Pith. sign in

REVIEW 3 major objections 6 minor 23 references

Splitting speaker from speech lets neural codecs stream in 660ms blocks

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · glm-5.2

2026-07-07 21:37 UTC pith:S4QF7OEL

load-bearing objection Probing analysis of TiCodec's TIRE module is solid and useful; Dual-TIRE is a reasonable extension with modest gains; but the streaming evaluation in Table 5 has metric values that are internally inconsistent with the rest of the paper and cannot be interpreted as written. the 3 major comments →

arxiv 2607.05250 v1 pith:S4QF7OEL submitted 2026-07-06 cs.CL

Streaming Neural Speech Codecs through Time-Invariant Representations

K\'elian Est\`eve , Salima Mhdaffar , Mickael Rouvier , Richard Dufour , Yannick Est\`eve This is my paper
classification cs.CL
keywords speechrepresentationsinformationneuralstreamingtime-invarianttirecodecs
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This paper investigates whether a neural speech codec that separates time-invariant information (speaker identity, acoustic environment) from time-varying content (phonetics, prosody) can operate in a streaming mode suitable for real-time speech generation. The codec, TiCodec, uses a module called TIRE to extract a compact global representation alongside frame-level tokens. The authors probe what TIRE actually captures, find that intermediate encoder layers encode complementary speaker- and environment-related information with little linguistic content, and propose Dual-TIRE, which connects two TIRE modules to different encoder layers. They then test TiCodec in a streaming configuration where audio is processed as successive 660ms blocks with no overlap or smoothing, and report that reconstruction quality degrades only marginally compared to offline processing of the full utterance. The broader claim is that factorized representations, where slowly varying global attributes are pulled out of the frame-level token stream, are a practical path toward low-latency codec-based speech generation.

Core claim

The central finding is that TiCodec's time-invariant representation, extracted by the TIRE module, primarily encodes acoustic-scene and paralinguistic information while retaining little linguistic content, and that different encoder layers capture complementary aspects of this invariant information. Exploiting this complementarity through Dual-TIRE (two TIRE modules at layers 2 and 3) improves speaker similarity on out-of-domain data from an average of 0.681 to 0.701. Separately, the codec can process audio as contiguous 660ms blocks in a streaming fashion with near-identical reconstruction quality to offline processing (MOS 0.618 vs. 0.621), because the invariant representation provides a稳定

What carries the argument

The TIRE (Time-Invariant Representation Extraction) module extracts a compact global representation from an encoder layer, which is then quantized and replicated along the time axis to condition the decoder. Dual-TIRE extends this by connecting two independent TIRE modules to encoder layers 2 and 3, each quantized separately and reinjected at the corresponding decoder layer. The consistency loss during training pushes TIRE to produce similar representations for two segments from the same source. The streaming mechanism simply processes contiguous 660ms blocks independently and concatenates outputs.

Load-bearing premise

The streaming evaluation assumes that processing contiguous 660ms blocks with no overlap or smoothing is a sufficient test of streaming capability, but since the model was trained on 660ms segments, this test may simply confirm that inference matches the training distribution rather than demonstrating generalization to real streaming conditions with shorter or variable block sizes.

What would settle it

If TIRE representations were shown to encode significant linguistic content (e.g., high accuracy on keyword spotting or language identification), the core premise that factorization cleanly separates invariant from time-varying information would be undermined. Additionally, if streaming inference at block lengths shorter than the training segment produced substantial quality degradation, the streaming capability claim would be weakened.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • If factorized codecs can stream at 660ms latency, codec-based speech generation systems (where a language model predicts codec tokens) could operate in real-time conversational settings without buffering entire utterances.
  • Separating invariant from time-varying information could reduce the number of tokens a language model must predict per second, since global attributes need not be regenerated at every frame.
  • The probing methodology (frozen representations fed to diagnostic classifiers across five tasks) provides a template for auditing what information other neural codec intermediate representations actually encode.
  • Layer-dependent training strategies, where different TIRE branches use different segment sampling policies, suggest that multi-level invariant extraction can be tuned for specific attributes like speaker similarity.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The streaming test may be too easy: the model was trained on 660ms segments, so block-by-block inference at the same length may simply match the training distribution rather than demonstrate robust generalization to streaming conditions. A stronger test would use block lengths shorter than the training segment or evaluate on continuous audio with varying durations.
  • The near-identical streaming and offline metrics (MOS 0.618 vs. 0.621) could indicate that the offline mode gains little from full-utterance context, which would mean either the invariant representation is already sufficient or that the model is not exploiting long-range dependencies.
  • The absence of overlap or smoothing at block boundaries means any artifacts would appear as discontinuities; the fact that metrics hold suggests the invariant representation provides enough global conditioning to maintain coherence across blocks, but perceptual evaluation of boundary artifacts is not reported.
  • Dual-TIRE's improvement on speaker similarity but degradation on PESQ across all out-of-domain datasets suggests the two TIRE branches may be competing for representational capacity, and a more principled fusion mechanism (e.g., attention-weighted combination) might resolve the trade-off.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. This paper investigates the time-invariant representations (TIRE) of TiCodec through probing tasks, proposes a Dual-TIRE architecture that extracts invariant representations from two encoder layers, and evaluates TiCodec in a streaming inference setting using 660ms processing blocks. The probing analysis shows that TIRE captures acoustic and paralinguistic information but little linguistic content. Dual-TIRE is shown to improve speaker similarity on out-of-domain data. The streaming evaluation claims that block-by-block decoding does not significantly degrade reconstruction quality.

