GAN vocoders. source: §4.7, Table 9, §G evidence: In zero-shot TTS with ARDiT-TTS acoustic features, PeriodWave+FreeU achieves MOS 4.07 versus BigVGAN's 4.03 and BigVSAN's 3.99; the iterative refinement allows the vocoder to correct imperfections in generated Mel-spectrograms rather than propagating them. confidence: high relevance: high - claim_id: flow_matching_vocoders_can_decode_neural_codec_tokens_with_streaming role: supports claim: Flow-matching vocoders can decode neural codec tokens with streaming generation and minimal quality degradation. source: §5, Table 10, Table 11 evidence: PeriodWave trained for parallel generation from Mimi (Q=8) tokens achieves CER 2.5% versus Mimi decoder's 3.07%; streaming with single-token delay and 2-step sampling maintains comparable quality (CER 2.45%, UTMOS 3.85 versus parallel 3.93). confidence: high relevance: low limitations: - 'Synthesis speed is the principal limitation: at 16 steps, PeriodWave runs at 7.48× real-time, substantially slower than one-step GAN vocoders (HiFi-GAN: 166×, BigVGAN: 38×). For latency-sensitive applications, the 2-step variant (56×) is practical but still 3× slower than HiFi-GAN.' - High-frequency reproduction remains challenging even with multi-band modeling and FreeU. The single-loss (CFM objective only) training means the model lacks the spectral feedback that GAN discriminators provide, and M-STFT metrics are generally worse than GAN baselines despite better perceptual scores. The authors plan to incorporate short-time Fourier convolution blocks or modified spectral objectives to address this. - The codec streaming mode uses a non-causal architecture with a one-token look-ahead, which introduces a small latency penalty. In-context streaming generation for longer sequences is identified as future work. Evaluation is confined to English speech and music; generalisation to other languages, accents, and audio domains remains untested in this paper. caveats: [] - id: iclr-2025-uxDFlPGRLX published_date: "2025-04-24" entry_date: '2026-07-30' year: 2025 venue: ICLR task: - codec architecture: - flow-matching - GAN relevance: high evidence_role: - core_evidence - evaluation_caution current_role: active_evidence method_family: - flow_matching_waveform_generation - adversarial_waveform_vocoders claims: - claim_id: a_flow_matching_based_stochastic_postfilter_conditioned_on_a_deterministic role: supports claim: A flow matching-based stochastic postfilter conditioned on a deterministic codec's decoder output can replace adversarial training while achieving comparable subjective quality to a GAN-based codec. source: §5.2, Figure 6 evidence: MUSHRA listening tests (Test A, 11 expert listeners) show no significant difference between FlowDec-75m and DAC-75 score distributions at matched bitrates of 4.5 and 7.5 kbit/s. confidence: high relevance: high - claim_id: coupling_the_flow_matching_source_distribution_to_the_conditioning_signal role: supports claim: Coupling the flow matching source distribution to the conditioning signal (rather than sampling it independently) removes the need for minibatch optimal-transport solvers and improves postfilter sample quality at low inference budgets. source: §5.1, Table 4 evidence: At NFE=6, the proposed coupled formulation achieves FAD×100 of 1.62 versus 145.3 for the diffusion-based ScoreDec baseline and approximately 29 for an alternative constant-σ flow matching formulation, on the same underlying codec and test set. confidence: high relevance: low - claim_id: improving_perceptual_distance_metrics_fad_via_generative_postfiltering_trades_off role: complicates claim: Improving perceptual distance metrics (FAD) via generative postfiltering trades off against intrusive distortion metrics relative to discriminator-trained codecs. source: §5.1, Figure 4, Figure 5 evidence: Retrained non-adversarial DAC (NDAC) generally outperforms FlowDec on SI-SDR and fwSSNR even though FlowDec achieves better FAD, consistent with the perception-distortion tradeoff; the gap is small in the perceptually weighted fwSSNR but clear in SI-SDR. confidence: high relevance: high - claim_id: generative_postfilters_trained_with_vanilla_score_or_flow_matching_formulations role: complicates claim: Generative postfilters trained with vanilla score- or flow-matching formulations using a fixed-variance or independent prior can fail to converge to the target signal or require expensive multi-step inference to reach acceptable quality. source: §5.1, Table 4 evidence: ScoreDec (diffusion-based postfilter) produces unusable audio quality at a reduced inference budget of 6 function evaluations (FAD×100 = 145.3, SI-SDR = -27.23), only becoming competitive at roughly 50 evaluations; a constant-σ flow matching variant also underperforms the proposed coupled formulation at NFE=6. confidence: high relevance: medium limitations: - The proposed codec, like the DAC baseline it builds on, uses a noncausal architecture and is not streaming-capable, limiting applicability to real-time communication settings where low-latency, causal processing is required. - The authors note this could be addressed with a causal DNN design in future work, but no such variant is evaluated in this paper. Additionally, the postfilter and underlying codec are trained in two separate stages; joint training is identified as a possible quality improvement but is left unexplored due to potential training instability. The NCSN++ backbone used for the postfilter was originally designed for images, and the authors suggest audio-specific architectures could further improve quality, indicating the current network design is not optimized end-to-end for the audio domain. Listening test sample sizes (11 and 10 raters) and clip counts (21 ten-second clips) are modest for the audio-type breakdown analysis reported in the appendix. caveats: [] - id: 2025.findings-naacl.130 published_date: "2025-04-29" entry_date: '2026-07-30' year: 2025 venue: NAACL task: - TTS architecture: - transformer-enc-dec - GAN relevance: medium evidence_role: - architecture_variant - control_evidence - evaluation_caution current_role: active_evidence method_family: - transformer_conditioned_vocoder_systems - adversarial_waveform_vocoders claims: - claim_id: in_video_to_speech_synthesis_discrete_acoustic_unit_intermediate_representations role: supports claim: In video-to-speech synthesis, discrete acoustic-unit intermediate representations preserve speech content but discard speaker-identifying acoustic detail relative to continuous Mel-spectrogram representations. source: §3, Table 1, Figure 1 evidence: With ground-truth input held fixed, Unit-HiFiGAN scores SECS 0.555 / EER 40.52 versus HiFi-GAN's SECS 0.894 / EER 22.96, and a Mel-spectrogram visual comparison shows the unit-based vocoder output diverges from ground truth in the frequency domain. confidence: high relevance: medium - claim_id: audio_visual_pre_trained_visual_encoders_can_supply_enough_speaker role: supports claim: Audio-visual pre-trained visual encoders can supply enough speaker-identity information from silent video to make explicit speaker embeddings unnecessary in video-to-speech synthesis. source: §5.2, Table 4 evidence: DiVISe attains the best or near-best SECS scores on LRS2 (0.609) and LRS3 (0.624) among all compared methods, including several that use audio speaker embeddings during training or inference (SVTS, Multi-Task). confidence: high relevance: medium - claim_id: gains_in_objective_intelligibility_metrics_from_architectural_changes_do_not role: complicates claim: Gains in objective intelligibility metrics from architectural changes do not necessarily transfer to subjective audio-quality preference, especially when speaker-identity cues are removed from the listening context. source: Appendix C, Table 16 evidence: In an audio-only MOS test without speaker reference images, Unit-HiFiGAN scores higher (4.37±0.11) than the Mel-based HiFi-GAN (4.24±0.12), the inverse of the speaker-matching and intelligibility rankings observed elsewhere in the paper. confidence: high relevance: medium - claim_id: the_benefit_of_audio_visual_pre_training_for_video_to role: refines claim: The benefit of audio-visual pre-training for video-to-speech synthesis is not uniform across model components; it most strongly aids the component responsible for output representation choice rather than uniformly improving all downstream metrics. source: §5.4.2, Table 7 evidence: Removing pre-training degrades DiVISe across all four reported metrics (SECS, EER, ESTOI, WER), but for ReVISE removing pre-training mainly hurts intelligibility (WER 36.03 to 77.24) while leaving speaker metrics comparatively unaffected (SECS 0.5384 to 0.5304). confidence: high relevance: medium limitations: - Reported WER remains high (35.68-36.24% in the full-resource setting), and the paper's own subjective MOS audio-quality test (without speaker context) ranks the proposed Mel-based vocoder below the unit-based baseline, indicating the speaker-preservation gains are not accompanied by a clear audio-quality win in isolation. - The paper requires the same heavy mouth-region preprocessing pipeline as AV-HuBERT, which the authors note limits real-time applicability (§8, Limitations). Evaluation is restricted to English-only corpora (LRS2, LRS3); the authors explicitly flag multilingual generalization as untested due to compute constraints. The model's WER substantially trails dedicated ASR or text-conditioned TTS systems, reflecting the inherent difficulty of inferring content purely from lip movements without textual or acoustic priors. Latency analysis (§6) shows the conformer module, while improving intelligibility, reduces throughput relative to a no-conformer variant, an explicit accuracy/speed trade-off the paper surfaces but does not resolve. caveats: [] - id: 2025.findings-naacl.279 published_date: "2025-04-29" entry_date: '2026-07-30' year: 2025 venue: NAACL task: - TTS architecture: - autoregressive-LM - GAN - hybrid relevance: medium evidence_role: - architecture_variant - control_evidence - evaluation_caution current_role: active_evidence method_family: - autoregressive_waveform_and_token_decoders - adversarial_waveform_vocoders - hybrid_signal_neural_vocoders claims: - claim_id: few_shot_fine_tuning_on_a_small_amount_of_target role: supports claim: Few-shot fine-tuning on a small amount of target-speaker studio audio substantially improves naturalness, clarity, and speaker similarity over a zero-shot pretrained multilingual TTS model, even when the zero-shot model was pretrained on thousands of hours of in-language speech. source: §4, Table 1 evidence: Fine-tuning on 20 minutes per speaker (4 speakers, 10 epochs) raises SMOS from 4.456 (BnTTS-0, zero-shot) to 4.624 (BnTTS-n, few-shot) and SECS (reference) from 0.529 to 0.548 on BnStudioEval. confidence: high relevance: low - claim_id: expressive_conversationally_prosodic_tts_output_can_score_worse_on_asr role: complicates claim: Expressive, conversationally-prosodic TTS output can score worse on ASR-based intelligibility metrics than flatter-sounding commercial systems, even when subjective listening scores favor the expressive system. source: §4, "High CER in Text Generation", Tables 1–2 evidence: BnTTS-n and BnTTS-0 show higher CER than Google Cloud TTS and Azure TTS on both BnStudioEval and BnTTSTextEval despite BnTTS-n achieving the highest SMOS, Naturalness, and Clarity scores overall; the authors attribute this to ASR systems being better suited to standardized, less expressive speech patterns. confidence: high relevance: low - claim_id: autoregressive_codec_based_tts_models_trained_primarily_on_longer_utterances role: complicates claim: Autoregressive codec-based TTS models trained primarily on longer utterances can fail to generate short sequences reliably under default sampling settings. source: §4, "Effect of Sampling and Prompt Length on Short Speech Generation", Table 3 evidence: Under default generation settings (T=0.85, TopK=50), the model achieves only 0.699 DurationEquality and 0.081 CER on a 23-sample short-text subset (under 30 characters); adjusting prompt length and sampling temperature/top-K improves this to 0.827 DurationEquality and 0.015 CER. confidence: high relevance: low - claim_id: adapting_a_multilingual_zero_shot_tts_architecture_to_a_new role: refines claim: Adapting a multilingual zero-shot TTS architecture to a new low-resource language via continual pretraining can match or exceed existing language-specific commercial and open-source systems without requiring training from scratch. source: §3 "BnTTS model", §7 "Limitations", Tables 1–2 evidence: BnTTS-n, initialized from the pretrained XTTS checkpoint and continually pretrained on 3.85k hours of Bengali data, outperforms IndicTTS, Google Cloud TTS, and Azure TTS on SMOS, Naturalness, and Clarity across both evaluation datasets, while the authors note training from scratch was not attempted due to resource constraints. confidence: high relevance: low limitations: - The system struggles to adapt to speakers with unique vocal traits absent from training, and fails to generate short sequences (under two words or 20 characters) in some cases even after mitigation; these failures stem from limitations inherited from the underlying XTTS foundation model rather than being fully resolved by the Bangla adaptation. - The authors did not train XTTS from scratch for Bangla, relying instead on continual pretraining due to resource constraints, leaving the ceiling of a from-scratch Bangla model untested. A substantial portion of the pretraining corpus (3636 of 3856 hours) is pseudo-labeled and not manually reviewed, which may introduce label noise not fully captured by the reported metrics. The evaluation is limited to four in-house speakers for the few-shot speaker-adaptation setting, and dialectal variation within Bangla is not addressed. caveats: [] - id: 2025.naacl-long.110 published_date: "2025-04-29" entry_date: '2026-07-30' year: 2025 venue: NAACL task: - TTS architecture: - flow-matching relevance: medium evidence_role: - architecture_variant current_role: active_evidence method_family: - flow_matching_waveform_generation claims: - claim_id: a_mel_conditioned_prior_that_matches_the_energy_distribution_of role: supports claim: A mel-conditioned prior that matches the energy distribution of the target signal reduces transportation cost in flow matching vocoders and substantially improves one-step generation quality. source: §3.1, Table 4 evidence: A mel-conditioned prior that matches the energy distribution of the target signal reduces transportation cost in flow matching vocoders and substantially improves one-step generation quality. confidence: high relevance: low - claim_id: reparameterizing_the_flow_matching_objective_to_directly_predict_clean_output role: supports claim: Reparameterizing the flow matching objective to directly predict clean output, rather than random derivatives, enables auxiliary perceptual losses (STFT, mel) and periodic activation functions that cannot otherwise be applied. source: §3.2, Table 4 evidence: Reparameterizing the flow matching objective to directly predict clean output, rather than random derivatives, enables auxiliary perceptual losses (STFT, mel) and periodic activation functions that cannot otherwise be applied. confidence: high relevance: low - claim_id: consistency_distillation_can_be_adapted_to_flow_matching_vocoders_with role: supports claim: Consistency distillation can be adapted to flow matching vocoders with dynamic priors, enabling single-step inference at near-GAN speeds without adversarial training. source: §3.3, Table 3 evidence: Consistency distillation can be adapted to flow matching vocoders with dynamic priors, enabling single-step inference at near-GAN speeds without adversarial training. confidence: high relevance: high - claim_id: flow_matching_vocoders_trained_with_auxiliary_perceptual_losses_generalize_better role: supports claim: Flow matching vocoders trained with auxiliary perceptual losses generalize better to out-of-distribution audio (music) than diffusion vocoders trained without them. source: §4.4, Table 2 evidence: Flow matching vocoders trained with auxiliary perceptual losses generalize better to out-of-distribution audio (music) than diffusion vocoders trained without them. confidence: high relevance: low limitations: - The model is a vocoder (mel-spectrogram to waveform); it does not address acoustic modeling (text to mel). The 19.5M-parameter model is evaluated only on English speech. The out-of-distribution music results, while strong, use the vocal track for ASR-based metrics, limiting the scope of the claim. Whether the reparameterization technique generalizes to latent flow matching (i.e., operating in a lower-dimensional space) is unexplored. caveats: [] - id: 2025.naacl-long.591 published_date: "2025-04-29" entry_date: '2026-07-30' year: 2025 venue: NAACL task: - TTS architecture: - transformer-enc-dec relevance: medium evidence_role: - architecture_variant - evaluation_caution current_role: active_evidence method_family: - transformer_conditioned_vocoder_systems claims: - claim_id: monotonic_alignment_can_be_learned_end_to_end_as_a role: supports claim: Monotonic alignment can be learned end-to-end as a latent property of an encoder-decoder TTS model via backpropagation, without requiring forced alignments or dynamic programming during training. source: §3.3, §3.4 evidence: VAT's alignment layer learns continuous alignment positions through interpolated relative position biases (IRPBs); alignment trajectories emerge from joint training with no external supervision, and generalise to utterances far longer than the training distribution. confidence: high relevance: low - claim_id: augmenting_cross_attention_with_a_learned_monotonic_alignment_position_enables role: supports claim: Augmenting cross-attention with a learned monotonic alignment position enables unbounded length generalisation in encoder-decoder TTS without degrading naturalness relative to an unmodified Transformer baseline. source: §5.1, §5.2, §5.3, Table 1 evidence: VAT achieves near-zero CER on inputs up to 1500 characters (~90 seconds) despite training only on utterances up to 9.6 seconds, while matching the T5 baseline in side-by-side naturalness evaluations (SxS -0.06 ± 0.14 on Lessac, 0.01 ± 0.14 on LibriTTS). confidence: high relevance: low - claim_id: standard_mos_evaluations_are_insufficient_to_surface_robustness_failures_in role: complicates claim: Standard MOS evaluations are insufficient to surface robustness failures in autoregressive TTS because raters cannot detect dropped or repeated words without access to target transcripts. source: §5.1, §5.2, Table 1 evidence: The T5 baseline achieves overlapping MOS with VAT (3.75 vs. 3.68 on Lessac) while producing a CER of 10.2 versus VAT's 3.3; the perceptual quality rating is statistically indistinguishable despite systematic robustness failures. confidence: high relevance: low - claim_id: autoregressive_transformer_tts_without_explicit_alignment_guidance_fails_on_repeated role: complicates claim: Autoregressive Transformer TTS without explicit alignment guidance fails on repeated words even within training sequence length limits. source: §5.4 evidence: The T5 baseline makes errors on 14 of 27 (52%) repeated-word test phrases, including phrases with as few as 2 repetitions of a single word, while VAT makes zero errors across all 27 templates. confidence: high relevance: low - claim_id: duration_based_tts_achieves_better_asr_measured_character_error_rates role: refines claim: Duration-based TTS achieves better ASR-measured character error rates than expressive autoregressive models, but the gain is attributable to hyper-intelligible, monotone synthesis rather than superior text coverage. source: §5.3, Table 1 evidence: NAT achieves CER 3.3 on LibriTTS, below both VAT (4.6) and ground truth (3.6), yet VAT is preferred over NAT in naturalness side-by-sides because NAT's unsupervised duration predictor produces robotic, monotonous prosody. confidence: high relevance: low limitations: - Training speed is affected by the need to compute alignment positions serially during training, imposing a 12-20% slowdown relative to the T5 baseline depending on model scale. All experiments use English and a speaker-conditioned (non-zero-shot) setting; generalisation to other languages and to audio-prompted zero-shot scenarios is untested. Evaluation compares against T5, Tacotron-GMMA, and NAT; no direct comparison with codec LM systems (VALL-E, SPEAR-TTS, MQTTS) is provided, which the paper attributes to incompatible dataset scales and evaluation protocols. Hyper-parameter choices for the alignment layer, IRPB initialization, and maximum distance penalty are reported but not systematically ablated. caveats: [] - id: 2025.naacl-short.65 published_date: "2025-04-29" entry_date: '2026-07-30' year: 2025 venue: NAACL task: - TTS architecture: - hybrid - GAN relevance: medium evidence_role: - architecture_variant - control_evidence current_role: active_evidence method_family: - hybrid_signal_neural_vocoders - adversarial_waveform_vocoders claims: - claim_id: ssl_feature_spaces_from_pre_trained_models_encode_cross_speaker role: supports claim: SSL feature spaces from pre-trained models encode cross-speaker structure that enables zero-shot voice transfer through nearest-neighbor retrieval, without speaker-specific training data. source: §2.1, Table 1 evidence: kNN-TTS uses WavLM-Large layer 6 features, where frames from different speakers that are linearly close share phonetic information while preserving speaker identity; kNN retrieval over these features achieves SECS 0.72 and competitive MOS scores trained only on 24h of single-speaker LJSpeech data. confidence: high relevance: medium - claim_id: zero_shot_multi_speaker_tts_competitive_with_large_multi_speaker role: supports claim: Zero-shot multi-speaker TTS competitive with large multi-speaker end-to-end systems can be achieved with single-speaker transcribed training data by delegating speaker identity to inference-time retrieval. source: §4, Table 1 evidence: GlowkNN-TTS (24h training, single speaker) achieves N-MOS and S-MOS within the confidence intervals of HierSpeech++ (2,796h, 7299 speakers) and XTTS (27,282h, multi-speaker) on LibriSpeech test-clean. confidence: high relevance: low - claim_id: retrieval_based_zero_shot_tts_requires_substantially_more_reference_audio role: complicates claim: Retrieval-based zero-shot TTS requires substantially more reference audio from the target speaker than embedding-based approaches to achieve sufficient quality. source: §Limitations, Figure 3b evidence: kNN-TTS requires approximately 30 seconds of target speaker audio for suitable intelligibility and around 1 minute for speaker similarity to plateau, whereas competing embedding-based systems show diminishing returns beyond 10-30 seconds of reference audio. confidence: high relevance: low - claim_id: frame_level_knn_speaker_transfer_does_not_address_speaker_specific role: complicates claim: Frame-level kNN speaker transfer does not address speaker-specific duration and rhythm, leaving prosodic timing patterns fixed to the training speaker. source: §Limitations "Rhythmic variations" evidence: In kNN-TTS, utterance duration is determined entirely by the single-speaker Text-to-SSL model; frame-by-frame retrieval substitutes voice quality but does not adapt speaking rate or rhythm to the target speaker. confidence: high relevance: medium limitations: - 'The reference audio requirement is a practical limitation: kNN-TTS needs approximately 30 seconds of target speaker audio for usable intelligibility, which is notably higher than embedding-based competitors that can function with shorter clips. This restricts applicability in truly few-shot or single-utterance zero-shot scenarios.' - Duration adaptation to the target speaker is not addressed; the speaking rate and rhythm of the output always reflect the training speaker (LJSpeech). The paper proposes Urhythmic-style rhythm modeling as future work. Evaluation is English-only, and while the authors note potential for cross-lingual transfer (via kNN-VC cross-lingual capabilities), this is not demonstrated. Using mel-spectrogram features as an alternative to SSL features was ablated and found completely ineffective, confirming the dependency on WavLM's particular representational structure. caveats: [] - id: 2025.americasnlp-1.1 published_date: "2025-05-04" entry_date: '2026-07-30' year: 2025 venue: AmericasNLP 2025 task: - TTS architecture: - transformer-enc-dec - GAN relevance: medium evidence_role: - architecture_variant - control_evidence - evaluation_caution current_role: active_evidence method_family: - transformer_conditioned_vocoder_systems - adversarial_waveform_vocoders claims: - claim_id: transfer_learning_from_a_closely_related_language_can_enable_intelligible role: supports claim: Transfer learning from a closely related language can enable intelligible TTS for an indigenous language with fewer than 5 hours of training data. source: §4.1, §5.1 evidence: Transfer learning from a closely related language can enable intelligible TTS for an indigenous language with fewer than 5 hours of training data. confidence: high relevance: low - claim_id: subjective_evaluation_by_native_speakers_is_necessary_when_objective_metrics role: supports claim: Subjective evaluation by native speakers is necessary when objective metrics such as PESQ and POLQA cannot be applied due to absent high-quality reference recordings. source: §5.3 evidence: Subjective evaluation by native speakers is necessary when objective metrics such as PESQ and POLQA cannot be applied due to absent high-quality reference recordings. confidence: high relevance: low - claim_id: pronunciation_drift_in_a_native_speaker_living_outside_the_language role: supports claim: Pronunciation drift in a native speaker living outside the language community can cause natural speech to score lower on intelligibility than synthetic speech on standardised evaluation. source: §5.4 evidence: Pronunciation drift in a native speaker living outside the language community can cause natural speech to score lower on intelligibility than synthetic speech on standardised evaluation. confidence: high relevance: low - claim_id: gan_based_vocoders_initialized_from_a_universal_pretrained_model_converge role: supports claim: GAN-based vocoders initialized from a universal pretrained model converge rapidly when the target speaker is similar to the pretraining distribution. source: §4.2, §5.2 evidence: GAN-based vocoders initialized from a universal pretrained model converge rapidly when the target speaker is similar to the pretraining distribution. confidence: high relevance: high limitations: - '- Corpus is only 4h 37m from a single female speaker, limiting prosodic and phonetic diversity and voice coverage. - Transfer learning from Spanish may introduce Spanish phonotactic biases that are not present in Shipibo-Konibo. - Evaluation relied entirely on subjective metrics (no PESQ/POLQA due to lack of reference data); subjectivity and evaluator bias are acknowledged. - Two writing conventions for Shipibo-Konibo exist; the 2015 standard was used, limiting applicability in communities using older orthographies. - The speaker''s extended residence in Lima introduced pronunciation drift that complicates the natural vs. synthetic comparison. - The model has not been tested on adaptation to other Panoan or Amazonian languages yet, though the framework is intended to generalize.' caveats: [] - id: '2507.01611' published_date: "2025-07-02" entry_date: '2026-07-30' year: 2025 venue: arXiv task: - TTS architecture: - GAN relevance: high evidence_role: - core_evidence - evaluation_caution current_role: active_evidence method_family: - adversarial_waveform_vocoders claims: - claim_id: explicit_spectral_envelope_modeling_in_a_neural_vocoder_enables_reliable role: supports claim: Explicit spectral envelope modeling in a neural vocoder enables reliable pitch modification beyond the training frequency range, where purely neural waveform predictors fail. source: §V.B, §V.D, Tables II, VI, Fig. 6 evidence: QHARMA-GAN maintains f0 RMSE of 0.08-0.11 Hz across pitch-scale factors from 0.5x to 2x on VCTK, while Vocos fails catastrophically (0.65-0.67 Hz); soprano singing voices from an unseen language are reproduced with full harmonic range, whereas HiFi-GAN and Vocos generate only low-frequency harmonics. confidence: high relevance: high - claim_id: incorporating_classical_signal_processing_constraints_into_a_neural_vocoder_reduces role: supports claim: Incorporating classical signal-processing constraints into a neural vocoder reduces data requirements without sacrificing subjective quality. source: §V.D, Table VII evidence: QHARMA-GAN trained on 919 LJSpeech utterances achieves MOS 3.85 compared to HiFi-GAN's 3.53 under the same data constraint; the gap is attributed to HiFi-GAN's overfitting on insufficient data, while QHARMA-GAN's hybrid structure provides analytical priors that reduce the burden on the neural component. confidence: high relevance: high - claim_id: hybrid_vocoders_that_analytically_reconstruct_phase_can_achieve_higher_subjective role: complicates claim: Hybrid vocoders that analytically reconstruct phase can achieve higher subjective naturalness than end-to-end neural vocoders despite lower PESQ scores, suggesting objective spectral distance metrics do not fully capture perceptual frequency smoothness. source: §V.B, Table III evidence: QHARMA-GAN scores MOS 4.21 on VCTK vs HiFi-GAN's 4.08, while PESQ is comparable (3.14 vs 3.14) and both trail QHM's PESQ of 3.45 despite QHM scoring the lowest subjective MOS (4.07); the divergence is attributed to QHARMA-GAN's ability to reduce frequency distortion that HiFi-GAN exhibits in some samples. confidence: high relevance: medium - claim_id: v_uv_detection_errors_are_a_critical_bottleneck_for_hybrid role: complicates claim: V/UV detection errors are a critical bottleneck for hybrid pitch-modification vocoders, particularly at extreme pitch-raising factors. source: §V.B, Table IV evidence: QHARMA-GAN outperforms WORLD in pitch lowering (MOS 3.01 vs 2.98 at rho=0.5x on VCTK) but underperforms WORLD in pitch raising (2.72 vs 2.82 at rho=2x), attributed to V/UV misclassification reducing harmonic component count in unvoiced segments during upward pitch shifts. confidence: high relevance: medium limitations: - 'Accurate V/UV detection is an unsolved dependency: QHARMA-GAN''s pitch modification quality is bounded by the accuracy of the V/UV classifier, which the paper treats as a fixed external component. The authors identify this as "an urgent research topic for the future," but no solution is proposed.' - QHARMA-GAN requires an additional f0 predictor module when used as part of a TTS backend, unlike mel-spectrogram-to-waveform vocoders that take mel-spectrograms directly as input. This additional module adds pipeline complexity and introduces a potential error source in the f0 prediction step. - PESQ scores on VCTK are below QHM (3.14 vs 3.45) and below Vocos (3.15), indicating that the ARMA-based amplitude estimation introduces spectral distortion despite improving perceptual frequency smoothness. The objective-subjective gap is noted but not fully explained. - Evaluations use proprietary metrics (UTMOS) alongside standard ones; no listening test confidence intervals are reported for the modification experiments (Table IV), making statistical significance of cross-method differences difficult to assess. caveats: [] - id: '2507.08319' published_date: "2025-07-11" entry_date: '2026-07-30' year: 2025 venue: arXiv task: - TTS architecture: - transformer-enc-dec relevance: medium evidence_role: - architecture_variant - control_evidence current_role: active_evidence method_family: - transformer_conditioned_vocoder_systems claims: - claim_id: active_learning_with_model_informed_informativeness_criteria_produces_more_data role: supports claim: Active learning with model-informed informativeness criteria produces more data-efficient TTS corpora than model-agnostic static data selection at equal corpus size. source: §III.B, Table I evidence: The proposed method achieves 71.0% synthesizable speaker ratio at 3,943 utterances versus 66.6% for the static baseline of the same size, measured by pseudo-MOS exceedance rate. confidence: high relevance: low - claim_id: data_selection_methods_that_optimize_synthesizable_speaker_coverage_may_reduce role: complicates claim: Data selection methods that optimize synthesizable speaker coverage may reduce synthesis quality for speakers already near the quality threshold. source: §III.B, Figure 3 evidence: The proposed method underperforms the static baseline for speakers with pseudo-MOS above 2.6, because the informativeness criterion explicitly excludes speakers the current model can already synthesize well. confidence: high relevance: medium - claim_id: tts_corpus_quality_estimators_remain_effective_when_trained_on_a role: supports claim: TTS corpus quality estimators remain effective when trained on a small fraction of candidate data rather than requiring full candidate set pre-processing. source: §III.B, Figure 7 evidence: The data quality estimator trained on 10% of the YouTube candidate pool shows a strong correlation with the estimator trained on 100% of the data, with no significant change in selection behavior. confidence: high relevance: low - claim_id: diffusion_models_better_capture_complex_speaker_embedding_distributions_than_gaussian role: supports claim: Diffusion models better capture complex speaker embedding distributions than Gaussian mixture models for speaker generation tasks with limited training samples. source: §III.B, Figure 5 evidence: The diffusion model achieves Wasserstein distance to held-out test speakers that is more than twice the standard deviation lower than GMMs across all cluster sizes (M=1..10). confidence: high relevance: low limitations: - All experiments are conducted on Japanese YouTube data only. Generalizability to other languages, recording conditions, or web data sources is untested, and the quality threshold calibration requires a language-matched studio corpus. - The active learning loop is evaluated with only two iterations (K=2). Whether additional iterations continue to improve coverage or reach diminishing returns is not explored. The main baseline comparison is the authors' own prior static selection method; no comparison against other active learning or subset selection approaches from the computer vision or NLP literature is included. The evaluation uses pseudo-MOS (an automatic quality predictor) rather than subjective listening tests, so the absolute synthesis quality of the final corpora is not independently verified. caveats: [] - id: 2025.findings-acl.115 published_date: "2025-07-27" entry_date: '2026-07-30' year: 2025 venue: ACL task: - SCA - TTS architecture: - autoregressive-LM relevance: medium evidence_role: - architecture_variant - control_evidence current_role: active_evidence method_family: - autoregressive_waveform_and_token_decoders claims: - claim_id: grouping_discrete_semantic_tokens_during_autoregressive_generation_reduces_training_and role: supports claim: Grouping discrete semantic tokens during autoregressive generation reduces training and inference costs by alleviating the frequency mismatch between text and audio token streams. source: §5.4.1, Table 6 evidence: Semantic Group Modeling with G=3 reduces SLAM-Omni training from 126 to 60 GPU hours and ASR-WER from 18.23% (G=1) to 4.54%, with G=3 providing the best quality-efficiency trade-off across five group sizes tested. confidence: high relevance: medium - claim_id: multi_stage_pre_training_on_modality_specific_tasks_for_spoken role: refines claim: Multi-stage pre-training on modality-specific tasks for spoken dialogue systems can degrade instruction-following ability despite improving modality alignment metrics. source: §5.4.2, Table 7 evidence: ASR pre-training reduces ChatGPT Score from 39.32 to 34.02 and TTS pre-training to 27.22, while ASR-WER improves only marginally (4.38% / 4.53% vs. 4.54%), indicating that modality-specific pre-training shifts the model away from general instruction-following capabilities. confidence: high relevance: medium - claim_id: zero_shot_timbre_control_in_spoken_dialogue_systems_achieves_competitive role: complicates claim: Zero-shot timbre control in spoken dialogue systems achieves competitive speaker similarity but remains constrained by training data volume relative to dedicated TTS systems. source: §5.2, Table 5 evidence: SLAM-Omni reaches SIM-o of 0.517, comparable to FireRedTTS (0.486) but below CosyVoice2 (0.684), with the gap attributed to approximately 50x less training data; limited and less diverse data cause ambiguities in timbre rendering during vocoder synthesis. confidence: high relevance: low - claim_id: compressing_multi_turn_dialogue_history_to_text_representation_sacrifices_non role: complicates claim: Compressing multi-turn dialogue history to text representation sacrifices non-verbal paralinguistic context that may be important for maintaining dialogue coherence across turns. source: §3.5, §6 Limitations evidence: Historical Text Prompting stores only text history, explicitly trading away emotional and prosodic signals from previous dialogue turns for computational efficiency; the paper identifies this as a primary limitation affecting scenarios requiring sustained dialogue coherence. confidence: high relevance: medium - claim_id: semantic_token_based_speech_generation_in_spoken_dialogue_systems_provides role: supports claim: Semantic token-based speech generation in spoken dialogue systems provides tighter speech-text alignment than acoustic codec-based approaches, as measured by word error rate between generated speech and corresponding text. source: §5.1, Table 3 evidence: SLAM-Omni achieves the lowest ASR-WER (4.54%) among all evaluated spoken dialogue models, outperforming larger models including Moshi (7.18%), GLM-4-Voice (12.71%), and Freeze-Omni (16.32%), using single-layer semantic tokens rather than multi-codebook acoustic tokens. confidence: high relevance: low limitations: - The finding that single-stage training outperforms multi-stage pre-training is demonstrated only at 0.5B scale with limited data (400K utterances). The authors explicitly note that extending to larger LLMs would require substantially more training data, and whether the result holds at larger scales or with more diverse corpora is untested. - 'Historical text prompting provides computational efficiency at the cost of discarding non-verbal paralinguistic signals (emotion, prosody) from previous dialogue turns. This limits the system''s ability to maintain affective continuity across multi-turn conversations. The evaluation is also narrow: the custom 8-task benchmark measures general spoken interaction but not domain-specific or emotion-aware dialogue quality.' caveats: [] - id: '2507.20731' published_date: "2025-07-28" entry_date: '2026-07-30' year: 2025 venue: arXiv task: - TTS architecture: - GAN relevance: high evidence_role: - core_evidence - acceleration_evidence current_role: active_evidence method_family: - adversarial_waveform_vocoders claims: - claim_id: explicit_incorporation_of_known_linear_degradation_structure_the_mel_filterbank role: supports claim: Explicit incorporation of known linear degradation structure (the mel filterbank) into vocoder design reduces the learning burden on the neural component and improves parameter efficiency. source: §4.3, Table 4 evidence: Removing the RND decomposition and replacing it with direct black-box mel-to-spectrum mapping (ablation id3 vs. id1) degrades PESQ from 3.987 to 3.655 and causes near-total pitch tracking failure on LJSpeech. confidence: high relevance: high - claim_id: enforcing_orthogonality_between_range_space_and_null_space_components_in role: supports claim: Enforcing orthogonality between range-space and null-space components in T-F vocoder design is important; making the decomposition matrices learnable degrades reconstruction quality. source: §4.3, Table 4, Figure 6 evidence: Setting A and A† as learnable (ablation id4) degrades PESQ from 3.987 to 3.645 and V/UV F1 from 0.9714 to 0.1164 on LJSpeech, indicating that the learned matrices no longer maintain the orthogonality property. confidence: high relevance: high - claim_id: t_f_domain_vocoders_can_close_the_quality_gap_with role: supports claim: T-F domain vocoders can close the quality gap with large-scale time-domain GAN vocoders while remaining substantially more parameter-efficient and faster on CPU. source: §4.3, Tables 1, 2, 3 evidence: RNDVoC (3.14M parameters) achieves PESQ 4.226 and MUSHRA 80.74 on LibriTTS, matching or exceeding BigVGAN (112M parameters, PESQ 4.027, MUSHRA 79.33) trained for 1M steps, while running at nearly 10x real-time speed on CPU. confidence: high relevance: high - claim_id: the_quality_advantage_of_t_f_domain_vocoders_over_large role: complicates claim: The quality advantage of T-F domain vocoders over large-scale time-domain models depends on architectural choices for sub-band modeling; naive T-F vocoders remain inferior to well-tuned time-domain GAN vocoders at equivalent scale. source: §4.3, Table 1 evidence: Vocos (13.46M) achieves PESQ 3.522 on LJSpeech while BigVGAN (112M) reaches 4.107; only the proposed hierarchical sub-band design closes this gap at 3.14M parameters. confidence: high relevance: high - claim_id: extending_phase_loss_to_capture_differential_relations_across_all_eight role: supports claim: Extending phase loss to capture differential relations across all eight adjacent T-F bins improves objective vocoder metrics relative to sparse two-direction anti-wrapping losses. source: §4.3, Table 4 evidence: Replacing the proposed omnidirectional phase loss with the prior anti-wrapping formulation (ablation id2) reduces PESQ from 3.987 to 3.892 and MCD from 2.047 to 2.214 on LJSpeech. confidence: high relevance: high limitations: - The evaluation is restricted to single-speaker (LJSpeech) and multi-speaker clean-speech (LibriTTS) benchmarks. The paper includes an out-of-distribution spectral visualisation using a singing voice from MUSDB18, but no quantitative generalisation results are reported for singing, emotional speech, or noisy conditions. Whether the orthogonality-based framework degrades gracefully when the mel-filterbank prior is a poor match to the target signal (e.g., non-speech audio, very high-frequency content) is not examined. The MUSHRA listening test involves 35 participants drawn from an audio signal processing background, which may not reflect listener preferences in deployment scenarios with general audiences. caveats: [] - id: '2508.01796' published_date: "2025-08-03" entry_date: '2026-07-30' year: 2025 venue: arXiv task: - singing architecture: - diffusion - GAN relevance: high evidence_role: - core_evidence current_role: active_evidence method_family: - diffusion_vocoder_and_restoration_systems - adversarial_waveform_vocoders claims: - claim_id: mel_spectrogram_to_waveform_vocoders_systematically_underrepresent_high_frequency_spectral role: supports claim: Mel-spectrogram-to-waveform vocoders systematically underrepresent high-frequency spectral structure, producing a detectable signature that lightweight classifiers can exploit. source: §1, §4.2 evidence: Mel-spectrogram-to-waveform vocoders systematically underrepresent high-frequency spectral structure, producing a detectable signature that lightweight classifiers can exploit. confidence: high relevance: medium - claim_id: inserting_an_explicit_bandwidth_extension_step_prior_to_vocoding_can role: supports claim: Inserting an explicit bandwidth extension step prior to vocoding can improve both spectrogram visual realism and perceived audio quality, but only when the vocoder is redesigned to accommodate the higher-resolution intermediate representation. source: §4.4, Figure 8 evidence: Inserting an explicit bandwidth extension step prior to vocoding can improve both spectrogram visual realism and perceived audio quality, but only when the vocoder is redesigned to accommodate the higher-resolution intermediate representation. confidence: high relevance: high - claim_id: diffusion_models_with_transformer_backbones_can_be_adapted_to_time role: supports claim: Diffusion models with transformer backbones can be adapted to time-frequency spectral data by restricting self-attention to the time axis and using per-frequency learned embeddings to handle the lack of spatial invariance across frequency. source: §3.2 evidence: Diffusion models with transformer backbones can be adapted to time-frequency spectral data by restricting self-attention to the time axis and using per-frequency learned embeddings to handle the lack of spatial invariance across frequency. confidence: high relevance: low - claim_id: spectrogram_visual_realism_and_waveform_perceptual_quality_are_partially_independent role: supports claim: 'Spectrogram visual realism and waveform perceptual quality are partially independent dimensions: a system can improve on one while degrading