Significance. The probing analysis of TIRE representations across multiple encoder layers and information types is a useful diagnostic contribution. The Dual-TIRE architecture is a reasonable, low-parameter-cost extension. The exploration of segment selection strategies for TIRE training is a novel and practical investigation. However, the streaming evaluation, which is the paper's titular contribution, has verifiability issues that undermine its significance (see major comments).

major comments (3)
  1. Table 5 (streaming vs. offline, LibriTTS validation) reports metric values that are inconsistent with Table 1 (LibriTTS test-clean) for the same model family. Specifically, PESQ in Table 1 ranges 2.75–2.96, while Table 5 reports PESQ 0.728–0.767 — a ~4× discrepancy. STOI moves in the opposite direction (0.94 in Table 1 vs. 0.99 in Table 5). Additionally, MOS of 0.618 on a conventional 1–5 scale would indicate essentially unusable quality, and SI-SDR of 0.75 dB is near the silence baseline (0 dB). The paper provides no explanation for these discrepancies. If the Table 5 metrics are computed with a different PESQ variant, normalized to [0,1], or otherwise rescaled, the central streaming claim ('no significant degradation') is not verifiable because both offline and streaming modes could score equally poorly on a miscomputed metric. The authors must clarify how each metric in Table 5 was计算,
  2. Table 4, EMILIA-DE row (TIRE system): the values ViSQOL=4.382, PESQ=2.956, STOI=0.944 are identical to Table 1, Layer 2 row (ViSQOL=4.382, PESQ=2.956, STOI=0.944). This exact match across different datasets (LibriTTS test-clean vs. EMILIA-DE) is implausible and suggests a data-entry or copy-paste error. Since Table 4 supports the claim that Dual-TIRE improves out-of-domain generalization (average Sim 0.701 vs. 0.681), any incorrect row affects the averaged results and the cross-corpus comparison. The authors should verify all values in Table 4.
  3. Section 5: The streaming evaluation uses 660ms blocks, which is the same segment length used during training. The paper acknowledges this ('Each block lasts 660ms, the same duration as during the training process') but does not discuss whether this constitutes a meaningful test of streaming generalization. Since the model was trained on 660ms segments, block-by-block inference at the same length may simply confirm that inference matches training conditions. The streaming claim would be substantially strengthened by testing with different block sizes, overlapping windows, or smoothing at block boundaries, and by comparing against at least one other codec in streaming mode.
minor comments (6)
  1. Table 3: The paper states Dual-TIRE 'improves speech reconstruction quality and speaker similarity' (abstract), but Table 3 shows PESQ decreases for both Dual-TIRE variants compared to single-TIRE (2.931 and 2.923 vs. 2.956). The abstract should be revised to reflect the trade-off rather than a uniform improvement.
  2. Section 4.1: No statistical significance tests are reported for any comparison. Given that many differences in Tables 1–4 are small (e.g., ViSQOL 4.382 vs. 4.387), confidence intervals or significance tests would strengthen the claims.
  3. Table 4: MCD values for EMILIA subsets (1.5–1.9) are much higher than for VCTK (0.62–0.63) and LibriTTS (0.73). A brief discussion of why MCD differs so dramatically across corpora would help interpretation.
  4. Figure 3: The y-axis scale and task labels are difficult to read. Consider enlarging or using a table format for the probing results.
  5. Section 2.2: The choice of layers 2 and 3 for Dual-TIRE is motivated by Table 1, but the paper does not report results for other layer combinations (e.g., layers 1+2, 1+3, 3+4). A brief justification for not testing other combinations would be helpful.
  6. The paper does not report bitrate information for the codec configurations evaluated, which is relevant for comparing reconstruction quality across systems.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for a careful and constructive report. The referee raises three major comments concerning (1) apparent metric inconsistencies between Tables 1 and 5, (2) a suspected copy-paste error in Table 4, and (3) the validity of the streaming evaluation given that the block size matches the training segment length. We address each below. Two of the three comments identify genuine errors that require correction; the third motivates additional experiments that we will incorporate.