on the other depending on how the vocoder processes the extended spectral representation.' source: §4.4 evidence: 'Spectrogram visual realism and waveform perceptual quality are partially independent dimensions: a system can improve on one while degrading on the other depending on how the vocoder processes the extended spectral representation.' confidence: high relevance: high limitations: - All training and evaluation data is assembled from a mix of privately recorded singing audio and public datasets under a custom pipeline; no standard public benchmark is used for final comparison, making it difficult to situate results relative to the broader SVS or TTS literature. - The 32-step DPM++ sampling in the LSE model adds meaningful latency compared to a direct mel-to-waveform vocoder; the paper does not report inference-time benchmarks. The work focuses entirely on singing voice at 44.1 kHz — generalisation to speech or other domains is not tested. The spectrogram realism evaluation relies on a classifier trained on the same data distribution, which could favour the proposed method if the discriminator training is not sufficiently diverse. The adversarial-attack framing (evading fake-spectrogram detectors) raises questions about dual use that are not discussed. caveats: [] - id: '2508.05207' published_date: "2025-08-07" entry_date: '2026-07-30' year: 2025 venue: arXiv task: - codec architecture: - GAN relevance: medium evidence_role: - architecture_variant - acceleration_evidence current_role: active_evidence method_family: - adversarial_waveform_vocoders claims: - claim_id: operating_in_the_time_frequency_domain_enables_neural_codecs_to role: supports claim: Operating in the time-frequency domain enables neural codecs to achieve higher perceptual quality for full-band audio than equivalent waveform-domain architectures, especially at low bit rates. source: §2, §4, Table 1 evidence: Operating in the time-frequency domain enables neural codecs to achieve higher perceptual quality for full-band audio than equivalent waveform-domain architectures, especially at low bit rates. confidence: high relevance: medium - claim_id: cross_channel_phase_coherence_in_multi_channel_neural_codecs_requires role: supports claim: Cross-channel phase coherence in multi-channel neural codecs requires joint processing of audio channels in at least some encoder layers, and neither fully independent nor fully joint encoding is optimal. source: §2 evidence: Cross-channel phase coherence in multi-channel neural codecs requires joint processing of audio channels in at least some encoder layers, and neither fully independent nor fully joint encoding is optimal. confidence: high relevance: medium - claim_id: multi_scale_spectral_discriminators_are_effective_for_training_high_quality role: supports claim: Multi-scale spectral discriminators are effective for training high-quality neural codecs without requiring waveform-domain discriminators. source: §3 evidence: Multi-scale spectral discriminators are effective for training high-quality neural codecs without requiring waveform-domain discriminators. confidence: high relevance: medium - claim_id: biased_quantizer_dropout_towards_low_codebook_counts_during_training_improves role: supports claim: Biased quantizer dropout towards low codebook counts during training improves codec quality at the low bit rate end without sacrificing high bit rate performance. source: §3.1.1 evidence: Biased quantizer dropout towards low codebook counts during training improves codec quality at the low bit rate end without sacrificing high bit rate performance. confidence: high relevance: low - claim_id: real_time_streaming_neural_codec_inference_at_48_khz_stereo role: supports claim: Real-time streaming neural codec inference at 48 kHz stereo is achievable on a desktop CPU with an 80 ms architectural latency when using causal convolutions and a minimal look-ahead. source: §1, §2 evidence: Real-time streaming neural codec inference at 48 kHz stereo is achievable on a desktop CPU with an 80 ms architectural latency when using causal convolutions and a minimal look-ahead. confidence: high relevance: low limitations: - Training data is proprietary and the only baseline is DAC; results cannot be reproduced and the comparison does not include SoundStream, EnCodec, or Mimi, leaving SpectroStream's position in the broader codec landscape unclear. - Evaluation is restricted to music (MUSDB18) despite the paper's "general audio" framing. Speech quality at 48 kHz stereo is not reported. The A/B preference protocol does not include a MUSHRA-style anchor, making absolute quality judgements difficult. The latency of 80 ms is described as suitable for streaming but is not benchmarked against real-time constraints in actual deployment. The delayed-fusion fusion point is treated as a design choice found empirically — no ablation is provided to quantify the quality/coherence trade-off as a function of fusion layer depth. caveats: [] - id: '2508.06870' published_date: "2025-08-09" entry_date: '2026-07-30' year: 2025 venue: arXiv task: - TTS architecture: - transformer-enc-dec - GAN relevance: medium evidence_role: - architecture_variant - control_evidence current_role: minor method_family: - transformer_conditioned_vocoder_systems - adversarial_waveform_vocoders claims: - claim_id: standard_neural_tts_pipelines_acoustic_model_neural_vocoder_can_produce role: supports claim: Standard neural TTS pipelines (acoustic model + neural vocoder) can produce intelligible speech for previously unsupported tonal scripts with as little as 40 minutes of single-speaker training data. source: §5.3.5, §5.4 evidence: Standard neural TTS pipelines (acoustic model + neural vocoder) can produce intelligible speech for previously unsupported tonal scripts with as little as 40 minutes of single-speaker training data. confidence: high relevance: high - claim_id: the_absence_of_explicit_tone_modelling_in_a_tacotron_2 role: supports claim: The absence of explicit tone modelling in a Tacotron 2 pipeline leads to degraded prosody on complex utterances in tonal languages, even when the training corpus is small. source: §2.1, §5.4 evidence: The absence of explicit tone modelling in a Tacotron 2 pipeline leads to degraded prosody on complex utterances in tonal languages, even when the training corpus is small. confidence: high relevance: medium - claim_id: developing_a_phoneme_mapping_between_a_novel_script_and_an role: supports claim: Developing a phoneme mapping between a novel script and an established phoneme set (such as ARPAbet) is a necessary and non-trivial prerequisite before standard TTS toolkits can be applied to indigenous writing systems. source: §3.1.4, §6.2 evidence: Developing a phoneme mapping between a novel script and an established phoneme set (such as ARPAbet) is a necessary and non-trivial prerequisite before standard TTS toolkits can be applied to indigenous writing systems. confidence: high relevance: low limitations: - The system is trained on a single speaker with only 40 minutes of audio, and no comparison to any baseline is possible. MOS scores are self-contained and cannot be situated relative to other low-resource systems. - Tonal phonology is not explicitly modelled. Manipuri distinguishes at least two contrastive tones, and the falling tone is marked orthographically in Meitei Mayek, but the paper does not describe any mechanism to condition acoustic output on tone — the tone diacritic is treated as another character in the phoneme table. This likely explains the reported flatness on longer utterances. - The dataset is very small (818 utterances, one speaker) and remains unreleased at time of publication, limiting reproducibility. The practical path forward requires more data, speaker diversity, and either explicit prosody conditioning or fine-tuning from a related-language pre-trained model. caveats: [] - id: '2508.07711' published_date: "2025-08-11" entry_date: '2026-07-30' year: 2025 venue: arXiv task: - TTS architecture: - GAN relevance: high evidence_role: - core_evidence - evaluation_caution current_role: active_evidence method_family: - adversarial_waveform_vocoders claims: - claim_id: adversarial_training_is_not_necessary_for_high_quality_mel_spectrogram role: supports claim: Adversarial training is not necessary for high-quality mel-spectrogram vocoders that explicitly predict the phase spectrum; frequency-weighted phase supervision can substitute for discriminator-provided phase signal. source: §IV.A, Table I evidence: Adversarial training is not necessary for high-quality mel-spectrogram vocoders that explicitly predict the phase spectrum; frequency-weighted phase supervision can substitute for discriminator-provided phase signal. confidence: high relevance: high - claim_id: the_necessity_of_gan_in_neural_vocoders_is_contingent_on role: supports claim: 'The necessity of GAN in neural vocoders is contingent on the presence of explicit phase prediction: removing adversarial training without explicit phase modelling results in a quality gap that targeted improvements do not fully close.' source: §IV.B, Table III evidence: 'The necessity of GAN in neural vocoders is contingent on the presence of explicit phase prediction: removing adversarial training without explicit phase modelling results in a quality gap that targeted improvements do not fully close.' confidence: high relevance: high - claim_id: eliminating_gan_discriminators_from_vocoder_training_substantially_reduces_wall_clock role: supports claim: Eliminating GAN discriminators from vocoder training substantially reduces wall-clock training time and storage requirements with no measurable subjective quality cost for explicit-phase architectures. source: §IV.A, Table I evidence: Eliminating GAN discriminators from vocoder training substantially reduces wall-clock training time and storage requirements with no measurable subjective quality cost for explicit-phase architectures. confidence: high relevance: high - claim_id: automatic_mos_predictors_utmos_and_subjective_mos_scores_can_rank role: supports claim: Automatic MOS predictors (UTMOS) and subjective MOS scores can rank vocoder systems differently, making UTMOS an unreliable sole proxy for naturalness evaluation. source: §IV.A, Table I evidence: Automatic MOS predictors (UTMOS) and subjective MOS scores can rank vocoder systems differently, making UTMOS an unreliable sole proxy for naturalness evaluation. confidence: high relevance: high limitations: - All experiments use VCTK at 16 kHz with a constrained multi-speaker English setup. Generalisation to 24 kHz or 44.1 kHz targets, non-English speakers, noisy or out-of-domain conditions, and downstream TTS pipelines (as opposed to ground-truth mel reconstruction) is untested and explicitly deferred to future work. - 'The comparison to BigVGAN is acknowledged to be partially unfair: BigVGAN is trained and optimised for general-purpose, large-scale audio generation, so the efficiency figures reflect FreeGAN''s advantage in a more constrained setting. Whether FreeGAN''s approach scales to the quality levels BigVGAN achieves in zero-shot conditions remains open.' - The FWAW phase loss design relies on a heuristic exponential weighting scheme across frequency bins with a hyperparameter (ρ = 2.5); how sensitive results are to this choice and whether a learned weighting would improve further is not explored. - Code availability is not reported, which limits reproducibility. caveats: [] - id: '2508.12001' published_date: "2025-08-16" entry_date: '2026-07-30' year: 2025 venue: arXiv task: - TTS architecture: - VAE - GAN - hybrid relevance: high evidence_role: - core_evidence - control_evidence current_role: active_evidence method_family: - vae_codec_vocoder_decoders - adversarial_waveform_vocoders - hybrid_signal_neural_vocoders claims: - claim_id: applying_mixture_of_experts_routing_to_the_duration_predictor_of role: supports claim: Applying Mixture-of-Experts routing to the duration predictor of a non-autoregressive TTS system produces phoneme duration distributions that more closely match ground-truth speaker-specific prosodic variation than deterministic or stochastic single-path predictors. source: §IV-C, Table III evidence: Applying Mixture-of-Experts routing to the duration predictor of a non-autoregressive TTS system produces phoneme duration distributions that more closely match ground-truth speaker-specific prosodic variation than deterministic or stochastic single-path predictors. confidence: high relevance: low - claim_id: improving_duration_prediction_diversity_in_nar_tts_without_upgrading_the role: supports claim: Improving duration prediction diversity in NAR TTS without upgrading the vocoder degrades synthesis quality, because richer prosodic information exceeds the artifact-rejection capacity of standard GAN discriminators. source: §IV-A, Table I evidence: Improving duration prediction diversity in NAR TTS without upgrading the vocoder degrades synthesis quality, because richer prosodic information exceeds the artifact-rejection capacity of standard GAN discriminators. confidence: high relevance: high - claim_id: frequency_domain_and_multi_band_discriminators_combd_and_sbd_are role: supports claim: Frequency-domain and multi-band discriminators (CoMBD and SBD) are more effective than HiFi-GAN's multi-period discriminator at suppressing spectral artifacts when a more expressive duration predictor is used. source: §V-B, Table II evidence: Frequency-domain and multi-band discriminators (CoMBD and SBD) are more effective than HiFi-GAN's multi-period discriminator at suppressing spectral artifacts when a more expressive duration predictor is used. confidence: high relevance: high - claim_id: word_error_rate_is_an_unreliable_proxy_for_prosody_modeling role: supports claim: Word Error Rate is an unreliable proxy for prosody modeling quality in TTS, as simpler prosodic patterns can yield lower WER by reducing ASR system disruption. source: §IV-C, §III evidence: Word Error Rate is an unreliable proxy for prosody modeling quality in TTS, as simpler prosodic patterns can yield lower WER by reducing ASR system disruption. confidence: high relevance: low limitations: - '- The MoE-DP requires the discriminator upgrade to yield any net benefit; neither module alone suffices, which raises questions about robustness when either component is replaced or updated independently. - Evaluation is limited to English datasets; whether MoE-DP''s speaker-specific prosody gains transfer to more typologically diverse languages with different prosodic structures is untested. - The paper notes that current vocoders still struggle under highly diverse prosodic patterns (disharmony components persist even with CoMBD+SBD), pointing to an open challenge for future vocoder designs. - WER is shown to be an unreliable proxy for prosody quality, but the proposed duration-accuracy metric (MFA-based classification) requires ground-truth phoneme duration annotations, limiting its generalizability. - No code or demo links are provided, limiting reproducibility. - Inference RTF on CPU (0.046) is not compared directly against autoregressive or diffusion-based competitors that operate on the same hardware, making efficiency claims difficult to contextualize.' caveats: [] - id: interspeech-2025-0347 published_date: "2025-08-17" entry_date: '2026-07-30' year: 2025 venue: Interspeech task: - codec - singing architecture: - GAN - VAE relevance: medium evidence_role: - architecture_variant - acceleration_evidence - control_evidence current_role: active_evidence method_family: - adversarial_waveform_vocoders - vae_codec_vocoder_decoders claims: - claim_id: injecting_explicit_periodic_signals_into_a_neural_codec_decoder_enables role: supports claim: Injecting explicit periodic signals into a neural codec decoder enables independent F0 control during waveform reconstruction, decoupling pitch from the discrete token stream. source: §4.2, §4.3, Table 1, Figure 2 evidence: Period variants achieve substantially lower F0-RMSE than Base (which embeds pitch implicitly in tokens) across all pitch shift conditions, and MOS improves from 2.37 to 3.28 at no shift. confidence: high relevance: low - claim_id: including_singing_voice_data_in_codec_training_improves_f0_accuracy role: supports claim: Including singing voice data in codec training improves F0 accuracy at high pitch ranges that speech-only corpora do not cover. source: §4.2, Figure 2, Figure 3 evidence: Period+GT reduces F0-RMSE relative to Period when log F0 is shifted upward by 6-12 semitones, corresponding to the extended high-pitch coverage of GTSinger vs. LibriTTS. confidence: high relevance: low - claim_id: gradient_reversal_based_pitch_disentanglement_in_neural_codecs_does_not role: complicates claim: Gradient reversal-based pitch disentanglement in neural codecs does not reliably improve perceptual quality and may degrade it. source: §4.3, Table 1 evidence: Period-GRL scores 2.44 MOS vs. Period at 3.28 MOS at no shift; Period-GRL+GT scores 3.33 vs. Period+GT at 3.55. The authors note the effect varies with training data domain, leaving the mechanism unclear. confidence: high relevance: medium - claim_id: codec_training_on_data_with_a_wider_pitch_range_can role: complicates claim: Codec training on data with a wider pitch range can improve quality at high pitches but reduces quality at lower pitches that are underrepresented in the new data. source: §4.3, Table 1 evidence: At -6 semitone shift, Period+GT (3.29 MOS) scores lower than Period (3.42 MOS), attributed to the model allocating capacity to the wider pitch range of GTSinger at the cost of fidelity in the lower range. confidence: high relevance: low limitations: - All subjective evaluation is conducted on a proprietary Japanese children's song dataset recorded by two singers unseen during training. Results may not generalise to other singing styles, languages, or recording conditions. - 'The GRL disentanglement module did not improve naturalness in subjective evaluation and in some conditions worsened it; the authors acknowledge this interaction with training data domain requires clarification. The evaluation is a codec reconstruction task only: no downstream discrete-token-based singing synthesis system is presented, so the utility of PeriodCodec in a full SVS pipeline remains undemonstrated. Model size is not reported.' caveats: [] - id: interspeech-2025-0406 published_date: "2025-08-17" entry_date: '2026-07-30' year: 2025 venue: Interspeech task: - TTS architecture: - GAN relevance: medium evidence_role: - architecture_variant current_role: active_evidence method_family: - adversarial_waveform_vocoders claims: - claim_id: pretrained_monaural_denoising_vocoders_can_refine_geometrically_warped_speech_into role: supports claim: Pretrained monaural denoising vocoders can refine geometrically warped speech into perceptually natural binaural audio without exposure to any binaural training data. source: §3.3, §4.4 evidence: Pretrained monaural denoising vocoders can refine geometrically warped speech into perceptually natural binaural audio without exposure to any binaural training data. confidence: high relevance: medium - claim_id: supervised_mono_to_binaural_synthesis_models_trained_on_small_room role: supports claim: Supervised mono-to-binaural synthesis models trained on small room-specific datasets degrade significantly under out-of-distribution acoustic conditions, including different rooms and languages. source: §4.5, Table 2 evidence: Supervised mono-to-binaural synthesis models trained on small room-specific datasets degrade significantly under out-of-distribution acoustic conditions, including different rooms and languages. confidence: high relevance: medium - claim_id: geometric_interaural_time_delay_warping_and_amplitude_scaling_based_on role: supports claim: Geometric interaural time delay warping and amplitude scaling based on the inverse-square law are both necessary components for zero-shot binaural synthesis; removing either one substantially degrades perceptual quality. source: §5, Table 3 evidence: Geometric interaural time delay warping and amplitude scaling based on the inverse-square law are both necessary components for zero-shot binaural synthesis; removing either one substantially degrades perceptual quality. confidence: high relevance: medium - claim_id: standard_in_distribution_benchmarks_for_binaural_synthesis_are_insufficient_to role: supports claim: Standard in-distribution benchmarks for binaural synthesis are insufficient to assess generalisation; evaluating on held-out room conditions reveals large performance gaps that in-distribution results conceal. source: §4.2, §4.5 evidence: Standard in-distribution benchmarks for binaural synthesis are insufficient to assess generalisation; evaluating on held-out room conditions reveals large performance gaps that in-distribution results conceal. confidence: high relevance: medium limitations: - The method does not model room impulse responses or listener head shape (HRTF is implicit and generic). Phase accuracy (P-ℓ2) remains comparable to DSP baselines but does not approach supervised methods. The vocoder processes each channel independently, which could introduce cross-channel inconsistencies. Quality plateaus at 3 WaveFit iterations and does not benefit from more. Future work could integrate personalized HRTFs or light room estimation modules to close the remaining gap on objective metrics. caveats: [] - id: interspeech-2025-0455 published_date: "2025-08-17" entry_date: '2026-07-30' year: 2025 venue: Interspeech task: - TTS architecture: - flow-matching - VAE - hybrid relevance: medium evidence_role: - architecture_variant - control_evidence current_role: active_evidence method_family: - flow_matching_waveform_generation - vae_codec_vocoder_decoders - hybrid_signal_neural_vocoders claims: - claim_id: adversarial_post_training_applied_to_a_pre_trained_flow_matching role: supports claim: Adversarial post-training applied to a pre-trained flow matching decoder can reduce the required ODE sampling steps by 5–8x while maintaining competitive naturalness and improving intelligibility relative to the full-step baseline. source: §3.3, Table 1, Table 2 evidence: Adversarial post-training applied to a pre-trained flow matching decoder can reduce the required ODE sampling steps by 5–8x while maintaining competitive naturalness and improving intelligibility relative to the full-step baseline. confidence: high relevance: low - claim_id: latent_flow_matching_tts_trained_on_approximately_1k_hours_can role: supports claim: Latent flow matching TTS trained on approximately 1K hours can match or surpass the intelligibility of zero-shot systems trained on 60–100K hours, though speaker similarity remains lower at this data scale. source: §4.2, Table 2 evidence: Latent flow matching TTS trained on approximately 1K hours can match or surpass the intelligibility of zero-shot systems trained on 60–100K hours, though speaker similarity remains lower at this data scale. confidence: high relevance: low - claim_id: framing_zero_shot_tts_as_a_speech_infilling_task_within role: supports claim: Framing zero-shot TTS as a speech infilling task within a latent space enables prompt-conditioned generation without requiring a separate speaker encoder or d-vector lookup. source: §3.2 evidence: Framing zero-shot TTS as a speech infilling task within a latent space enables prompt-conditioned generation without requiring a separate speaker encoder or d-vector lookup. confidence: high relevance: low - claim_id: hybrid_classifier_free_guidance_decoupling_the_unconditional_branch_to_a role: supports claim: Hybrid classifier-free guidance — decoupling the unconditional branch to a frozen base model checkpoint — reduces error accumulation in few-step generators compared to applying a fixed large guidance scale during adversarial post-training. source: §3.3 evidence: Hybrid classifier-free guidance — decoupling the unconditional branch to a frozen base model checkpoint — reduces error accumulation in few-step generators compared to applying a fixed large guidance scale during adversarial post-training. confidence: high relevance: medium - claim_id: adversarial_post_training_generalises_across_flow_matching_representation_domains_the role: supports claim: 'Adversarial post-training generalises across flow matching representation domains: the same post-training procedure improves both latent-domain (APTTS) and mel-spectrogram-domain (Matcha-TTS) models.' source: §4.3, Table 3 evidence: 'Adversarial post-training generalises across flow matching representation domains: the same post-training procedure improves both latent-domain (APTTS) and mel-spectrogram-domain (Matcha-TTS) models.' confidence: high relevance: low limitations: - SIM-o scores lag behind large-scale trained models, suggesting prompt fidelity would improve with more data. The hybrid CFG requires inference calls to both the post-trained and base FM models, adding complexity. The fixed ODE time steps (empirically chosen) may not be optimal across all input lengths. The system is English-only; multilingual extension is unexplored. caveats: [] - id: interspeech-2025-0854 published_date: "2025-08-17" entry_date: '2026-07-30' year: 2025 venue: Interspeech task: - TTS architecture: - GAN - flow-matching - transformer-enc-dec relevance: high evidence_role: - core_evidence - control_evidence - evaluation_caution current_role: active_evidence method_family: - adversarial_waveform_vocoders - flow_matching_waveform_generation - transformer_conditioned_vocoder_systems claims: - claim_id: generative_postprocessing_models_for_spectrogram_enhancement_lose_effectiveness_in_low role: supports claim: Generative postprocessing models for spectrogram enhancement lose effectiveness in low-resource settings because the acoustic model produces stronger artifacts that are underrepresented in the available training pairs. source: §1, §2.1 evidence: Generative postprocessing models for spectrogram enhancement lose effectiveness in low-resource settings because the acoustic model produces stronger artifacts that are underrepresented in the available training pairs. confidence: high relevance: medium - claim_id: simulating_low_resource_speaker_conditions_from_subsampled_high_resource_data role: supports claim: Simulating low-resource speaker conditions from subsampled high-resource data provides a viable source of artifact-realistic training pairs for spectrogram postprocessors, yielding consistent naturalness improvements across both GAN and flow-matching architectures. source: §2.2, §4, Table 2 evidence: Simulating low-resource speaker conditions from subsampled high-resource data provides a viable source of artifact-realistic training pairs for spectrogram postprocessors, yielding consistent naturalness improvements across both GAN and flow-matching architectures. confidence: high relevance: high - claim_id: flow_matching_postprocessors_achieve_higher_absolute_naturalness_than_gan_based role: supports claim: Flow-matching postprocessors achieve higher absolute naturalness than GAN-based postprocessors on low-resource speakers, but show smaller and statistically insignificant gains from the proposed data generation strategy compared to GAN-based postprocessors. source: §4, Table 2 evidence: Flow-matching postprocessors achieve higher absolute naturalness than GAN-based postprocessors on low-resource speakers, but show smaller and statistically insignificant gains from the proposed data generation strategy compared to GAN-based postprocessors. confidence: high relevance: high - claim_id: noise_augmentation_of_the_acoustic_model_does_not_improve_spectrogram role: supports claim: Noise augmentation of the acoustic model does not improve spectrogram naturalness when ground-truth prosody is used at evaluation, as its primary benefit is prosody prediction rather than spectrogram texture quality. source: §4, Table 2 evidence: Noise augmentation of the acoustic model does not improve spectrogram naturalness when ground-truth prosody is used at evaluation, as its primary benefit is prosody prediction rather than spectrogram texture quality. confidence: high relevance: low - claim_id: multi_stimulus_hidden_reference_tests_alone_may_be_insufficient_to role: supports claim: Multi-stimulus hidden-reference tests alone may be insufficient to detect when a TTS system exceeds vocoded ground-truth quality; ranking-by-elimination tests can reveal preferences for synthetic over reference conditions that MUSHRA-style scoring cannot capture. source: §3.3, §4 evidence: Multi-stimulus hidden-reference tests alone may be insufficient to detect when a TTS system exceeds vocoded ground-truth quality; ranking-by-elimination tests can reveal preferences for synthetic over reference conditions that MUSHRA-style scoring cannot capture. confidence: high relevance: low limitations: - The method requires training multiple acoustic models (4 in the experiments), increasing compute cost proportional to the number of high-resource speakers used. Evaluation uses ground-truth prosody, making it unclear how the approach fares in full end-to-end synthesis with predicted prosody. Only two low-resource speakers (one male, one female) are tested; generalization to different accent or language scenarios is not validated. The authors themselves acknowledge the noise-augmentation baseline performs worse here because it improves prosody — the interactions between prosody control and spectrogram naturalness are not fully resolved. caveats: [] - id: interspeech-2025-0998 published_date: "2025-08-17" entry_date: '2026-07-30' year: 2025 venue: Interspeech task: - VC architecture: - diffusion - GAN relevance: medium evidence_role: - architecture_variant - control_evidence current_role: active_evidence method_family: - diffusion_vocoder_and_restoration_systems - adversarial_waveform_vocoders claims: - claim_id: multi_stage_speech_restoration_pipelines_that_separate_noise_suppression_from role: supports claim: Multi-stage speech restoration pipelines that separate noise suppression from speaker-guided generation outperform single-stage generative approaches under severe degradation conditions. source: §4.3, Table 3 evidence: GSR+VC substantially outperforms standalone GSR or standalone VC across all metrics on both VCTK-DEMAND and UNIVERSE; the gain is largest on the UNIVERSE set, which simulates more severe distortions including band-limiting, reverberation, codec artefacts, and packet drops. confidence: high relevance: medium - claim_id: self_supervised_discrete_speech_representations_provide_more_robust_content_features role: supports claim: Self-supervised discrete speech representations provide more robust content features for voice conversion than raw mel-spectrograms when the input speech is degraded. source: §4.3, Table 3 evidence: VC (SSL) using HuBERT+VQ consistently outperforms VC (Mel) using direct mel-spectrogram input, with the gap widening on the more challenging UNIVERSE dataset where VC (Mel) shows significant quality degradation. confidence: high relevance: low - claim_id: diffusion_based_voice_conversion_models_cannot_reliably_handle_degraded_input role: complicates claim: Diffusion-based voice conversion models cannot reliably handle degraded input without a dedicated pre-processing stage, even when conditioned on clean speaker embeddings. source: §4.3, Table 3 evidence: VC (SSL) in standalone mode achieves lower scores than GSR+VC on both evaluation sets; the VC module performs markedly worse when the input is noisy without the GSR front-end, confirming that speaker-embedding guidance alone does not compensate for noisy content features. confidence: high relevance: low - claim_id: enrollment_dependent_speaker_guidance_for_speech_restoration_limits_applicability_to role: complicates claim: Enrollment-dependent speaker guidance for speech restoration limits applicability to settings where clean reference speech from the same speaker is available in advance. source: §3, §4.1 evidence: The system assumes short, uncorrelated segments of clean speech are obtained beforehand for speaker embedding extraction; the paper does not evaluate performance when such enrollment audio is unavailable or mismatched. confidence: high relevance: medium limitations: - The system requires a clean enrollment utterance from the target speaker, which may not be available in all real-world scenarios. No comparison with Miipher could be performed due to unavailability of that model, leaving the relationship between this approach and the closest prior work unquantified. Evaluation uses only automated perceptual quality metrics (NISQA, UTMOS, WV-MOS, DNSMOS) without any human listening tests, so it is unclear whether the metric gains translate to perceived quality improvements. Both the GSR and VC models are trained on separate datasets, and the system has not been evaluated on the joint training configuration proposed as future work. Performance on languages other than English is not examined. caveats: [] - id: interspeech-2025-1081 published_date: "2025-08-17" entry_date: '2026-07-30' year: 2025 venue: Interspeech task: - VC architecture: - GAN relevance: high evidence_role: - core_evidence - acceleration_evidence - control_evidence current_role: active_evidence method_family: - adversarial_waveform_vocoders claims: - claim_id: explicit_content_restoration_mechanisms_can_compensate_for_linguistic_information_degraded role: supports claim: Explicit content-restoration mechanisms can compensate for linguistic information degraded by instance normalization in speaker disentanglement, preserving intelligibility without SSL feature extractors. source: §2.1, Table 2 evidence: The SNCR dual restoration loss, which decomposes the IN residual into content-relevant and content-irrelevant parts via phoneme-entropy constraints, reduces WER from 5.57% (without SNCR) to 3.78% compared to a baseline trained with neither IN nor dual restoration loss. confidence: high relevance: medium - claim_id: gan_based_zero_shot_voice_conversion_systems_can_match_the role: supports claim: GAN-based zero-shot voice conversion systems can match the speaker similarity of diffusion-based systems at substantially lower parameter count and faster inference. source: §3.2, Table 1 evidence: The proposed 50M-parameter GAN system achieves SECS 0.25 versus DDDM-VC's SECS 0.27, while running at RTF 0.11 versus 0.29 and using 366M fewer parameters; naturalness (nMOS 3.86 vs. 3.52) also favors the GAN system. confidence: high relevance: high - claim_id: integrating_self_attention_layers_into_cnn_based_discriminators_improves_adversarial role: supports claim: Integrating self-attention layers into CNN-based discriminators improves adversarial training quality for mel-spectrogram reconstruction in voice conversion. source: §3.3, Table 2 evidence: Ablation Base4 (discriminator without self-attention) degrades nMOS and WER compared to the full system, confirming that attention-enhanced discrimination contributes to speech quality gains beyond the convolutional baseline. confidence: high relevance: high - claim_id: speaker_similarity_metrics_in_zero_shot_vc_evaluations_can_be role: complicates claim: Speaker similarity metrics in zero-shot VC evaluations can be inflated when the test set speakers overlap with a baseline system's training set, making fair cross-system comparison difficult. source: §3.2 evidence: VQMIVC achieves SECS 0.23 (close to the proposed 0.25) but was trained on VCTK, which overlaps with the VCTK test speakers; its WER of 21.4% reveals its conversions are not actually intelligible, suggesting its SECS reflects in-distribution behavior rather than genuine zero-shot capability. confidence: high relevance: medium limitations: - The system trains and evaluates on English only (LibriLight and VCTK). Whether the SNCR mechanism generalizes to cross-lingual voice conversion or languages with different phonological properties is untested. The ablation evaluates components one at a time, leaving interactions between SNCR and consistency loss unexplored. The mel-spectrogram reconstruction approach limits the system to fixed-rate synthesis, and no streaming or low-latency deployment scenario is discussed. Inference speed is reported on GPU only; embedded or CPU deployment characteristics are not assessed. caveats: [] - id: interspeech-2025-1122 published_date: "2025-08-17" entry_date: '2026-07-30' year: 2025 venue: Interspeech task: - TTS architecture: - GAN - transformer-enc-dec relevance: medium evidence_role: - architecture_variant current_role: active_evidence method_family: - adversarial_waveform_vocoders - transformer_conditioned_vocoder_systems claims: - claim_id: quantisation_aware_training_is_more_effective_than_parameter_reduction_via role: supports claim: Quantisation-aware training is more effective than parameter reduction via channel shrinkage for achieving compact TTS models at equivalent model sizes. source: §3.3.1, Table 1 evidence: Quantisation-aware training is more effective than parameter reduction via channel shrinkage for achieving compact TTS models at equivalent model sizes. confidence: high relevance: low - claim_id: the_vocoder_stage_of_a_tts_pipeline_is_more_sensitive role: supports claim: The vocoder stage of a TTS pipeline is more sensitive to aggressive weight quantisation than the acoustic model, making it a disproportionate source of quality degradation. source: §3.3.1 evidence: The vocoder stage of a TTS pipeline is more sensitive to aggressive weight quantisation than the acoustic model, making it a disproportionate source of quality degradation. confidence: high relevance: high - claim_id: extremely_low_bit_1_2_bit_weight_quantisation_can_preserve role: supports claim: Extremely low-bit (1–2 bit) weight quantisation can preserve intelligible and reasonably natural speech synthesis when applied selectively to convolutional layers. source: §3.3.1, Table 1 evidence: Extremely low-bit (1–2 bit) weight quantisation can preserve intelligible and reasonably natural speech synthesis when applied selectively to convolutional layers. confidence: high relevance: medium - claim_id: hardware_aware_packing_schemes_are_necessary_to_realise_the_theoretical role: supports claim: Hardware-aware packing schemes are necessary to realise the theoretical storage savings of sub-byte quantisation on devices that operate in 8-bit units. source: §2.3 evidence: Hardware-aware packing schemes are necessary to realise the theoretical storage savings of sub-byte quantisation on devices that operate in 8-bit units. confidence: high relevance: medium limitations: - All experiments use a single TTS architecture (JETS with convolutional blocks) trained on a single corpus (LibriTTS-R). Whether these findings generalise to transformer-heavy or autoregressive architectures, or to lower-resource languages, is not addressed. - The subjective evaluation uses only 15 raters and 30 utterances, which is at the lower end of statistical power for MOS comparisons. The index distribution is heavily non-uniform (indices 0, 121, 242 — representing all-zero, all-one, all-minus-one blocks — are overrepresented), suggesting that further compression via entropy coding (e.g. Huffman) is feasible but unexplored. Inference speed does not improve relative to the full 32-bit base model because weight reconstruction from indices adds overhead; achieving both compact storage and fast inference simultaneously requires dedicated hardware or kernel-level support that is not demonstrated. caveats: [] - id: interspeech-2025-1747 published_date: "2025-08-17" entry_date: '2026-07-30' year: 2025 venue: Interspeech task: - VC architecture: - diffusion - GAN relevance: medium evidence_role: - architecture_variant - control_evidence current_role: active_evidence method_family: - diffusion_vocoder_and_restoration_systems - adversarial_waveform_vocoders claims: - claim_id: performing_knowledge_distillation_in_the_conversion_process_rather_than_the role: supports claim: Performing knowledge distillation in the conversion process rather than the reconstruction process is necessary to avoid identity-mapping collapse when the content encoder is simultaneously trainable. source: §4.2, Table 1, Table 2 evidence: Baseline reconstruction-distillation with trainable CNN (FastVoiceGrad + pϕ) achieves UTMOS 3.45 and SECS 0.718; switching to conversion-based ADCD raises these to 4.03 and 0.853, respectively. Direct distillation (an alternative that avoids conversion-process training) degrades to UTMOS 3.35 with a 1-layer encoder, confirming the mechanism failure. confidence: high relevance: medium - claim_id: simultaneously_distilling_both_the_diffusion_backbone_and_the_content_encoder role: supports claim: Simultaneously distilling both the diffusion backbone and the content encoder of a one-step diffusion VC model can reduce inference cost substantially with little or no quality degradation relative to distilling only the diffusion module. source: §4.3, Table 3, Table 5 evidence: FasterVoiceGrad (joint distillation) achieves UTMOS 4.03 and SECS 0.853 vs. FastVoiceGrad's 3.96 and 0.847, while delivering 6.6x GPU RTF and 1.8x CPU RTF improvement across both VCTK and LibriTTS evaluations. confidence: high relevance: low - claim_id: neural_speaker_encoder_based_similarity_metrics_may_not_reflect_perceptual role: complicates claim: Neural speaker-encoder-based similarity metrics may not reflect perceptual speaker identity as judged by human listeners, particularly when source speaker residuals remain after conversion. source: §4.3, Table 3, Table 4 evidence: FasterVoiceGrad scores higher SECS (0.853 vs. 0.847) than FastVoiceGrad but lower sMOS (2.66 vs. 2.84); the authors attribute the discrepancy to the CNN content encoder not explicitly removing source speaker characteristics, which humans but not speaker encoders detect. confidence: high relevance: medium - claim_id: contrastive_speaker_emphasis_via_inverse_score_distillation_mitigates_the_trade role: refines claim: Contrastive speaker emphasis via inverse score distillation mitigates the trade-off between content preservation and speaker conversion when joint distillation is used. source: §4.2, Table 1 evidence: Adding inverse score distillation on top of reconversion-based distillation improves SECS from 0.844 to 0.853 without degrading CER (1.2) or UTMOS (4.03), recovering the speaker similarity lost by reconversion's content preservation bias. confidence: high relevance: medium limitations: - The CNN content encoder does not explicitly separate source speaker identity from linguistic content. Subjective sMOS scores (2.66) are measurably lower than FastVoiceGrad's (2.84), suggesting residual source voice characteristics persist and limit perceptual speaker conversion quality despite competitive automatic metrics. - Subjective evaluation involved only 11 participants and 90 speaker/sentence pairs, which constrains the statistical power of the sMOS comparisons. The GPU speedup (6.6x) is larger than the CPU speedup (1.8x), reflecting architectural factors that may affect deployment on edge hardware differently. The study does not evaluate against modern flow-matching or large-scale VC baselines, leaving the system's position in the broader landscape uncharacterised. Future directions suggested include accent conversion and real-time VC applications. caveats: [] - id: interspeech-2025-1763 published_date: "2025-08-17" entry_date: '2026-07-30' year: 2025 venue: Interspeech task: - VC architecture: - GAN - diffusion relevance: high evidence_role: - core_evidence current_role: active_evidence method_family: - adversarial_waveform_vocoders - diffusion_vocoder_and_restoration_systems claims: - claim_id: intermediate_vocoder_features_after_partial_upsampling_are_sufficient_for_waveform role: supports claim: Intermediate vocoder features after partial upsampling are sufficient for waveform-quality adversarial discrimination of acoustic feature generators. source: §3.2, Tables 1 and 5, Figure 3 evidence: FVG + VPFD₁ (single upsampling step, 8x) matches full waveform discriminator performance on UTMOS, DNSMOS, CER, and SECS across both VCTK and LibriTTS; visualisation shows periodic waveform structures emerge after one upsampling step. confidence: high relevance: high - claim_id: freezing_pretrained_feature_extractors_in_projected_gan_discriminators_is_essential role: supports claim: Freezing pretrained feature extractors in projected GAN discriminators is essential for acoustic synthesis quality. source: §3.2, Table 2 evidence: 'Ablation shows that both pretraining and freezing V_feat are independently necessary: omitting either degrades UTMOS, DNSMOS, and SECS, with no change in training time or memory.' confidence: high relevance: high - claim_id: waveform_domain_discriminators_in_two_stage_tts_vc_training_are role: complicates claim: Waveform-domain discriminators in two-stage TTS/VC training are effective but impose resource costs that make them impractical outside well-resourced settings. source: §1, §3.3, Tables 1 and 3 evidence: VWD (MPD + MRD applied to vocoder output) requires 47 hours and 66.3 GB GPU memory on VCTK; mel-spectrogram discriminators avoid this cost