read point-by-point responses

  1. Referee: Table 5 (streaming vs. offline, LibriTTS validation) reports metric values inconsistent with Table 1 (LibriTTS test-clean). PESQ in Table 1 ranges 2.75–2.96, while Table 5 reports PESQ 0.728–0.767. STOI moves in the opposite direction (0.94 in Table 1 vs. 0.99 in Table 5). MOS of 0.618 on a 1–5 scale would indicate unusable quality, and SI-SDR of 0.75 dB is near silence baseline. The authors must clarify how each metric in Table 5 was computed.

    Authors: The referee is correct that the metrics in Table 5 are not directly comparable to those in Table 1, and we acknowledge that the manuscript fails to explain this. After reviewing our experimental code, we can confirm the following: (1) PESQ in Table 5 was computed using the wide-band PESQ implementation from a different library than the one used for Table 1, and the raw scores were normalized to [0,1] — this was not stated in the paper and we will correct it. (2) STOI values near 0.99 in Table 5 reflect the fact that STOI was computed on a per-block basis on 660ms segments, where high values are expected for clean reconstructed speech at short durations; this differs from the utterance-level STOI in Table 1. (3) The MOS values were not obtained from human listeners but are NISQA-style predicted MOS scores; the label 'Mean Opinion Score' without this qualification is misleading and will be corrected. (4) SI-SDR values near 0 dB are unexpectedly low and we are re-examining the computation pipeline for a possible implementation issue. We agree that without clarification of these methodological differences, the streaming claim is not verifiable. In the revision, we will: (a) recompute all Table 5 metrics using the same metric implementations and scales as Table 1 so that direct comparison is possible, (b) clearly state the metric implementations used, and (c) report both offline and streaming results on the same dataset partition with consistent computation. If any metric cannot be recomputed in a compatible manner, we will remove it rather than present an unverifiable comparison. revision: yes

  2. Referee: Table 4, EMILIA-DE row (TIRE system): ViSQOL=4.382, PESQ=2.956, STOI=0.944 are identical to Table 1, Layer 2 row. This exact match across different datasets is implausible and suggests a copy-paste error. The authors should verify all values in Table 4.

    Authors: The referee is correct. The EMILIA-DE row in the TIRE section of Table 4 contains values that are identical to the Layer 2 row in Table 1, which is indeed implausible across different datasets. This is a data-entry error: the LibriTTS test-clean results were inadvertently copied into the EMILIA-DE row during table preparation. We have re-examined our raw results files and confirmed that the correct EMILIA-DE values for the TIRE system differ from what is reported. We will correct the EMILIA-DE row with the actual values and will re-verify every entry in Table 4 against our experiment logs. We will also recompute the TIRE average row and confirm that the qualitative conclusion (Dual-TIRE improves average speaker similarity) still holds with the corrected values. If the correction changes any reported average or conclusion, we will update the text accordingly. revision: yes

  3. Referee: The streaming evaluation uses 660ms blocks, the same segment length used during training. The paper acknowledges this but does not discuss whether this constitutes a meaningful test of streaming generalization. The streaming claim would be strengthened by testing with different block sizes, overlapping windows, or smoothing at block boundaries, and by comparing against at least one other codec in streaming mode.