but fail to match waveform-domain quality on DNSMOS regardless of model size. confidence: high relevance: low - claim_id: for_vocoder_based_feature_projection_a_minimum_of_one_upsampling role: refines claim: For vocoder-based feature projection, a minimum of one upsampling step is necessary to produce the periodic structures required for effective adversarial discrimination. source: §3.2, Table 1, Figure 3 evidence: VPFD₀ (no upsampling) significantly degrades DNSMOS (3.66 vs. 3.79) and SECS (0.843 vs. 0.851) relative to VPFD₁; Figure 3 shows periodic structures absent in mel spectrograms and zero-upsampling features appear only after one step. confidence: high relevance: high limitations: - VPFD is validated only on one VC system (FastVoiceGrad on VoiceGrad) and never on a TTS system, despite TTS being a stated target application. The subjective evaluation involves only 11 participants, which is small for a listening test. All experiments use 22.05 kHz audio with 80-dimensional log-mel spectrograms and HiFi-GAN V1 as the vocoder; performance with other vocoders, sample rates, or codec-based acoustic representations is unknown. The approach requires a high-quality pretrained vocoder whose intermediate features provide sufficient discriminative power, which may not hold for vocoders trained on limited or mismatched data. caveats: [] - id: interspeech-2025-1819 published_date: "2025-08-17" entry_date: '2026-07-30' year: 2025 venue: Interspeech task: - TTS architecture: - GAN relevance: high evidence_role: - core_evidence - acceleration_evidence - control_evidence current_role: active_evidence method_family: - adversarial_waveform_vocoders claims: - claim_id: input_side_multi_stream_decomposition_via_analysis_filters_can_improve role: supports claim: Input-side multi-stream decomposition via analysis filters can improve both quality and efficiency in harmonic-prior 2D convolutional vocoders. source: §2.2, §4.2 evidence: MS-Wavehax splits the input waveform into four subscale signals before STFT, producing sub-spectrograms with locally consistent spatial structure; it achieves the highest MOS in both analysis-synthesis and TTS tasks while requiring roughly half the MACs of standard Wavehax and only 0.332M parameters. confidence: high relevance: medium - claim_id: batch_inference_real_time_factor_is_a_poor_proxy_for role: supports claim: Batch-inference real-time factor is a poor proxy for streaming throughput; architecture choices interact with chunk size and cache behaviour in ways that can reverse the efficiency ranking. source: §3.2, Table 2, Figure 2 evidence: At small chunk sizes MS-Wavehax achieves the lowest RTF among all vocoders; at larger chunk sizes Vocos overtakes it because 1D filter matrices are constant-sized and cache-resident, while Wavehax's 2D im2col matrices grow with chunk size T, causing more frequent main-memory access. confidence: high relevance: low - claim_id: multi_stream_decomposition_does_not_uniformly_improve_quality_across_all role: complicates claim: Multi-stream decomposition does not uniformly improve quality across all frame-level vocoder architectures. source: §4.2 evidence: MS-Vocos underperforms standard Vocos in MOS evaluation; the authors attribute this to a latent representation narrower than the output complex spectrum, which makes estimating inter-stream dependencies particularly difficult. confidence: high relevance: high - claim_id: replacing_layer_normalization_with_batch_normalization_in_gan_vocoders_enables role: supports claim: Replacing layer normalization with batch normalization in GAN vocoders enables streaming compatibility and can also yield objective quality improvements. source: §2.3 evidence: Layer normalization computes statistics over the full sequence and is incompatible with incremental streaming; switching to batch normalization removed this constraint and improved PESQ and UTMOS scores in internal experiments on all tested vocoder configurations. confidence: high relevance: high - claim_id: causal_vocoder_processing_with_minimal_lookahead_can_match_non_causal role: supports claim: Causal vocoder processing with minimal lookahead can match non-causal quality in single-speaker TTS settings. source: §4.2, Figure 4 evidence: Causal MS-Wavehax with a one-frame lookahead achieves nearly identical MOS to non-causal MS-Wavehax on the JSUT TTS task; the gap observed in the multi-speaker A/S task disappears, suggesting that causal constraints function as an effective regularizer in single-speaker conditions. confidence: high relevance: high limitations: - Evaluations are conducted exclusively on Japanese speech corpora (JVS and JSUT) on a single CPU architecture (AMD EPYC), so generalisation to multilingual or GPU-accelerated deployment is not demonstrated. All models were conditioned on mel-spectrograms extracted from clean studio recordings; behaviour on noisy or spontaneous speech is not assessed. The paper does not include a real-time demo or end-to-end latency measurement with an acoustic model frontend, so the practical time-to-first-audio-byte in a complete TTS pipeline remains uncharacterised. caveats: [] - id: interspeech-2025-2031 published_date: "2025-08-17" entry_date: '2026-07-30' year: 2025 venue: Interspeech task: - TTS architecture: - transformer-enc-dec relevance: medium evidence_role: - architecture_variant - control_evidence current_role: active_evidence method_family: - transformer_conditioned_vocoder_systems claims: - claim_id: for_zero_shot_tts_across_language_families_assigning_g2p_rules role: supports claim: For zero-shot TTS across language families, assigning G2P rules by phonotactic family rather than script family produces substantially more natural and intelligible synthesis. source: §5.3, Tables 1, 2 evidence: Sanskrit synthesis with Dravidian rules (Kannada system, MOS 4.12, WER 4.6%) substantially outperforms synthesis with Indo-Aryan rules (Hindi system, MOS 3.25, WER 28.0%), because schwa deletion applied via IA rules violates Sanskrit's phonotactics despite its Devanagari script. confidence: high relevance: low - claim_id: a_shared_phoneme_inventory_extended_via_phonetic_substitution_can_enable role: supports claim: A shared phoneme inventory extended via phonetic substitution can enable zero-shot synthesis for target languages with sounds absent from all source training sets. source: §4.2 evidence: The CLS was extended to represent nukta-diacritic consonants and borrowed aspiration in Kurukh by mapping them to nearest-neighbour base characters, enabling intelligible synthesis of Kurukh without any Kurukh training data. confidence: high relevance: medium - claim_id: dialect_specific_zero_shot_tts_requires_matching_the_source_synthesiser role: supports claim: Dialect-specific zero-shot TTS requires matching the source synthesiser to the dialect's phonotactic affiliation, not its surface script or nominal language grouping. source: §5.3, Table 1 evidence: 'Two Konkani dialects yield opposite synthesiser preferences: Maharashtrian Konkani (IA phonotactics) is best with Marathi at MOS 3.64, while Canara Konkani (DR phonotactics) is best with Kannada at MOS 3.34, despite both dialects being classified under the same language label.' confidence: high relevance: low - claim_id: cross_family_zero_shot_synthesis_degrades_for_target_languages_with role: complicates claim: Cross-family zero-shot synthesis degrades for target languages with phonological features structurally absent from all available source synthesisers. source: §5.3 evidence: Kurukh synthesis struggles specifically with glottal stops across both Hindi and Kannada systems, and question intonation contours were misrendered as assertions in some utterances, because neither source language encodes these features. confidence: high relevance: medium - claim_id: linguistic_proximity_for_zero_shot_tts_transfer_is_captured_more role: refines claim: Linguistic proximity for zero-shot TTS transfer is captured more precisely by phonotactic family than by script family or geographic proximity. source: §5.3, Tables 1, 3 evidence: 'Sanskrit belongs to the IA family but achieves lower MOS with IA synthesisers (3.25) than with DR synthesisers (4.12), and MCD scores independently confirm the same pattern (IA: 8.48 vs. DR: 7.15/6.94). Phonotactic family membership predicts transfer quality more reliably than script or family taxonomy.' confidence: high relevance: low limitations: - The evaluation relies exclusively on four monolingual source synthesisers (Hindi, Kannada, Marathi, Telugu), and the method is not tested against any large multilingual TTS baseline that might implicitly have learned cross-family phonotactics from scale. For Kurukh, the ground truth audio is Bible recordings with faint background music (cleaned with VoiceFixer), which may depress the GT MOS and make the synthesised system comparisons harder to interpret. The approach requires linguistic expertise to correctly classify the phonotactic family of a new target language, which limits its applicability as an automated pipeline for new low-resource languages. - The anomalous MCD result for Maharashtrian Konkani (higher distortion with Marathi despite better naturalness MOS) is unresolved and may indicate that MCD is less reliable for measuring synthesis quality under phonotactic family mismatch conditions. caveats: [] - id: interspeech-2025-2151 published_date: "2025-08-17" entry_date: '2026-07-30' year: 2025 venue: Interspeech task: - VC - evaluation architecture: - GAN relevance: high evidence_role: - core_evidence - control_evidence - evaluation_caution - infrastructure current_role: active_evidence method_family: - adversarial_waveform_vocoders claims: - claim_id: objective_voice_conversion_evaluation_metrics_that_require_ground_truth_utterances role: supports claim: Objective voice conversion evaluation metrics that require ground-truth utterances (MCD, F0-RMSE, FFE, GPE) are only feasible when a parallel corpus exists, making dataset parallelism a prerequisite for rigorous VC benchmarking. source: §1.3 evidence: The authors identify the absence of parallel Farsi data as the direct reason these objective metrics cannot be applied to existing Farsi corpora; FaVC is designed specifically to make them feasible. confidence: high relevance: low - claim_id: gan_based_voice_conversion_systems_trained_on_parallel_corpora_in role: supports claim: GAN-based voice conversion systems trained on parallel corpora in low-resource languages can reach naturalness and speaker similarity scores comparable to training on established English benchmarks. source: §4.3, Table 2 evidence: CycleGAN-VC, StarGAN-VC, StarGAN-VC2, and StarGANv2-VC trained on FaVC achieve MOS-N scores of 3.3–3.94 and MOS-P of 2.5–4.1, within the range of the same models trained on VCC2018 under identical configuration. confidence: high relevance: high - claim_id: non_english_voice_conversion_research_is_limited_by_the_near role: complicates claim: Non-English voice conversion research is limited by the near-absence of parallel, phonemically balanced corpora outside English. source: §1.2, §2.2 evidence: Despite Farsi being spoken by over 100 million people, no publicly available parallel Farsi VC dataset existed prior to FaVC; existing Farsi datasets (DeepMine, Farsi ESD, FarsDAT) are non-parallel and target ASR, TTS, or emotion recognition. confidence: high relevance: low - claim_id: incorporating_film_and_adain_conditioning_into_a_tts_decoder_supports role: supports claim: Incorporating FiLM and AdaIN conditioning into a TTS decoder supports effective one-shot voice conversion with results competitive with many-to-many GAN baselines. source: §4.1, §4.3, Table 2 evidence: Modified StyleTTS-VC with FiLM+AdaIN achieves MOS-N 3.94 and MOS-P 3.87 on FaVC (MCD 5.58), the highest naturalness score among the five evaluated models and competitive with StarGANv2-VC on speaker similarity (MOS-P 4.1). confidence: high relevance: high limitations: - The dataset is small (11 speakers, 405 utterances each), which limits the training capacity for data-hungry approaches. Evaluations across FaVC and English datasets conflate language, speaker population, and recording condition differences; the cross-language parity finding should be treated as indicative rather than controlled. No statistical significance testing is reported for MOS comparisons. Cross-lingual VC scenarios (Farsi to English or vice versa) are not evaluated, despite being flagged as a future direction. The dataset is currently distributed upon request rather than through an open public repository, which constrains reproducibility. caveats: [] - id: interspeech-2025-2573 published_date: "2025-08-17" entry_date: '2026-07-30' year: 2025 venue: Interspeech task: - TTS - evaluation architecture: - flow-matching - GAN relevance: medium evidence_role: - architecture_variant - control_evidence - infrastructure current_role: active_evidence method_family: - flow_matching_waveform_generation - adversarial_waveform_vocoders claims: - claim_id: accurate_grapheme_to_phoneme_conversion_is_a_critical_prerequisite_for role: supports claim: Accurate grapheme-to-phoneme conversion is a critical prerequisite for TTS quality in morphologically complex languages, with phonemizer errors directly degrading perceived naturalness and pronunciation. source: §4.2, Table 2 evidence: The modified espeak-ng phonemizer reduced mispronounced words by 38% (157 to 97) and improved naturalness MOS by 0.17 points for Egyptian dialect; objective MCD scores were unchanged, confirming the perceptual significance of phonemizer accuracy beyond spectrogram-level metrics. confidence: high relevance: low - claim_id: diacritization_vowelization_of_arabic_text_is_essential_for_intelligible_tts role: supports claim: Diacritization (vowelization) of Arabic text is essential for intelligible TTS; training without vowels causes consonant clustering and robotic-sounding output. source: §4.3, Table 3 evidence: MSA naturalness MOS dropped from 3.21 (vowelized) to 2.67 (unvowelized) and pronunciation from 3.95 to 2.53; listening revealed that missing vowels caused consonant clustering that reduced intelligibility and produced unnatural speech. confidence: high relevance: low - claim_id: code_switching_tts_is_more_difficult_than_monolingual_tts_even role: complicates claim: Code-switching TTS is more difficult than monolingual TTS even when both use the same acoustic model and vocoder trained on equivalent data quantities. source: §4.2, Table 2 evidence: CS pronunciation MOS (4.23 with modified phonemizer) is lower than Egyptian (4.57) and MSA (4.27) despite identical training duration and model architecture, with the phrase-level language-switching strategy introducing boundary artefacts not present in monolingual conditions. confidence: high relevance: high - claim_id: recording_the_same_speaker_across_multiple_dialects_and_an_l2 role: supports claim: Recording the same speaker across multiple dialects and an L2 language enables controlled disentanglement of speaker identity from dialect-specific acoustic characteristics in TTS datasets. source: §2.2 evidence: SawtArabi records a native Egyptian speaker in EGY, MSA, English, and CS conditions, allowing comparison of L2-speaking characteristics against native dialect production within a constant speaker identity. confidence: high relevance: low limitations: - The corpus is single-speaker (one male voice talent), meaning baseline systems cannot generalise to multi-speaker Arabic TTS and MOS comparisons are specific to this speaker's voice characteristics. Reproducibility of baseline MOS scores by external labs depends on access to similar listener pools proficient in Egyptian Arabic. - The corpus size (4 hours total, 1 hour per condition) is small relative to modern TTS training norms. Matcha-TTS baselines trained on 45 minutes of speech per dialect are unlikely to achieve the quality attainable with larger corpora, which limits the benchmark's utility for measuring state-of-the-art systems. The paper does not include multi-speaker baselines or zero-shot TTS evaluation, leaving open how well the corpus scales to speaker-adaptive or large-model settings. Future work plans to expand to additional Arabic dialects but does not address multi-speaker coverage. caveats: [] - id: interspeech-2025-2586 published_date: "2025-08-17" entry_date: '2026-07-30' year: 2025 venue: Interspeech task: - TTS architecture: - transformer-enc-dec - GAN relevance: medium evidence_role: - architecture_variant - control_evidence current_role: active_evidence method_family: - transformer_conditioned_vocoder_systems - adversarial_waveform_vocoders claims: - claim_id: restricting_style_quantization_to_acoustically_informative_voiced_speech_regions_improves role: supports claim: Restricting style quantization to acoustically informative (voiced) speech regions improves both style expressiveness and prosodic accuracy in reference-based TTS. source: §4.4.1, Table 4.3 evidence: On ESD, removing voiced extraction raises RMSE_F0 from 8.27 to 11.48 Hz and WER from 12.64 to 14.06; the full model achieves the best style similarity MOS (3.84) over all baselines. confidence: high relevance: low - claim_id: pairing_style_disentanglement_with_a_complementary_prosody_preserving_loss_stabilises role: supports claim: Pairing style disentanglement with a complementary prosody-preserving loss stabilises training and prevents prosody degradation from aggressive content removal. source: §4.4.2, Table 4.3 evidence: Removing only the SP loss raises pitch error to 9.74 Hz; removing both SD and SP degrades nMOS to 3.66 and WER to 15.38, with the SP-only-removed condition showing worse prosody than the case with no disentanglement losses at all. confidence: high relevance: medium - claim_id: asymmetric_biased_self_attention_in_unvoiced_region_filling_allowing_information role: supports claim: Asymmetric (biased) self-attention in unvoiced region filling, allowing information flow from voiced to unvoiced positions but not the reverse, outperforms symmetric or fully blocked alternatives for prosodic continuity. source: §4.4.3, Table 4.4 evidence: Standard self-attention in the UF module degrades pitch error to 16.38 Hz and F1 v/uv to 0.6668; binary masking partially recovers at 13.19 Hz versus 8.27 Hz for biased attention. confidence: high relevance: medium - claim_id: style_transfer_quality_gains_from_region_specific_quantization_have_been role: complicates claim: Style transfer quality gains from region-specific quantization have been established only within a controlled emotional corpus, leaving open the question of whether the approach generalises to broader speaking styles or modern generative architectures. source: §4.1, Table 4.1 evidence: All baselines are FastSpeech 2 variants evaluated solely on ESD (10 speakers, 5 discrete emotions); no comparison with flow-matching, diffusion, or large-scale systems is provided. confidence: high relevance: medium limitations: - All evaluations are conducted on a single English emotional speech corpus (ESD, 10 speakers, 5 emotions). Baselines are restricted to FastSpeech 2-based systems; no comparison with flow-matching or diffusion TTS is included, and there is no evidence of generalisation to multilingual or out-of-domain speech. - Model size is not reported. The voiced/unvoiced segmentation depends on pre-extracted V/UV flags, introducing a dependency on external pitch tracking that may degrade under noisy or spontaneous speech conditions. Non-parallel style transfer performance, while positive, consistently trails the parallel setting, indicating residual sensitivity to content mismatch even after disentanglement. The rotation trick hyperparameters and UF module depth are fixed without sensitivity analysis. caveats: [] - id: interspeech-2025-2726 published_date: "2025-08-17" entry_date: '2026-07-30' year: 2025 venue: Interspeech task: - codec architecture: - GAN - hybrid relevance: medium evidence_role: - architecture_variant current_role: active_evidence method_family: - adversarial_waveform_vocoders - hybrid_signal_neural_vocoders claims: - claim_id: staged_codec_training_that_separates_quantizer_optimization_from_decoder_optimization role: supports claim: Staged codec training that separates quantizer optimization from decoder optimization can improve single-codebook reconstruction quality beyond joint training. source: §3.6, Table 3 evidence: 'DS-Codec''s two-stage framework (mirror Stage 1 to train the quantizer, non-mirror Stage 2 to specialize the decoder) outperforms APCodec+''s single-stage joint training at both stages: Stage 1 UTMOS 4.123 vs. 4.113, PESQ 2.768 vs. 2.632; final model UTMOS 4.214 vs. 4.186 on LibriSpeech.' confidence: high relevance: low - claim_id: a_mirrored_encoder_decoder_constraint_during_quantizer_training_reduces_the role: supports claim: A mirrored encoder-decoder constraint during quantizer training reduces the input-output MSE of the quantization module, producing more robust codebooks. source: §3.5, Figure 2 evidence: VQ loss curves during Stage 1 show the mirrored structure achieves lower quantization MSE than the non-mirrored structure across training epochs, even though the non-mirrored structure has lower VQ loss; the paper interprets smaller MSE as higher codebook fidelity. confidence: high relevance: medium - claim_id: product_quantization_over_multiple_small_sub_codebooks_enables_large_effective role: supports claim: Product quantization over multiple small sub-codebooks enables large effective codebook sizes while preserving the single-token-per-frame interface required by LLM-based TTS. source: §2.2.2, Table 1 evidence: DS-Codec-PQ combines four 16-code VQ modules to produce a 65,536-code effective codebook, indexed as a single integer, achieving UTMOS 4.214 and PESQ 2.882 at 1.28kbps with 80 tokens/second. confidence: high relevance: low - claim_id: a_stronger_decoder_does_not_straightforwardly_compensate_for_a_weaker role: complicates claim: A stronger decoder does not straightforwardly compensate for a weaker quantizer when both are trained jointly. source: §3.6, Table 3 evidence: APCodec+'s joint training with a non-mirrored (stronger) decoder in Stage 1 yields lower Stage 1 PESQ (2.632) than DS-Codec's mirror Stage 1 with a weaker mirrored decoder (PESQ 2.768), suggesting quantizer quality dominates reconstruction fidelity early in training. confidence: high relevance: medium limitations: - Evaluation is limited to English read speech (LibriSpeech) and one supplementary in-domain set (LJSpeech). Performance on noisy, spontaneous, or multilingual speech is untested. Model size is not reported, making it impossible to assess parameter efficiency relative to BigCodec (159M) or DAC (74M). The UTMOS and PESQ metrics used are objective proxies for perceptual quality; no formal subjective listening study is reported. - The Stage 2 improvement from retaining decoder weights vs. reinitializing them is stated but not ablated directly — the comparison to APCodec+ involves multiple differences (stage design, weight retention, architecture), so the isolated effect of weight retention is unclear. caveats: [] - id: interspeech-2025-2739 published_date: "2025-08-17" entry_date: '2026-07-30' year: 2025 venue: Interspeech task: - TTS architecture: - GAN relevance: high evidence_role: - core_evidence current_role: active_evidence method_family: - adversarial_waveform_vocoders claims: - claim_id: frequency_domain_learnable_filters_can_simultaneously_suppress_aliasing_and_blurring role: supports claim: Frequency-domain learnable filters can simultaneously suppress aliasing and blurring artifacts in GAN vocoders, replacing specialised anti-aliasing upsampling modules. source: §2.2, §3.3, Table 1 evidence: GAFilter, a temporally invariant STFT-domain multiplicative filter, improves all objective metrics over BigVGAN on LibriTTS-dev while allowing removal of the anti-aliasing upsampling-downsampling operations. confidence: high relevance: high - claim_id: frequency_domain_artifact_filtering_in_gan_vocoders_generalises_to_out role: supports claim: Frequency-domain artifact filtering in GAN vocoders generalises to out-of-distribution speakers and acoustic domains without domain-specific adaptation. source: §3.3, Table 3 evidence: AF-Vocoder, trained only on LibriTTS, outperforms BigVGAN in PESQ and periodicity error on VCTK, LJSpeech, and MUSDB18-HD. confidence: high relevance: high - claim_id: slicing_adversarial_network_training_frameworks_can_increase_perceptual_artifact_severity role: complicates claim: Slicing adversarial network training frameworks can increase perceptual artifact severity in GAN vocoders despite achieving stronger objective periodicity scores. source: §3.3, Table 1 evidence: Under BigVSAN's SAN framework, the authors observe more audible artifacts; AF-Vocoder-SAN achieves UTMOS 3.517, below the 3.650 of AF-Vocoder V1 trained with standard least-squares GAN loss. confidence: high relevance: high - claim_id: frequency_domain_inductive_biases_improve_gradient_stability_during_gan_vocoder role: supports claim: Frequency-domain inductive biases improve gradient stability during GAN vocoder training, reducing early-stage gradient variance. source: §3.3, Figure 3 evidence: Gradient curves show lower deviation and faster stabilisation for AF-Vocoder versus BigVGAN in early training steps, attributed to GAFilter filtering out anomalous spectral components. confidence: high relevance: high limitations: - The subjective evaluation uses only 8 raters and a single MUSDB18-HD test set, which limits the strength of the perceptual claims. Whether the perceptual gains hold across diverse speakers, languages, or acoustic conditions at larger rater scale is not established. The paper does not report inference latency or real-time factor, leaving practical throughput trade-offs relative to BigVGAN unquantified. GAFilter is evaluated only on mel-spectrogram vocoders; its interaction with neural codec decoders or codec-native TTS pipelines is not explored. The SAN framework's tendency to increase audible artifacts under GAFilter integration is noted but not explained mechanistically. caveats: [] - id: '2508.18006' published_date: "2025-08-25" entry_date: '2026-07-30' year: 2025 venue: arXiv task: - TTS architecture: - GAN relevance: medium evidence_role: - architecture_variant - control_evidence - evaluation_caution current_role: active_evidence method_family: - adversarial_waveform_vocoders claims: - claim_id: adapter_based_fine_tuning_prevents_catastrophic_forgetting_in_tts_models role: supports claim: Adapter-based fine-tuning prevents catastrophic forgetting in TTS models while achieving better naturalness and speaker similarity than full fine-tuning. source: §4.3, Table 1; §4.4, Table 2 evidence: Adapter-based fine-tuning prevents catastrophic forgetting in TTS models while achieving better naturalness and speaker similarity than full fine-tuning. confidence: high relevance: low - claim_id: in_gan_based_tts_language_agnostic_speaker_representations_are_encoded role: supports claim: In GAN-based TTS, language-agnostic speaker representations are encoded primarily in the vocoder rather than the acoustic model, as evidenced by the outsized impact of vocoder-only adapters on accent nativeness. source: §4.4, Table 2 evidence: In GAN-based TTS, language-agnostic speaker representations are encoded primarily in the vocoder rather than the acoustic model, as evidenced by the outsized impact of vocoder-only adapters on accent nativeness. confidence: high relevance: high - claim_id: the_number_and_diversity_of_speakers_in_target_language_pre role: supports claim: The number and diversity of speakers in target-language pre-training data is a critical prerequisite for high-quality cross-lingual speaker adaptation; single-speaker pre-training leads to severe degradation. source: §4.4, Table 2 evidence: The number and diversity of speakers in target-language pre-training data is a critical prerequisite for high-quality cross-lingual speaker adaptation; single-speaker pre-training leads to severe degradation. confidence: high relevance: medium - claim_id: phoneme_substitution_rate_derived_from_a_mispronunciation_detection_model_provides role: supports claim: Phoneme substitution rate, derived from a mispronunciation detection model, provides a reliable and reproducible objective proxy for accent nativeness in cross-lingual TTS, correlating well with human MUSHRA accent judgements. source: §3.3 evidence: Phoneme substitution rate, derived from a mispronunciation detection model, provides a reliable and reproducible objective proxy for accent nativeness in cross-lingual TTS, correlating well with human MUSHRA accent judgements. confidence: high relevance: low limitations: - All evaluations are purely objective — no subjective listening test (MOS, MUSHRA) is reported for the proposed systems. The NORESQA-MOS (TorchAudio-SQUIM) values used as "MOS" are automatic predictions, not human ratings; the paper's decision to discard subjective evaluation limits the strength of quality claims. - The study is restricted to two languages (British English and Castilian Spanish) and a single source speaker per adaptation direction. It is unclear whether the findings generalise to phonologically more distant language pairs, accents, or speakers with more unusual voice characteristics. The backbone model is proprietary and not publicly released, limiting reproducibility. Adapter hyperparameters (bottleneck dimension 16, kernel sizes [3, 5, 3]) were chosen via early experiments but no systematic search is reported. caveats: [] - id: '2507.14534' published_date: "2025-08-30" entry_date: '2026-07-30' year: 2025 venue: arXiv task: - VC architecture: - GAN - hybrid relevance: medium evidence_role: - architecture_variant - acceleration_evidence - control_evidence - evaluation_caution current_role: active_evidence method_family: - adversarial_waveform_vocoders - hybrid_signal_neural_vocoders claims: - claim_id: streaming_voice_conversion_quality_degrades_significantly_when_offline_hubert_representations role: supports claim: Streaming voice conversion quality degrades significantly when offline HuBERT representations are naively replaced by causal alternatives, but knowledge distillation into an Emformer backbone can recover content accuracy with acceptable latency. source: §III.B, Table III evidence: Streaming voice conversion quality degrades significantly when offline HuBERT representations are naively replaced by causal alternatives, but knowledge distillation into an Emformer backbone can recover content accuracy with acceptable latency. confidence: high relevance: low - claim_id: causal_temporal_upsampling_via_pixel_shuffle_eliminates_the_checkerboard_artifacts role: supports claim: Causal temporal upsampling via pixel shuffle eliminates the checkerboard artifacts introduced by zero-padding non-causal vocoders, without sacrificing subjective quality. source: §III.D, Table III evidence: Causal temporal upsampling via pixel shuffle eliminates the checkerboard artifacts introduced by zero-padding non-causal vocoders, without sacrificing subjective quality. confidence: high relevance: medium - claim_id: explicit_style_modeling_with_clustering_based_vector_quantization_improves_zero role: supports claim: Explicit style modeling with clustering-based vector quantization improves zero-shot speaker similarity in streaming VC beyond what timbre embeddings alone provide. source: §III.C, Table I evidence: Explicit style modeling with clustering-based vector quantization improves zero-shot speaker similarity in streaming VC beyond what timbre embeddings alone provide. confidence: high relevance: low - claim_id: online_voice_conversion_systems_can_achieve_speaker_similarity_comparable_to role: supports claim: Online voice conversion systems can achieve speaker similarity comparable to offline systems when style transfer is modeled at chunk level rather than at the global utterance level. source: §IV.B, Table I evidence: Online voice conversion systems can achieve speaker similarity comparable to offline systems when style transfer is modeled at chunk level rather than at the global utterance level. confidence: high relevance: low limitations: - '- Evaluation is in English only; cross-lingual style transfer is untested. - The reference speaker must be fully available before streaming begins, limiting applications where reference is also captured in real-time. - CER is slightly higher than StreamVC because StreamVC reuses source pitch; Conan introduces some pitch variation that ASR penalizes as CER. - Model size is not reported; the cost of the Emformer + main model + CSV in production deployment is unclear. - Perceptual evaluation used 15 listeners per pair on Amazon Mechanical Turk style tasks; larger-scale evaluation would be needed to establish statistical robustness.' caveats: [] - id: '2509.02244' published_date: "2025-09-02" entry_date: '2026-07-30' year: 2025 venue: arXiv task: - codec architecture: - VAE - GAN relevance: medium evidence_role: - architecture_variant - acceleration_evidence current_role: minor method_family: - vae_codec_vocoder_decoders - adversarial_waveform_vocoders claims: - claim_id: single_stage_vector_quantization_without_residual_refinement_can_achieve_intelligibility role: supports claim: Single-stage vector quantization without residual refinement can achieve intelligibility and perceptual quality competitive with multi-level RVQ codecs at comparable bitrates. source: §5.2, Table 1 evidence: Single-stage vector quantization without residual refinement can achieve intelligibility and perceptual quality competitive with multi-level RVQ codecs at comparable bitrates. confidence: high relevance: medium - claim_id: training_a_hifi_gan_vocoder_on_codec_reconstructed_spectrograms_rather role: supports claim: Training a HiFi-GAN vocoder on codec-reconstructed spectrograms rather than clean references improves robustness to codec artefacts. source: §4.2 evidence: Training a HiFi-GAN vocoder on codec-reconstructed spectrograms rather than clean references improves robustness to codec artefacts. confidence: high relevance: high - claim_id: patchwise_quantization_of_mel_spectrograms_produces_a_2d_discrete_token role: supports claim: Patchwise quantization of mel spectrograms produces a 2D discrete token grid compatible with low-latency streaming at practical bitrates without architectural complexity. source: §3, §5.4 evidence: Patchwise quantization of mel spectrograms produces a 2D discrete token grid compatible with low-latency streaming at practical bitrates without architectural complexity. confidence: high relevance: low - claim_id: objective_metrics_such_as_mcd_can_diverge_from_perceptual_quality role: supports claim: Objective metrics such as MCD can diverge from perceptual quality metrics like PESQ when comparing codecs across different sampling rates and architectures. source: §5.2, Table 1 evidence: Objective metrics such as MCD can diverge from perceptual quality metrics like PESQ when comparing codecs across different sampling rates and architectures. confidence: high relevance: medium limitations: - The training dataset is described only as a "multilingual speech corpus" without further detail — speaker count, language distribution, total hours, and data sources are unreported. This makes the results impossible to reproduce and limits the generalisability of claims about performance on "diverse" speech. - No subjective evaluation (MOS or MUSHRA) is included; all comparisons are objective-only, which is particularly limiting for a perceptual task like codec quality assessment. The paper does not ablate the patch size (4×4 is the only configuration tested) or codebook size. MCD comparisons between 16kHz and 24kHz systems may be confounded by the sampling rate difference. The codec's real-world performance at varying packet-loss rates or noisy input conditions is untested. Future directions flagged by the authors include reducing the frequency-axis dimensionality toward a 1D token sequence, joint codec-vocoder training, and downstream use with autoregressive language models. caveats: [] - id: '2509.04667' published_date: "2025-09-04" entry_date: '2026-07-30' year: 2025 venue: arXiv task: - VC architecture: - GAN - hybrid relevance: medium evidence_role: - architecture_variant - acceleration_evidence - control_evidence current_role: active_evidence method_family: - adversarial_waveform_vocoders - hybrid_signal_neural_vocoders claims: - claim_id: limited_lookahead_in_causal_speech_encoders_substantially_improves_linguistic_content role: supports claim: Limited lookahead in causal speech encoders substantially improves linguistic content preservation with minimal latency penalty compared to purely causal encoders. source: §V.A, §V.B, Tables I–II evidence: Wave+CL accuracy improves from 53.16% at zero lookahead to 78.99% at 140ms, while end-to-end latency increases from 84.3ms to 203ms; extending to 280ms adds only 0.7pp accuracy with 120ms additional delay. confidence: high relevance: low - claim_id: better_content_encoding_in_streaming_anonymization_introduces_a_fundamental_tension role: complicates claim: 'Better content encoding in streaming anonymization introduces a fundamental tension: improved linguistic clarity reduces speaker anonymization strength under adversarial threat models.' source: §V.E, Table III evidence: Adding the contextual layer drops lazy-informed EER from 36.61% to 20.35% at zero lookahead, meaning the cleaner representations are more discriminative for speaker recognition attacks, directly trading anonymization quality for intelligibility. confidence: high relevance: low - claim_id: token_quantization_via_k_means_clustering_achieves_near_chance_speaker role: supports claim: Token quantization via k-means clustering achieves near-chance speaker verification performance in streaming anonymization by removing fine-grained speaker cues from content representations. source: §V.C, §V.D, §V.E, Tables IV–V evidence: Applying a 256-centroid k-means bottleneck raises lazy-informed EER from ~12% to ~47% (Wave+CL, 140ms lookahead), at a cost of WER rising from 2.09% to 9.52% and MOS falling from 3.79 to 3.22. confidence: high relevance: low - claim_id: streaming_voice_anonymization_systems_can_approach_offline_pipeline_anonymization_performance role: supports claim: Streaming voice anonymization systems can approach offline-pipeline anonymization performance when evaluated under the lazy-informed threat scenario, while retaining real-time latency. source: §V.F, Table VI evidence: DarkStream achieves 22.68% semi-informed EER in streaming mode (140ms lookahead), matching VoicePrivacy 2024 baselines B3 (26.28%) and B5a (22.09%) that require full-utterance processing. confidence: high relevance: low - claim_id: direct_waveform_synthesis_in_streaming_voice_conversion_maintains_acceptable_naturalness role: complicates claim: Direct waveform synthesis in streaming voice conversion maintains acceptable naturalness without mel-spectrogram intermediate representations, but k-means quantization introduced for privacy causes perceivable quality degradation beyond what objective metrics capture. source: §V.D, Table V evidence: MOS drops from 3.79 (Wave+CL) to 3.22 (Wave+CL+KMeans) with quantization; WER degrades only modestly, indicating that intelligibility metrics underestimate the perceptual impact of quantization artifacts. confidence: high relevance: high limitations: - DarkStream does not explicitly disentangle static speaker traits (accent, age, sex) from dynamic attributes (emotion, speaking style), leaving indirect identity cues potentially intact. The semi-informed EER of 22.68% remains well above chance, indicating meaningful residual linkability for well-resourced adversaries. - 'The privacy/quality trade-off exposed by the quantization ablation is steep: each MOS point recovered (by disabling k-means) costs roughly 30pp EER under the lazy-informed scenario. Systems requiring both high quality and robust anonymization against semi-informed attackers have no current solution in this architecture. Comparison of perceptual quality against the offline VoicePrivacy baselines is not reported, so whether DarkStream''s naturalness advantage over batch-processing pipelines is real remains an open question. Evaluation is limited to English (LibriTTS), and generalization to accented or code-switched speech is untested.' caveats: [] - id: '2509.07038' published_date: "2025-09-08" entry_date: '2026-07-30' year: 2025 venue: arXiv task: - singing architecture: - diffusion relevance: medium evidence_role: - architecture_variant - control_evidence current_role: active_evidence method_family: - diffusion_vocoder_and_restoration_systems claims: - claim_id: explicit_conditioning_on_a_signal_derived_attribute_provides_substantially_more role: supports claim: Explicit conditioning on a signal-derived attribute provides substantially more controllability than implicit predictors trained on the same attribute as an auxiliary loss. source: §V.B, Table III evidence: Adding an energy predictor to the baseline reduces energy MAE from 0.33 to 0.30, while replacing the predictor with explicit phoneme-level energy input achieves 0.14 - a 57.6% reduction over baseline vs. 9% for the predictor. confidence: high relevance: medium - claim_id: phoneme_level_aggregation_of_frame_level_features_enables_user_friendly role: supports claim: Phoneme-level aggregation of frame-level features enables user-friendly control in singing synthesis at acceptable precision cost. source: §III.B, §V.A, Table I, Table II evidence: Phoneme-level energy (L values, matching lyric and note sequence length) achieves energy MAE of 0.14 vs. 0.03 for frame-level (T values, roughly 1,000 per utterance), with higher MOS (3.78 vs. 3.57), demonstrating that coarser granularity trades minor precision loss for substantially reduced control burden. confidence: high relevance: low - claim_id: ground_truth_signal_features_can_substitute_for_manually_annotated_expressive role: supports claim: Ground-truth signal features can substitute for manually annotated expressive attributes as conditioning signals in singing voice synthesis. source: §I, §III.B evidence: Frame-level energy is extracted directly from ground-truth mel-spectrograms without human annotation, yet conditioning on it enables effective dynamics control, reducing annotation costs relative to technique-labelled datasets used in prior controllable SVS work. confidence: high relevance: low - claim_id: perceptual_quality_and_controllability_precision_do_not_align_monotonically_across role: complicates claim: Perceptual quality and controllability precision do not align monotonically across conditioning granularities in SVS. source: §V.A, Table I, Table II evidence: The frame-level model achieves the best energy MAE (0.03) but lower MOS (3.57 ± 0.18) than the phoneme-level model (MOS 3.78 ± 0.19, energy MAE 0.14), suggesting that very tight energy matching may over-constrain generation in ways that reduce perceived naturalness. confidence: high relevance: medium limitations: - MOS evaluation involved only 10 listeners, which the authors acknowledge constrains statistical reliability and generalizability. Results should be treated as directional rather than definitive. - The study uses only the Chinese subset of GTSinger and two speakers, leaving multilingual and speaker-diversity generalisability untested. The baseline DDPM architecture is explicitly noted to fall below state-of-the-art SVS performance, so the absolute quality figures do not reflect what the conditioning mechanism would achieve on a competitive backbone. Expressive attributes beyond dynamics (timbre, vibrato, advanced singing techniques) are not addressed. Phoneme-level energy may not capture fine-grained intra-phoneme temporal variation present in skilled vocal performance. caveats: [] - id: '2509.08379' published_date: "2025-09-10" entry_date: '2026-07-30' year: 2025 venue: arXiv task: - VC architecture: - diffusion - flow-matching relevance: medium evidence_role: - architecture_variant - acceleration_evidence - control_evidence current_role: active_evidence method_family: - diffusion_vocoder_and_restoration_systems - flow_matching_waveform_generation claims: - claim_id: operating_diffusion_and_flow_matching_models_in_a_compressed_latent role: supports claim: Operating diffusion and flow-matching models in a compressed latent representation improves audio quality and inference speed in non-parallel voice conversion compared to operating directly on mel-spectrograms. source: §III.A, §IV.H, Table IV, Table X evidence: LatentVoiceGrad-DPM outperforms VoiceGrad-DPM in pMOS (3.93 vs. 3.86), subjective qMOS (4.09 vs. 3.83), sMOS (3.05 vs. 2.63), and GPU RTF (0.034 vs. 0.045) on VCTK zero-shot any-to-any conversion. confidence: high relevance: low - claim_id: flow_matching_achieves_comparable_voice_conversion_quality_to_diffusion_models role: supports claim: Flow matching achieves comparable voice conversion quality to diffusion models with substantially fewer inference steps. source: §IV.G, §IV.J, Tables VII, IX evidence: VoiceGrad-FM with L=3 Euler steps achieves pMOS 3.85 (versus 