    Authors: We agree that testing only at the training block size is a limitation of the current evaluation. The referee's suggestion is well-taken: demonstrating robustness across different block sizes, particularly sizes not seen during training, would substantially strengthen the streaming claim. In the revision, we will add experiments with at least two additional block sizes (e.g., 330ms and 990ms) to test whether block-by-block decoding degrades when the block size differs from training. We will also add an overlapping-window configuration with simple overlap-add smoothing at block boundaries and report the resulting metrics. Regarding comparison with another codec in streaming mode, we agree this would be informative; we will include EnCodec evaluated under the same block-by-block streaming protocol as a baseline, subject to computational feasibility within the revision period. If a full comparison is not feasible, we will at minimum report EnCodec offline results on the same data to contextualize the absolute quality level of our system. revision: partial

Circularity Check

0 steps flagged

No circularity found: derivation is self-contained against external benchmarks

full rationale

The paper presents empirical contributions built on TiCodec [9], which is cited as prior work by different authors (Ren et al., ICASSP 2024). The TIRE module, Dual-TIRE architecture, and streaming evaluation are independent extensions. The probing tasks use external datasets (VoxCeleb, TAU, MELD, Common Voice, Speech Commands) and the evaluation corpora include out-of-domain data (VCTK, EMILIA). No step in the derivation chain reduces to its own inputs by construction. The probing classifiers are trained on frozen TIRE representations against external labels, the segment selection strategies are ablated against held-out test sets, and the streaming comparison (Table 5) measures offline vs. streaming on the same model without fitting parameters to the streaming condition. The metric inconsistencies between Tables 1 and 5 (PESQ 2.9 vs 0.73, MOS 0.618) are a correctness risk, not a circularity issue — the streaming claim is not forced by definition or by a self-citation chain. No self-citation load-bearing pattern is present: the central premise rests on externally verifiable experiments, not on a uniqueness theorem or ansatz from the authors' own prior work.

Axiom & Free-Parameter Ledger

2 free parameters · 3 axioms · 0 invented entities

The paper introduces no new postulated entities. The TIRE module and TiCodec architecture are from prior work [9]. Dual-TIRE is an architectural extension using existing components. The free parameters are empirical choices (layer indices, segment strategies) selected via ablation rather than fitted constants.

free parameters (2)
  • TIRE connection layer index = Layers 2 and 3 for Dual-TIRE
    Selected empirically based on Table 1 results showing best trade-off at intermediate layers.
  • Segment selection strategy per layer = Layer 2: random unconstrained; Layer 3: cross-file same speaker same book
    Selected based on Table 2 results; the cross-file strategy for Layer 3 is chosen to improve speaker similarity.
axioms (3)
  • domain assumption Time-invariant speech information (speaker, environment) can be separated from time-varying content (phonetics) at the encoder level.
    Section 1.2 and Section 2.1: the entire TiCodec framework assumes this factorization is meaningful and achievable.
  • standard math Probing task performance indicates the presence of information in representations.
    Section 3: the probing approach assumes that classifier accuracy on frozen features reflects information content.
  • ad hoc to paper 660ms blocks are sufficient context for estimating time-invariant representations in streaming mode.
    Section 5: the streaming evaluation assumes that 660ms blocks (matching training segment length) provide adequate context for invariant representation extraction without explicit validation of this assumption.

pith-pipeline@v1.1.0-glm · 14342 in / 1913 out tokens · 118662 ms · 2026-07-07T21:37:17.616115+00:00 · methodology

0 comments
read the original abstract

Neural speech codecs are increasingly used as intermediate representations in codec-based speech generation systems. TiCodec introduces a factorized representation that separates time-varying speech content from time-invariant information through a Time-Invariant Representation Extraction (TIRE) module, potentially reducing the amount of information that must be modeled at the frame-level. In this work, we investigate the nature of the information captured by TIRE representations and their suitability for low-latency speech processing. Using a series of probing tasks, we analyze the influence of the encoder layer and show that intermediate layers capture complementary speaker- and environment-related information while containing little linguistic content. We further study several segment selection strategies for TIRE training and demonstrate that cross-file sampling improves the robustness of invariant representations. Based on these findings, we propose Dual-TIRE, a multi-level architecture that exploits the complementarity of different encoder layers and improves speech reconstruction quality and speaker similarity. Finally, we evaluate TiCodec in a streaming inference setting using successive 660ms processing blocks. Results show that streaming operation can be achieved without significant degradation in reconstruction performance, highlighting the potential of factorized neural codec representations for future low-latency speech generation systems.

Figures

Figures reproduced from arXiv: 2607.05250 by K\'elian Est\`eve, Mickael Rouvier, Richard Dufour, Salima Mhdaffar, Yannick Est\`eve.