3.86 for VoiceGrad-DPM at L=20), at GPU RTF 0.007 versus 0.045, a six-fold speedup with no quality penalty. confidence: high relevance: low - claim_id: adversarial_autoencoder_training_substantially_improves_audio_quality_and_speaker_similarity role: supports claim: Adversarial autoencoder training substantially improves audio quality and speaker similarity in latent-domain voice conversion compared to reconstruction-only training. source: §IV.F, Table V evidence: Adding adversarial loss during autoencoder training improves pMOS from 3.78 to 3.93 and SECS from 0.829 to 0.844 on VCTK zero-shot VC with the DPM generative model. confidence: high relevance: low - claim_id: discrete_acoustic_token_sequences_from_neural_audio_codecs_underperform_continuous role: complicates claim: Discrete acoustic token sequences from neural audio codecs underperform continuous spectral representations as the conversion domain for iterative generative voice conversion models. source: §IV.E, Tables II, III evidence: VoiceGrad with DAT features (DAC codec, dimensionality 1024) achieves pMOS 2.41 versus 3.86 for mel-spectrograms; LatentVoiceGrad-DAT shows CER 9.08 versus 2.99 for the mel-spectrogram variant, suggesting the combination of high-dimensional discrete tokens and score-based conversion is not yet effective. confidence: high relevance: low - claim_id: the_noise_injection_ratio_at_the_initial_point_of_the role: refines claim: The noise injection ratio at the initial point of the ODE in flow-matching voice conversion controls a three-way trade-off between audio quality, intelligibility, and speaker similarity. source: §IV.G, Table VI evidence: Systematic sweeps of r from 0 to 1 in VoiceGrad-FM and LatentVoiceGrad-FM show that increasing r raises SECS but degrades CER, with pMOS peaking around r=0.4-0.6; r=0.7 provides a practical balance across all three metrics. confidence: high relevance: low limitations: - Only objective metrics are reported for most ablations; the subjective listening test covers only 12 listeners and does not include FACodec as a comparison. The DAT evaluation is acknowledged as potentially sub-optimal due to untuned hyperparameters, so the conclusion that discrete tokens are unsuitable for this framework remains tentative. The model is evaluated on English speech only (VCTK and LibriTTS), and generalisation to other languages or noisy conditions is untested. The phoneme encoder relies on a supervised ASR model trained on English data, which ties the system to English without further adaptation. caveats: [] - id: '2509.13667' published_date: "2025-09-17" entry_date: '2026-07-30' year: 2025 venue: arXiv task: - TTS architecture: - GAN relevance: high evidence_role: - core_evidence current_role: active_evidence method_family: - adversarial_waveform_vocoders claims: - claim_id: knowledge_distillation_from_a_non_causal_teacher_can_substantially_recover role: supports claim: Knowledge distillation from a non-causal teacher can substantially recover the quality degradation introduced by causal convolutions in GAN-based neural vocoders. source: §V-A, §V-B, Table I, Table II evidence: DLL-APNet with full distillation reaches UTMOS 3.98, compared to 3.90 for causal APNet2 trained without distillation and 4.00 for non-causal APNet2; distillation depth ablation shows monotonic improvement from 3.73 (no blocks) to 3.98 (all 8 blocks) on VCTK unseen speakers. confidence: high relevance: high - claim_id: causal_adaptation_of_gan_based_vocoders_is_not_universally_applicable role: complicates claim: Causal adaptation of GAN-based vocoders is not universally applicable across all architectures. source: §V-A evidence: The causal variant of BigVGAN failed to converge during training and could not produce intelligible speech under the same causal substitution procedure applied to HiFi-GAN, iSTFTNet, APNet2, and Vocos, all of which converged. confidence: high relevance: high - claim_id: causal_constraints_in_neural_vocoders_disproportionately_degrade_pitch_accuracy_relative role: complicates claim: Causal constraints in neural vocoders disproportionately degrade pitch accuracy relative to spectral quality. source: §V-A, Table I evidence: 'Replacing standard convolutions with causal counterparts caused substantially larger relative degradations in F0-RMSE and V/UV error than in MCD and LAS-RMSE across all causal baseline vocoders on VCTK (e.g., causal HiFi-GAN: F0-RMSE 58.06 vs. 31.61 non-causal; causal iSTFTNet: F0-RMSE 54.43 vs. 32.87 non-causal).' confidence: high relevance: medium - claim_id: all_frame_level_vocoders_that_predict_spectral_features_before_waveform role: supports claim: All-frame-level vocoders that predict spectral features before waveform reconstruction are more resilient to causal adaptation than waveform-upsampling vocoders. source: §V-A, Table I evidence: Under the same causal substitution procedure on VCTK, causal APNet2 (UTMOS 3.90), causal Vocos (3.87), and DLL-APNet (3.98) outperform causal HiFi-GAN (3.88) and causal iSTFTNet (3.75), and computational cost (GFLOPS) is unchanged by causality conversion in all cases. confidence: high relevance: medium limitations: - Evaluation is entirely objective; no human listening test is reported. All perceptual quality claims rest on UTMOS (an automated MOS predictor), which may not fully capture quality degradation under causality constraints or differences between causal and non-causal models. - Evaluation uses a single English dataset (VCTK at 16 kHz) with 8 unseen speakers, so generalisation to other languages, higher sampling rates, or diverse speaking conditions is untested. Model size (parameter count) is not reported. The comparison against causal BigVGAN is limited to reporting convergence failure, with no diagnostic on why the architecture is more fragile to causal substitution. Future work is directed at reducing model size and computational overhead. caveats: [] - id: '2509.15085' published_date: "2025-09-18" entry_date: '2026-07-30' year: 2025 venue: arXiv task: - TTS architecture: - flow-matching relevance: medium evidence_role: - architecture_variant - acceleration_evidence current_role: active_evidence method_family: - flow_matching_waveform_generation claims: - claim_id: flow_matching_vocoders_can_match_or_exceed_gan_vocoders_in role: supports claim: Flow matching vocoders can match or exceed GAN vocoders in intrusive perceptual quality metrics even under a streaming constraint that reduces inference steps. source: §5, Table 1 evidence: MelFlow at N=5 (streaming) achieves PESQ 4.12 vs HiFi-GAN's 2.99 and SI-SDR -8.8 dB vs -29.9 dB on EARS-WHAM v2, while maintaining RTF < 1 on a consumer GPU. confidence: high relevance: high - claim_id: low_step_count_budgets_in_streaming_flow_diffusion_inference_degrade role: complicates claim: Low step-count budgets in streaming flow/diffusion inference degrade spectral fidelity even when perceptual quality remains competitive. source: §5, Table 1 evidence: MelFlow at N=5 shows LSD 1.0 and MCD 3.28 on EARS-WHAM v2, worse than HiFi-GAN (LSD 0.77, MCD 2.41), whereas the non-streaming N=25 variant recovers to LSD 0.7 and MCD 1.7. confidence: high relevance: low - claim_id: streaming_inference_for_multi_step_flow_and_diffusion_models_can role: supports claim: Streaming inference for multi-step flow and diffusion models can be achieved without train-test mismatch by maintaining N independent sets of per-layer causal cache buffers. source: §3.2 evidence: The N-buffer scheme produces outputs identical to offline batch processing (up to floating-point precision) because cached past activations never change, enabling standard parallel training alongside streaming deployment. confidence: high relevance: low - claim_id: high_algorithmic_latency_is_a_significant_barrier_for_diffusion_buffer role: complicates claim: High algorithmic latency is a significant barrier for diffusion-buffer streaming approaches, limiting practical streaming to methods that avoid coupling diffusion time to physical time. source: §2.2 evidence: The Diffusion Buffer achieves sub-second latency but requires at least 20 frames per buffer, giving a minimum effective algorithmic latency of ~340 ms at 16 ms hop size, which MelFlow avoids by using the per-layer cache scheme. confidence: high relevance: low limitations: - MelFlow operates at 16 kHz only; extension to 24 or 44.1 kHz is not evaluated and the step-count constraint tightens at higher sample rates. Spectral detail metrics (LSD, MCD) at N=5 are weaker than HiFi-GAN, suggesting a quality-latency tradeoff that few-step distillation might address (noted as future work). The training dataset is relatively small (~87 hours); behavior on highly diverse or noisy speech is untested. Comparison vocoders operating at 22-24 kHz are evaluated after downsampling to 16 kHz, introducing a resampling artifact that may affect fairness. caveats: [] - id: '2509.18806' published_date: "2025-09-23" entry_date: '2026-07-30' year: 2025 venue: arXiv task: - TTS architecture: - GAN relevance: medium evidence_role: - architecture_variant - acceleration_evidence - control_evidence current_role: active_evidence method_family: - adversarial_waveform_vocoders claims: - claim_id: dual_stream_architectures_for_joint_magnitude_phase_estimation_are_substantially role: supports claim: Dual-stream architectures for joint magnitude-phase estimation are substantially more prone to performance collapse on large, acoustically diverse multi-speaker training data than single-stream architectures, even when modeling unit, output parameterization, and loss function are matched. source: §2.1, Table 1, Fig. 1 evidence: APNet2 (dual-stream) collapses to PESQ 1.626 on LibriTTS versus Vocos (single-stream) at PESQ 3.487, despite near-parity on single-speaker LJSpeech; aligning APNet2 with Vocos on unit/output/loss (APNet2*) still leaves a large gap (PESQ 2.556 vs. 3.607). confidence: high relevance: medium - claim_id: introducing_weight_sharing_or_periodic_feature_exchange_between_otherwise_independent role: supports claim: Introducing weight sharing or periodic feature exchange between otherwise-independent magnitude and phase estimation streams substantially narrows the performance gap to fully shared-stream vocoder designs. source: §2.2, §4.1, Table 2 evidence: Partially-shared (6 shared layers) and shuffled dual-stream topology variants raise PESQ from 2.556 (fully separate) to 3.461 and 3.4 respectively, approaching the shared-stream Vocos-D baseline of 3.607, at comparable parameter counts. confidence: high relevance: high - claim_id: for_t_f_domain_vocoders_conditioned_on_mel_spectrograms_aligning role: refines claim: For T-F domain vocoders conditioned on mel-spectrograms, aligning the input representation with the target spectrum's range space (rather than using the compressed mel-spectrogram directly) improves reconstruction quality independently of any architectural change. source: §2.3, §4.2, Table 3 evidence: Replacing the raw mel-spectrogram with a pseudo-inverse mel projection as input raises the fully separate APNet2* variant's PESQ from 2.556 to 3.197 and UTMOS from 1.886 to 2.987, with no modification to the network topology. confidence: high relevance: medium - claim_id: an_auxiliary_loss_term_that_couples_the_gradient_updates_of role: supports claim: An auxiliary loss term that couples the gradient updates of otherwise-independent output branches can substitute for explicit architectural interaction between those branches. source: §2.4, §4.2, §4.3, Table 3, Fig. 4 evidence: Requiring the phase branch to also produce a magnitude-consistent term used in the magnitude loss (MI-RI) raises separate-stream APNet2*'s PESQ from 3.197 to 3.436 without altering the forward-pass magnitude prediction path, and produces phase-branch magnitude estimates with harmonic structure resembling the target that are absent in the unmodified model. confidence: high relevance: medium limitations: - Evaluation relies entirely on automatic objective metrics (PESQ, UTMOS, VISQOL, MCD, M-STFT, periodicity RMSE, V/UV F1, pitch RMSE); no human listening test (MOS or preference test) is reported, so perceptual validation of the reported gains is absent. - The analysis is confined to two vocoder families, Vocos and APNet2, that were deliberately chosen for their structural similarity in modeling unit, loss, and training pipeline; it is not established whether the topology-collapse finding or the three proposed fixes generalize to vocoders with substantially different modeling units (e.g. non-ConvNeXt backbones) or to time-domain vocoders. The paper also does not report inference-time latency or compute overhead for the pseudo-inverse mel projection or the added output branch, despite emphasizing efficiency as a motivation for T-F domain approaches over time-domain upsampling. caveats: [] - id: '2509.19025' published_date: "2025-09-23" entry_date: '2026-07-30' year: 2025 venue: arXiv task: - codec architecture: - GAN relevance: medium evidence_role: - architecture_variant current_role: active_evidence method_family: - adversarial_waveform_vocoders claims: - claim_id: deterministic_nearest_neighbor_codeword_selection_in_residual_vector_quantization_is role: supports claim: Deterministic nearest-neighbor codeword selection in residual vector quantization is fragile to small input perturbations, producing codeword reassignments that compound across RVQ stages into audible reconstruction artefacts. source: §2.1, Figure 1 evidence: Codeword-shift analysis on Encodec's first RVQ stage (24 kHz, 6 kbps) shows that although the noisy codeword usually matches the clean top-1 choice, a pronounced long tail of shifts (k>1) occurs when 15 dB SNR DEMAND noise is added to 120 clean VCTK utterances. confidence: high relevance: medium - claim_id: neural_speech_codec_noise_robustness_can_be_improved_without_any role: supports claim: Neural speech codec noise robustness can be improved without any paired noisy-clean training data by simulating perturbation-induced instability directly at the quantization step. source: §3.4, Table 1 evidence: Fine-tuning Encodec and WavTokenizer with distance-weighted probabilistic top-K sampling, trained exclusively on clean speech, improves SI-SDR, PESQ, STOI, and UTMOS at both 15 dB and 10 dB SNR with statistically significant paired t-tests (e.g., Encodec UTMOS 3.475 to 3.586 at 15 dB). confidence: high relevance: low - claim_id: training_on_paired_noisy_clean_data_achieves_stronger_robustness_than role: complicates claim: Training on paired noisy-clean data achieves stronger robustness than perturbation-only training under the specific noise conditions it was exposed to, but this advantage does not transfer to unseen noise types and comes at the cost of clean-speech quality. source: §3.5.1, Tables 2-3 evidence: A noise-exposed fine-tuning baseline (Closest*) outperforms the proposed clean-data-only method on some noisy-speech metrics (e.g., PESQ) under matched training/test noise, yet degrades clean-speech scores and is outperformed by the proposed method on all four metrics across three held-out noise types not present in either method's training. confidence: high relevance: medium - claim_id: introducing_training_time_quantization_perturbations_in_a_curriculum_like_schedule role: refines claim: Introducing training-time quantization perturbations in a curriculum-like schedule, from the finest residual quantizer stage toward the coarsest, stabilizes robustness training more effectively than perturbing all quantizer stages simultaneously. source: §3.5.3, Figure 2 evidence: PESQ and UTMOS improve steadily as probabilistic top-K sampling is progressively rolled out from the 6th to the 1st VQ of Encodec's RVQ, whereas applying the same sampling to all six VQs at once ("Direct Top-K") yields inferior results due to premature perturbation of core structural features. confidence: high relevance: medium limitations: - The evaluation is confined to a VCTK subset (a single-domain, studio-quality English read-speech corpus) and to relatively mild noise conditions (10-15 dB SNR from DEMAND). Whether the robustness gains extend to more severe noise, non-additive distortions (e.g., reverberation, codec cascading, packet loss), or multilingual/spontaneous speech is untested. The method is validated on only two codecs (Encodec, WavTokenizer); both use RVQ-style quantization, so its applicability to non-RVQ or single-codebook designs with different codebook geometries is unconfirmed. The paper itself notes extending the framework to more streamable architectures and integrating it with large speech-language models as future directions, but does not attempt either. caveats: [] - id: '2509.21968' published_date: "2025-09-26" entry_date: '2026-07-30' year: 2025 venue: arXiv task: - codec architecture: - GAN - VAE relevance: medium evidence_role: - architecture_variant - evaluation_caution current_role: active_evidence method_family: - adversarial_waveform_vocoders - vae_codec_vocoder_decoders claims: - claim_id: a_shared_single_codebook_can_be_structured_with_overlapping_nested role: supports claim: A shared single codebook can be structured with overlapping, nested domain partitions rather than rigid disjoint splits, improving both reconstruction and downstream generation quality relative to a rigid-split design. source: §4.4, Tables 3–4 evidence: Ablating codebook design at fixed codebook size and data scale, the nested codebook reduces reconstruction WER from 4.21 (rigid-split) to 3.99 and downstream TTS generation WER from 6.26 to 4.99 on LibriSpeech-PC test-clean. confidence: high relevance: medium - claim_id: distilling_frame_level_representations_from_multiple_domain_specific_self_supervised role: supports claim: Distilling frame-level representations from multiple domain-specific self-supervised teacher models into one acoustic codec can improve reconstruction and generation quality across all the covered domains simultaneously, not just the domain of a single teacher. source: §3.3, §4.4, Tables 3–4 evidence: Adding multi-domain distillation (WavLM for speech, MuQ for vocal/music, BEATs for sound) on top of the nested codebook further raises the speech-partition code-usage ratio from 37.1% to 38.2% and improves downstream TTS generation WER from 4.99 to 4.51. confidence: high relevance: low - claim_id: unifying_multiple_audio_domains_into_a_single_shared_quantization_codebook role: complicates claim: Unifying multiple audio domains into a single, shared quantization codebook does not close the gap with domain-specific single-layer codecs on every reconstruction metric, even when the unified model uses a larger codebook and lower token rate. source: §4.2, Table 1 evidence: On LibriSpeech test-clean, AUV's PESQ-WB (2.40) and SPK-SIM (0.81) trail dedicated speech codecs such as DAC (4.01, 0.95) and are roughly on par with, not clearly better than, BigCodec and UniCodec. confidence: high relevance: medium - claim_id: larger_unified_codebooks_intended_to_accommodate_more_audio_domains_can role: complicates claim: Larger unified codebooks intended to accommodate more audio domains can hurt downstream autoregressive generation quality by increasing the modeling burden on the generative model consuming the tokens, even when reconstruction quality is unaffected. source: §4.3, Table 4 evidence: Scaling the codebook from 16,384 to 20,480 entries (C0 vs. C2) yields slightly worse downstream generation WER (4.51 vs. 4.89) despite improved reconstruction metrics; training an autoregressive model on codes from a still-larger 131,072-entry codebook (MagiCodec) reportedly failed outright. confidence: high relevance: medium limitations: - The paper reports no total parameter count for the AUV encoder-decoder, limiting direct efficiency comparison with baselines whose sizes are known. Downstream generative evaluation is restricted to a single autoregressive TTS backbone (EmoVoice) trained on a comparatively small 1K-hour subset of LibriSpeech, so it is unclear whether the reconstruction and generation gains hold at larger downstream training scales or with non-autoregressive generators. Speaker similarity in the downstream TTS setting remains low in absolute terms (SPK-SIM ≈ 0.43–0.44) across all codecs tested, including AUV, suggesting the codec-level improvements shown here do not yet translate into strong speaker fidelity for generated speech. The domain-label input used during training but withheld at inference creates a train/inference mismatch whose effect on codebook index selection is only indirectly probed via the reported index-distribution statistics, not directly ablated. caveats: [] - id: '2509.24773' published_date: "2025-09-29" entry_date: '2026-07-30' year: 2025 venue: arXiv task: - TTS architecture: - flow-matching relevance: medium evidence_role: - architecture_variant current_role: active_evidence method_family: - flow_matching_waveform_generation claims: - claim_id: joint_training_of_two_video_conditioned_audio_generation_tasks_with role: supports claim: Joint training of two video-conditioned audio generation tasks with different output characteristics does not need to degrade either task's performance, provided the conditioning pathways for each task's signals are architecturally separated. source: §4.4, Figure 4 evidence: Learning curves comparing joint (V2S+VisualTTS), sound-only, and speech-only training at matched epochs show the joint-trained model reaches performance parity with each single-task model on FAD/DeSync (sound) and WER/UTMOS (speech), directly contrasting with a prior report that joint training of these tasks causes mutual degradation. confidence: high relevance: medium - claim_id: the_choice_between_cross_attention_and_channel_concatenation_conditioning_in role: refines claim: The choice between cross-attention and channel-concatenation conditioning in a DiT-based generative model should be determined by whether the conditioning signal is temporally dense/aligned or sparse/global, not fixed uniformly across all condition types. source: §4.3, Figure 3 evidence: A 5-variant ablation shows concatenation-based conditioning improves WER and lip-sync error for temporally-dense signals (transcript, sound/speech sync cues), while cross-attention conditioning improves FAD and video-audio semantic alignment for the sparse, global video-semantic feature; self-attention weight visualization shows a diagonal locality bias explaining the concatenation result. confidence: high relevance: medium - claim_id: feature_level_synthesis_of_paired_training_examples_constructed_by_mixing role: supports claim: Feature-level synthesis of paired training examples, constructed by mixing independently-sourced single-modality data in representation space, can substitute for scarce real multimodal joint-generation training data. source: §4.4, Table 4 evidence: Fine-tuning on synthetic sound-speech mixtures alone yields the best WER (19.4) and UTMOS (2.07) on the V2C-Animation joint-generation benchmark among all training-data configurations tested, exceeding fine-tuning on the small (8k-sample) real V2C corpus alone (36.5 WER). confidence: high relevance: medium - claim_id: unified_video_conditioned_generation_systems_that_depend_on_frozen_pretrained role: complicates claim: Unified video-conditioned generation systems that depend on frozen pretrained feature extractors for their conditioning signals inherit those extractors' representational limits, and systems adapted with synthetic joint data remain bounded by the distributional gap between synthetic and real joint data. source: §Limitations evidence: The authors state that reliance on frozen pretrained extractors (CLIP, Synchformer, AV-HuBERT) may limit capture of fine-grained audio-visual nuances, and that generation quality on real joint scenarios is inherently constrained by the distributional gap of the synthetic training data used for joint-generation fine-tuning. confidence: high relevance: medium limitations: - Beyond the frozen-extractor and synthetic-data-gap limitations the authors state directly, the joint sound-speech generation results are evaluated on a single benchmark (V2C-Animation) drawn entirely from animated content, and the joint-generation comparison is against pipeline baselines composed by the authors rather than against other end-to-end unified systems, since none are publicly available for this exact three-way task. Code was not released at the time of publication ("Demos and code will be soon released" per the paper), though a project page with generated examples is available. caveats: [] - id: '2510.00264' published_date: "2025-09-30" entry_date: '2026-07-30' year: 2025 venue: arXiv task: - codec - evaluation architecture: - GAN relevance: medium evidence_role: - architecture_variant - acceleration_evidence - evaluation_caution current_role: minor method_family: - adversarial_waveform_vocoders claims: - claim_id: residual_vector_quantized_convolutional_codecs_trained_end_to_end_with role: supports claim: Residual-vector-quantized convolutional codecs trained end-to-end with adversarial and multi-scale mel-spectrogram losses can be configured to satisfy joint sub-1 kbps-to-6 kbps bitrate, sub-50 ms latency, and sub-3000 MFLOPS compute budgets simultaneously. source: §3.1, §3.2, Table 2, Table 3 evidence: The Track 1 baseline totals 691.35 MFLOPS and 20 ms latency (within a 700 MFLOPS / 30 ms budget) and the Track 2 baseline totals 2546.2 MFLOPS and 40 ms latency (within a 2600 MFLOPS / 50 ms budget), both operating across the full 1-6 kbps range via quantizer dropout. confidence: high relevance: low - claim_id: joint_denoising_and_dereverberation_enhancement_codecs_require_substantially_more_compute role: complicates claim: Joint denoising-and-dereverberation ("enhancement") codecs require substantially more compute than transparency-only codecs even when the coding architecture and bitrate range are otherwise matched. source: §3.1, §3.2, Table 2, Table 3 evidence: The Track 2 (enhancement) baseline requires 2546.2 MFLOPS overall versus 691.35 MFLOPS for the Track 1 (transparency-only) baseline, roughly 3.7x more compute, despite sharing the same 6-layer, 1,024-codeword RVQ design and 1-6 kbps target range. confidence: high relevance: medium - claim_id: automatic_objective_codec_quality_metrics_can_disagree_substantially_with_one role: complicates claim: Automatic objective codec-quality metrics can disagree substantially with one another at low bitrates, complicating single-metric benchmarking of low-resource codecs. source: §4, Table 4, Table 5 evidence: At 1 kbps under clean conditions, SCOREQ_Ref reports 1.15/1.01 (Track 1/2, near floor) while Audiobox_AE_CE reports 3.9/3.96 for the same systems and conditions in the same evaluation table. confidence: high relevance: low - claim_id: filtering_public_speech_corpora_by_estimated_signal_quality_snr_reverberation role: supports claim: Filtering public speech corpora by estimated signal quality (SNR, reverberation, bandwidth) before training a low-resource codec substantially reduces the usable data volume relative to the source corpora. source: §2, Table 1 evidence: Quality- and diversity-based curation across LibriTTS, VCTK, EARS, Librivox (DNS5), MLS, and GLOBE retains 702.7 of 1,641.7 total hours (42.8% overall retention), ranging from 24.2% (LibriTTS) to 100% (EARS) by dataset. confidence: high relevance: low limitations: - All reported results (Table 4, Table 5) are automatic objective metrics on the challenge's open development test set; the paper reports no human listening-test (MOS/subjective) results, and the blind test set used for official challenge scoring had not yet been released at the time of writing, so these baseline numbers may not reflect final challenge-condition performance. - Checkpoint selection relies solely on validation multi-scale mel-spectrogram loss; the authors themselves note this prioritizes implementation simplicity over correlation with subjective quality, and suggest combining objective metrics that correlate better with listening tests as a likely improvement. The Track 2 validation set uses on-the-fly (rather than fixed, offline) noise and reverberation augmentation, which the authors acknowledge may increase variance in validation loss estimates. No ablations isolate the individual contributions of the loss terms, the data curation pipeline, or the quantizer-dropout bitrate-scalability scheme. caveats: [] claim_clusters: - id: gan_vocoders_match_autoregressive_quality_at_lower_latency claim: GAN vocoders can match autoregressive perceptual quality while reducing waveform-generation latency by orders of magnitude. status: emerging confidence: medium supporting_papers: - '2010.05646' contradicting_papers: [] refining_papers: - iclr-2025-tQ1PmLfPBL caveats: - Speed comparisons depend on hardware, batching, sampling rate, and whether upstream acoustic generation is included. last_reviewed: '2026-07-30' - id: discriminator_design_is_central_to_gan_vocoder_quality claim: Discriminator design is a primary determinant of GAN-vocoder quality and training stability. status: strongly_supported confidence: high supporting_papers: - '2010.05646' - '2210.13438' - '2508.05207' - '2508.07711' - '2508.12001' - interspeech-2025-1081 - interspeech-2025-1763 contradicting_papers: [] refining_papers: [] caveats: - Discriminator gains may depend on matching generator representation, loss balance, and data scale. last_reviewed: '2026-07-30' - id: multi_period_and_spectral_discriminators_capture_complementary_structure claim: Periodicity-aware and multi-resolution spectral discriminators capture complementary speech structure. status: strongly_supported confidence: high supporting_papers: - '2010.05646' - '2206.04658' - '2210.13438' - '2306.00814' - iclr-2025-tQ1PmLfPBL - '2508.05207' - '2508.12001' contradicting_papers: [] refining_papers: - interspeech-2025-1763 caveats: - No single discriminator family is uniformly best across waveform, spectrogram, and codec-decoder generators. last_reviewed: '2026-07-30' - id: frequency_domain_generation_improves_phase_and_efficiency claim: Time–frequency-domain vocoders can improve phase modeling, artifact control, and efficiency relative to direct waveform generation. status: strongly_supported confidence: high supporting_papers: - '2306.00814' - '2507.20731' - '2508.05207' - '2508.12001' - interspeech-2025-2739 contradicting_papers: [] refining_papers: - '2509.18806' caveats: - Frequency-domain systems remain sensitive to phase wrapping, representation alignment, and branch interaction. last_reviewed: '2026-07-30' - id: explicit_phase_modeling_can_reduce_adversarial_dependence claim: Explicit phase modeling can reduce or remove the need for adversarial training in high-quality vocoders. status: emerging confidence: medium supporting_papers: - '2508.07711' contradicting_papers: [] refining_papers: - '2507.01611' caveats: - Evidence is strongest for architectures designed around phase prediction rather than GAN objectives removed post hoc. last_reviewed: '2026-07-30' - id: periodic_and_antialiasing_biases_improve_robustness claim: Periodic activations and anti-aliasing constraints improve GAN-vocoder robustness and suppress high-frequency artifacts. status: strongly_supported confidence: high supporting_papers: - '2206.04658' - '2306.00814' - interspeech-2025-2739 contradicting_papers: [] refining_papers: [] caveats: - Architectural robustness still depends on diverse training data and matched bandwidth. last_reviewed: '2026-07-30' - id: predicted_feature_mismatch_requires_vocoder_adaptation claim: Vocoders benefit from training or adaptation on predicted and reconstructed conditioning features rather than clean references alone. status: strongly_supported confidence: high supporting_papers: - '1712.05884' - '2010.05646' - interspeech-2025-0854 - '2509.02244' contradicting_papers: [] refining_papers: [] caveats: - Adaptation can overfit the artifact distribution of one acoustic model or codec. last_reviewed: '2026-07-30' - id: data_diversity_drives_universal_vocoding claim: Speaker, language, recording, and domain diversity are major drivers of universal vocoder generalization. status: strongly_supported confidence: high supporting_papers: - '2306.00814' - '2507.20731' - interspeech-2025-2739 - '2508.18006' contradicting_papers: [] refining_papers: [] caveats: - Architecture, model scale, filtering, and dataset diversity are often changed together. last_reviewed: '2026-07-30' - id: gan_vocoder_scaling_requires_training_stabilization claim: Scaling GAN-vocoder capacity improves quality only when accompanied by architectural and optimization stabilization. status: emerging confidence: medium supporting_papers: - '2010.05646' contradicting_papers: [] refining_papers: [] caveats: - Naive capacity increases can destabilize adversarial training or obscure data-quality effects. last_reviewed: '2026-07-30' - id: signal_processing_constraints_improve_extrapolation_and_efficiency claim: Explicit signal-processing structure can improve vocoder extrapolation, parameter efficiency, and data efficiency. status: strongly_supported confidence: high supporting_papers: - '2507.01611' - '2507.20731' - interspeech-2025-0406 - interspeech-2025-1819 contradicting_papers: [] refining_papers: - '2509.18806' caveats: - Analytic components introduce their own estimation errors, including V/UV and phase failures. last_reviewed: '2026-07-30' - id: causal_streaming_vocoders_can_approach_offline_quality claim: Causal GAN vocoders can approach offline quality with limited lookahead, normalization changes, or teacher distillation. status: emerging confidence: medium supporting_papers: - interspeech-2025-1819 - '2507.14534' contradicting_papers: [] refining_papers: - '2509.13667' caveats: - Causal adaptation remains architecture-dependent and often degrades pitch more than spectral quality. last_reviewed: '2026-07-30' - id: flow_and_diffusion_trade_speed_for_iterative_quality claim: Flow-matching and diffusion vocoders can improve iterative waveform refinement but trade quality against inference cost. status: strongly_supported confidence: high supporting_papers: - iclr-2025-tQ1PmLfPBL - iclr-2025-uxDFlPGRLX - 2025.naacl-long.110 - interspeech-2025-0854 - interspeech-2025-1081 - '2509.15085' contradicting_papers: [] refining_papers: - iclr-2025-tQ1PmLfPBL - iclr-2025-uxDFlPGRLX caveats: - One-step distillation narrows the latency gap, but matched subjective comparisons remain limited. last_reviewed: '2026-07-30' - id: gan_decoders_support_low_bitrate_neural_codecs claim: Adversarial decoders improve perceptual reconstruction in low-bitrate neural audio codecs. status: strongly_supported confidence: high supporting_papers: - '2210.13438' - '2305.02765' - '2306.00814' - '2409.05377' - '2508.05207' - '2509.02244' - '2509.19025' - '2509.21968' contradicting_papers: [] refining_papers: - interspeech-2025-0347 - '2509.04667' - '2509.21968' caveats: - Codec quality depends jointly on quantizer rate, latent structure, decoder capacity, and evaluation domain. last_reviewed: '2026-07-30' - id: quantization_and_compression_disproportionately_stress_vocoders claim: Aggressive quantization and compressed conditioning disproportionately stress vocoder quality. status: strongly_supported confidence: high supporting_papers: - '2210.13438' - '2305.02765' - '2409.05377' - interspeech-2025-1122 - interspeech-2025-2726 - '2509.02244' contradicting_papers: [] refining_papers: - '2509.04667' - '2509.18806' - '2509.19025' caveats: - Objective reconstruction metrics may miss perceptually salient artifacts introduced by quantization. last_reviewed: '2026-07-30' - id: objective_metrics_can_misrank_vocoder_quality claim: Automatic spectral and MOS-prediction metrics can misrank vocoder perceptual quality. status: emerging confidence: medium supporting_papers: - '2508.07711' - '2509.02244' contradicting_papers: [] refining_papers: - '2507.01611' caveats: - Metric disagreement varies with sampling rate, bandwidth, phase, and intended signal transformation. last_reviewed: '2026-07-30' - id: vocoder_bottlenecks_interact_with_upstream_acoustic_models claim: Vocoder capacity and artifacts interact with upstream acoustic-model representations and prosodic diversity. status: strongly_supported confidence: high supporting_papers: - '1712.05884' - '2010.05646' - '2508.01796' - '2508.06870' - '2508.07711' - '2508.12001' - interspeech-2025-0854 - interspeech-2025-1122 - '2508.18006' contradicting_papers: [] refining_papers: - interspeech-2025-2573 - '2509.04667' - '2509.18806' caveats: - End-to-end gains cannot always be attributed cleanly to the acoustic model or vocoder alone. last_reviewed: '2026-07-30' - id: gan_vocoders_remain_effective_in_low_resource_and_cross_domain_settings claim: GAN vocoders remain effective in low-resource, cross-lingual, voice-conversion, and non-standard acoustic settings. status: strongly_supported confidence: high supporting_papers: - '2010.05646' - interspeech-2025-1081 - interspeech-2025-2151 contradicting_papers: [] refining_papers: [] caveats: - Evidence spans heterogeneous tasks and rarely controls training data, language, and upstream models jointly. last_reviewed: '2026-07-30' method_families: - id: adversarial_waveform_vocoders name: Adversarial waveform vocoders summary: GAN vocoders synthesize waveforms or time–frequency representations with discriminator-guided perceptual training and predominantly one-pass inference. papers: - '2010.05646' - '2104.00355' - '2106.15561' - '2206.04658' - '2210.13438' - '2305.02765' - '2306.00814' - '2409.05377' - '2411.01156' - 2025.chipsal-1.18 - '2502.05512' - iclr-2025-uxDFlPGRLX - 2025.findings-naacl.130 - 2025.findings-naacl.279 - 2025.naacl-short.65 - 2025.americasnlp-1.1 - '2507.01611' - '2507.20731' - '2508.01796' - '2508.05207' - '2508.06870' - '2508.07711' - '2508.12001' - interspeech-2025-0347 - interspeech-2025-0406 - interspeech-2025-0854 - interspeech-2025-0998 - interspeech-2025-1081 - interspeech-2025-1122 - interspeech-2025-1747 - interspeech-2025-1763 - interspeech-2025-1819 - interspeech-2025-2151 - interspeech-2025-2573 - interspeech-2025-2586 - interspeech-2025-2726 - interspeech-2025-2739 - '2508.18006' - '2507.14534' - '2509.02244' - '2509.04667' - '2509.13667' - '2509.18806' - '2509.19025' - '2509.21968' - '2510.00264' open_questions: - Which discriminator and generator inductive biases transfer reliably across speakers, languages, sampling rates, and non-speech audio? - id: transformer_conditioned_vocoder_systems name: Transformer-conditioned vocoder systems summary: Transformer encoder–decoder pipelines pair learned acoustic representations with neural vocoders for end-to-end or modular speech generation. papers: - '1712.05884' - '2006.04558' - '2106.15561' - 2025.chipsal-1.18 - 2025.findings-naacl.130 - 2025.naacl-long.591 - 2025.americasnlp-1.1 - '2507.08319' - '2508.06870' - interspeech-2025-0854 - interspeech-2025-1122 - interspeech-2025-2031 - interspeech-2025-2586 open_questions: - How should acoustic intermediates be matched to vocoder capacity under low-resource, multilingual, and quantized deployment? - id: flow_matching_waveform_generation name: Flow-matching waveform generation summary: Flow-matching vocoders and postfilters replace or complement adversarial generation with iterative or distilled continuous transport. papers: - '2106.15561' - iclr-2025-tQ1PmLfPBL - iclr-2025-uxDFlPGRLX - 2025.naacl-long.110 - interspeech-2025-0455 - interspeech-2025-0854 - interspeech-2025-2573 - '2509.08379' - '2509.15085' - '2509.24773' open_questions: - Can flow-matching vocoders retain their quality advantage at the one-step and streaming inference budgets where GANs are strongest? - id: vae_codec_vocoder_decoders name: VAE and codec vocoder decoders summary: VAE and neural-codec systems use adversarial or hybrid decoders to reconstruct waveform detail from compressed continuous or discrete latents. papers: - '2104.00355' - '2106.15561' - '2305.02765' - '2409.05377' - '2508.12001' - interspeech-2025-0347 - interspeech-2025-0455 - '2509.02244' - '2509.21968' open_questions: - How should quantizer structure, decoder capacity, and adversarial losses be co-designed across bitrates and audio domains? - id: diffusion_vocoder_and_restoration_systems name: Diffusion vocoder and restoration systems summary: Diffusion systems generate or restore waveforms iteratively, often trading inference cost for robust stochastic reconstruction. papers: - '2106.15561' - '2508.01796' - interspeech-2025-0998 - interspeech-2025-1747 - interspeech-2025-1763 - '2509.07038' - '2509.08379' open_questions: - Which distillation and conditioning strategies preserve restoration quality without the full iterative cost? - id: hybrid_signal_neural_vocoders name: Hybrid signal-processing and neural vocoders summary: Hybrid vocoders combine learned generation with explicit phase, filterbank, source–filter, codec, or signal-degradation structure. papers: - 2025.findings-naacl.279 - 2025.naacl-short.65 - '2508.12001' - interspeech-2025-0455 - interspeech-2025-2726 - '2507.14534' - '2509.04667' open_questions: - Which analytic constraints improve extrapolation and efficiency without imposing brittle estimation bottlenecks? - id: autoregressive_waveform_and_token_decoders name: Autoregressive waveform and token decoders summary: Autoregressive systems generate waveform samples or acoustic tokens sequentially and provide a quality reference for faster vocoders. papers: - '1609.03499' - '2106.15561' - '2411.01156' - '2502.05512' - 2025.findings-naacl.279 - 2025.findings-acl.115 open_questions: - Where do autoregressive decoders remain preferable once latency, context length, and acoustic-token hierarchy are controlled? reassessment_queue: - id: gan_vocoders_match_autoregressive_quality_at_lower_latency type: claim_status reason: Speed and quality comparisons use different hardware, sampling rates, and upstream pipelines. trigger: Matched multi-vocoder studies report listener quality and end-to-end latency on common hardware. due: 2026-10 current_assessment: emerging watch_for: - Matched hardware benchmarks - End-to-end latency and listener tests - id: explicit_phase_modeling_can_reduce_adversarial_dependence type: claim_status reason: Evidence is concentrated in explicit-phase systems and may not generalize to ordinary waveform generators. trigger: Independent architectures reproduce GAN-free quality under matched training data and subjective tests. due: 2026-10 current_assessment: emerging watch_for: - GAN-free phase-vocoder replications - Matched subjective evaluations - id: objective_metrics_can_misrank_vocoder_quality type: benchmark_validity reason: Metric disagreement recurs, but the direction depends on bandwidth, phase, and task. trigger: Cross-domain listener-calibration studies identify robust metric combinations for vocoder ranking. due: 2026-10 current_assessment: emerging watch_for: - Cross-domain metric calibration - Phase- and bandwidth-aware perceptual metrics - id: flow_and_diffusion_trade_speed_for_iterative_quality type: claim_status reason: Distillation and streaming constraints are rapidly narrowing the iterative-model latency gap. trigger: One-step flow or diffusion vocoders match GAN latency and listener preference across datasets. due: 2026-10 current_assessment: strongly_supported watch_for: - One-step iterative vocoder tests - Streaming subjective comparisons - id: data_diversity_drives_universal_vocoding type: claim_status reason: Data diversity is usually confounded with model scale and architectural changes. trigger: Controlled studies vary domain and speaker diversity while holding architecture and training budget fixed. due: 2026-10 current_assessment: strongly_supported watch_for: - Controlled diversity ablations - Out-of-domain speech and non-speech tests open_questions: - Which discriminator combination gives the best transferable quality–compute trade-off across speech, music, and codec decoding? - When can explicit phase and signal-processing constraints replace adversarial supervision without reducing perceptual quality? - How should vocoders be trained against predicted, quantized, bandwidth-limited, and otherwise mismatched conditioning features? - Can one-step flow or diffusion vocoders consistently match GAN latency while retaining iterative-generation robustness? - Which evaluation suite reliably captures phase, pitch, bandwidth, transient, and out-of-domain vocoder artifacts? - How should causal, low-bit, and hardware-aware vocoders trade model size, lookahead, energy use, and listener quality? trend_notes: - GAN vocoders have shifted from generic waveform discriminators toward periodic, spectral, multi-band, and projected discrimination. - Frequency-domain and explicit-phase vocoders increasingly use signal structure to reduce aliasing, parameter count, and adversarial dependence. - GAN decoders remain central to neural codecs, where discriminator design is being co-optimized with quantizer structure and bitrate. - Flow matching and diffusion increasingly appear as vocoders and postfilters, with distillation targeting GAN-like inference budgets. - Evaluation is expanding beyond in-domain MOS toward mismatch robustness, streaming, quantization, bandwidth, and metric disagreement.