Figure 1
Figure 1. Figure 1: TiCodec architecture, illustrating the factorization of speech into time￾dependent representations encoded via RVQ and time-invariant representations ex￾tracted by the TIRE module [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Dual-TIRE architecture. Two TIRE modules extract time-invariant represen￾tations from encoder layers 2 and 3, which are quantized independently and reinjected into the corresponding decoder layers, while frame-level content is encoded via RVQ. This dual extraction and injection mechanism is designed to enrich the global conditioning available to the decoder by combining invariant representations ex￾tracted… view at source ↗
Figure 3
Figure 3. Figure 3: Probing evaluation of TIRE representations extracted from encoder Layers 1–4 across five downstream classification tasks, highlighting the types of information cap￾tured at each layer and the progressive loss of informative content in deeper compressed representations [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

23 extracted references · 23 canonical work pages · 5 internal anchors

  1. [1]

    Sound- stream: An end-to-end neural audio codec,

    N. Zeghidour, A. Luebs, A. Omran, J. Skoglund, and M. Tagliasacchi, “Sound- stream: An end-to-end neural audio codec,”IEEE/ACM Transactions on Audio, Speech, and Language Processing, vol. 30, pp. 495–507, 2021

    work page 2021

  2. [2]

    High Fidelity Neural Audio Compression

    A. D´ efossez, J. Copet, G. Synnaeve, and Y. Adi, “High fidelity neural audio compression,” 2022. [Online]. Available: https://arxiv.org/abs/2210.13438

    work page internal anchor Pith review Pith/arXiv arXiv 2022

  3. [3]

    High-fidelity audio compression with improved rvqgan,

    R. Kumar, P. Seetharaman, A. Luebs, I. Kumar, and K. Kumar, “High-fidelity audio compression with improved rvqgan,”Advances in Neural Information Pro- cessing Systems, vol. 36, pp. 27 980–27 993, 2023

    work page 2023

  4. [4]

    Audiolm: a language modeling approach to audio generation,

    Z. Borsos, R. Marinier, D. Vincent, E. Kharitonov, O. Pietquin, M. Sharifi, D. Rob- lek, O. Teboul, D. Grangier, M. Tagliasacchiet al., “Audiolm: a language modeling approach to audio generation,”IEEE/ACM transactions on audio, speech, and lan- guage processing, vol. 31, pp. 2523–2533, 2023

    work page 2023

  5. [5]

    Text-free prosody-aware generative spoken language modeling,

    E. Kharitonov, A. Lee, A. Polyak, Y. Adi, J. Copet, K. Lakhotia, T.-A. Nguyen, M. Riviere, A. Mohamed, E. Dupouxet al., “Text-free prosody-aware generative spoken language modeling,” inACL, 2022. Streaming Neural Speech Codecs through Time-Invariant Representations 15

    work page 2022

  6. [6]

    Neural Codec Language Models are Zero-Shot Text to Speech Synthesizers

    C. Wang, S. Chen, Y. Wu, Z. Zhang, L. Zhou, S. Liu, Z. Chen, Y. Liu, H. Wang, J. Liet al., “Neural codec language models are zero-shot text to speech synthesiz- ers,”arXiv preprint arXiv:2301.02111, 2023

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  7. [7]

    Voicebox: Text-guided multilingual universal speech generation at scale,

    M. Le, A. Vyas, B. Shi, B. Karrer, L. Sari, R. Moritz, M. Williamson, V. Manohar, Y. Adi, J. Mahadeokaret al., “Voicebox: Text-guided multilingual universal speech generation at scale,”Advances in neural information processing systems, vol. 36, pp. 14 005–14 034, 2023

    work page 2023

  8. [8]

    Naturalspeech 3: zero-shot speech synthesis with factorized codec and diffusion models,

    Z. Ju, Y. Wang, K. Shen, X. Tan, D. Xin, D. Yang, Y. Liu, Y. Leng, K. Song, S. Tanget al., “Naturalspeech 3: zero-shot speech synthesis with factorized codec and diffusion models,” inProceedings of the 41st International Conference on Ma- chine Learning, 2024, pp. 22 605–22 623

    work page 2024

  9. [9]

    Fewer-token neural speech codec with time-invariant codes,

    Y. Ren, T. Wang, J. Yi, L. Xu, J. Tao, C. Y. Zhang, and J. Zhou, “Fewer-token neural speech codec with time-invariant codes,” inICASSP. IEEE, 2024, pp. 12 737–12 741

    work page 2024

  10. [10]

    Voxceleb: Large-scale speaker verification in the wild,

    A. Nagrani, J. S. Chung, W. Xie, and A. Zisserman, “Voxceleb: Large-scale speaker verification in the wild,”Computer Speech & Language, vol. 60, p. 101027, 2020

    work page 2020

  11. [11]

    A multi-device dataset for urban acoustic scene classification

    A. Mesaros, T. Heittola, and T. Virtanen, “A multi-device dataset for urban acous- tic scene classification,”arXiv preprint arXiv:1807.09840, 2018

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  12. [12]

    Meld: A multimodal multi-party dataset for emotion recognition in conversations,

    S. Poria, D. Hazarika, N. Majumder, G. Naik, E. Cambria, and R. Mihalcea, “Meld: A multimodal multi-party dataset for emotion recognition in conversations,” in ACL, 2019

    work page 2019

  13. [13]

    Common voice: A massively-multilingual speech corpus,

    R. Ardila, M. Branson, K. Davis, M. Kohler, J. Meyer, M. Henretty, R. Morais, L. Saunders, F. Tyers, and G. Weber, “Common voice: A massively-multilingual speech corpus,” inProceedings of the twelfth language resources and evaluation conference, 2020, pp. 4218–4222

    work page 2020

  14. [14]

    Speech Commands: A Dataset for Limited-Vocabulary Speech Recognition

    P. Warden, “Speech commands: A dataset for limited-vocabulary speech recogni- tion,”arXiv preprint arXiv:1804.03209, 2018

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  15. [15]

    Libritts: A corpus derived from librispeech for text-to-speech,

    H. Zen, V. Dang, R. Clark, Y. Zhang, R. J. Weiss, Y. Jia, Z. Chen, and Y. Wu, “Libritts: A corpus derived from librispeech for text-to-speech,”Interspeech, 2019

    work page 2019

  16. [16]

    Librispeech: an asr corpus based on public domain audio books,

    V. Panayotov, G. Chen, D. Povey, and S. Khudanpur, “Librispeech: an asr corpus based on public domain audio books,” in2015 IEEE international conference on acoustics, speech and signal processing (ICASSP). IEEE, 2015, pp. 5206–5210

    work page 2015

  17. [17]

    Cstr vctk corpus: English multi- speaker corpus for cstr voice cloning toolkit,

    C. Veaux, J. Yamagishi, K. MacDonaldet al., “Cstr vctk corpus: English multi- speaker corpus for cstr voice cloning toolkit,”University of Edinburgh. The Centre for Speech Technology Research (CSTR), vol. 6, p. 15, 2017

    work page 2017

  18. [18]

    Emilia: An extensive, multilingual, and diverse speech dataset for large- scale speech generation,

    H. He, Z. Shang, C. Wang, X. Li, Y. Gu, H. Hua, L. Liu, C. Yang, J. Li, P. Shi et al., “Emilia: An extensive, multilingual, and diverse speech dataset for large- scale speech generation,” inSpoken Language Technology Workshop (SLT). IEEE, 2024, pp. 885–890

    work page 2024

  19. [19]

    Visqol v3: An open source production ready objective speech and audio metric,

    M. Chinen, F. S. C. Lim, J. Skoglund, N. Gureev, F. O’Gorman, and A. Hines, “Visqol v3: An open source production ready objective speech and audio metric,”

    work page

  20. [20]

    ViSQOL v3: An Open Source Production Ready Objective Speech and Audio Metric

    [Online]. Available: https://arxiv.org/abs/2004.09584

    work page internal anchor Pith review Pith/arXiv arXiv 2004

  21. [21]

    Perceptual evaluation of speech quality (pesq)-a new method for speech quality assessment of telephone networks and codecs,

    A. Rix, J. Beerends, M. Hollier, and A. Hekstra, “Perceptual evaluation of speech quality (pesq)-a new method for speech quality assessment of telephone networks and codecs,” inICASSP, 2001

    work page 2001

  22. [22]

    An algorithm for intel- ligibility prediction of time–frequency weighted noisy speech,

    C. H. Taal, R. C. Hendriks, R. Heusdens, and J. Jensen, “An algorithm for intel- ligibility prediction of time–frequency weighted noisy speech,”IEEE Transactions on Audio, Speech, and Language Processing, vol. 19, no. 7, pp. 2125–2136, 2011

    work page 2011

  23. [23]

    Mel-cepstral distance measure for objective speech quality assess- ment,

    R. Kubichek, “Mel-cepstral distance measure for objective speech quality assess- ment,” inProceedings of IEEE Pacific Rim Conference on Communications Com- puters and Signal Processing, vol. 1, 1993, pp. 125–128 vol.1

    work page 1993