Pith. sign in

REVIEW 4 major objections 6 minor 32 references

A lightweight and robust method for blind wideband-to-fullband extension of speech

T0 review · 4 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read A blind bandwidth-extension model with roughly 370K parameters significantly improves Opus 1.5 wideband speech at 6, 9, and 12 kb/s, and at 9 kb/s it statistically ties EVS at 9.6 kb/s and Opus 1.4 at 18 kb/s.

desk verdict Solid engineering paper: a genuinely light blind BWE with credible P.808 gains, but the headline 'meets guided BWE' needs a properly reported listening test and an equivalence analysis. read the letter →

arxiv 2412.11392 v4 pith:4YBWSGD7 submitted 2024-12-16 eess.AS cs.SD

Jan Büthe , Jean-Marc Valin This is my paper · ORCID
classification eess.AScs.SD
keywords blindbandwidthextensionwideband-to-fullbandlow-complexityspeechcodingOpuscodecadversarialtrainingP.808listeningtestspectralfoldingenhancement
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper tries to establish that blind bandwidth extension—reconstructing missing high frequencies without any side information from the encoder—can be light enough for mobile devices and good enough to rival guided bandwidth extension. It presents a hybrid model with roughly 370K parameters and about 140 MFLOPS that combines classical DSP operations with a small neural network controlling time-varying filters and shaping. When placed after the Opus 1.5 codec, the model significantly improves perceived quality at 6, 9, and 12 kb/s in a P.808 DCR listening test, and at 9 kb/s it is statistically tied with EVS at 9.6 kb/s and Opus 1.4 at 18 kb/s. The payoff, if true, is that low-bitrate speech calls can get fullband quality through a backward-compatible decoder-side enhancement.

What carries the argument

The load-bearing object is a hybrid signal path made of classical, interpretable DSP components controlled by a small recurrent feature encoder. The wideband signal passes through adaptive convolutional filters (AdaConv) for pre- and post-filtering and adaptive temporal shaping (AdaShape) that multiplies the signal by a learned periodic weight sequence—spectral folding in the broad sense—while a nonlinearity $f(x) = x \sin(\log |x|)$ generates harmonic extension for voiced speech. A 72-dimensional feature vector computed from an ERB-scale log-magnitude spectrogram and complex phase differences, processed by a GRU at 200 Hz, adapts the time-varying filters and weights. The whole signal path has a lookahead of only 0.27 ms because it sits on the libopus two-stage 16-to-48 kHz upsampler, and training combines regression losses with a frequency-domain adversarial discriminator.

What would settle it

Run the same P.808 DCR protocol on speech recorded with phone microphones or mixed with background noise at 9 kb/s Opus; if the bandwidth-extended condition no longer significantly beats the wideband baseline, or if EVS at 9.6 kb/s clearly outscores it, the central claim of robustness and guided-quality equivalence would be contradicted.

Watch

Extended reading notes

Core claim

The central claim is that a blind bandwidth-extension system built from classical DSP building blocks and a small learned controller can meet the quality of codecs that use guided, side-information bandwidth extension. Specifically, the paper reports that adding the proposed model to Opus 1.5 wideband speech gives significant quality improvements at 6, 9, and 12 kb/s on unseen EARS test material, and that the 9 kb/s condition is statistically tied with 3GPP EVS at 9.6 kb/s and with Opus 1.4 at 18 kb/s. The paper further claims this is achieved with a model small enough and low-delay enough to ride on common wideband speech codecs, and that the model learns a division of labor: spectral folding extends unvoiced parts while a nonlinearity extends voiced parts.

Load-bearing premise

The evaluation assumes that a P.808 DCR listening test on clean anechoic EARS speech processed by Opus 1.5 at decoder complexity 10 represents real deployment; the paper offers no evidence on noisy speech, other codecs, or device microphones, and the EVS tie is one indirect comparison rather than a general proof.

Editorial extensions

If this is right

  • Adding the proposed bandwidth extension to Opus 1.5 significantly raises perceived quality at every tested bitrate, so low-bitrate Opus calls could be upgraded without changing the encoder or bitstream.
  • At 9 kb/s, blind bandwidth extension plus Opus matches the quality of EVS at 9.6 kb/s, suggesting side-information bandwidth extension is not necessary to reach that quality in this configuration.
  • The same configuration matches Opus 1.4 at 18 kb/s, implying roughly a 45–50% bitrate reduction for equivalent quality in the comparison.
  • Because the model is decoder-side and delay-light, it can be paired with existing wideband codecs rather than replacing them, preserving compatibility.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If this result transfers to noisy speech and device microphone captures, decoder-side blind bandwidth extension could become a universal drop-in upgrade for legacy wideband telephony, where no encoder cooperation is possible.
  • The observed split—folding for unvoiced speech and nonlinearity for voiced speech—suggests a classical design rule: switch between the two extension mechanisms according to voicing, which could be tested with a rule-based version of the same signal path.
  • The statistical tie with EVS rests on a single listening test on clean anechoic EARS speech; a head-to-head P.808 test across bitrates, codecs, and noise conditions would be the natural next experiment to see whether the tie holds.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 6 minor

Summary. The paper proposes BBWENet, a blind wideband-to-fullband bandwidth-extension model with ~370K parameters and a claimed complexity of ~140 MFLOPS (~70 MMACS), built around a DSP-inspired structure: upsampling, a nonlinearity, and an AdaShape-based spectral-folding path, all steered by a small feature encoder with a GRU. The model is trained on TTS data with augmentations and a combination of regression and adversarial losses. In a P.808 DCR listening test on the EARS dataset, the authors report that adding the model to Opus 1.5 significantly improves quality at 6, 9, and 12 kb/s, and that at 9 kb/s Opus 1.5+BWE is statistically tied with EVS at 9.6 kb/s and Opus 1.4 at 18 kb/s, which they interpret as showing that blind BWE can meet the quality of guided BWE. The paper also presents a model-inspection analysis showing that the model uses folding for unvoiced speech and nonlinear extension for voiced speech.

Significance. The main contribution is a low-delay, low-complexity blind BWE with a transparent DSP-inspired architecture and a public C/Python implementation. If the listening-test results are fully reported and the complexity figure is documented, the paper would provide a useful backward-compatible quality improvement for existing wideband codecs and a solid reference point for blind BWE. Strengths include evaluation on a held-out dataset (EARS), use of a standardized P.808 DCR methodology, a model-inspection analysis that gives insight into the dual extension mechanism, and release of code. The significance is conditional on the statistical support for the 'meets guided BWE' claim and on a reproducible complexity estimate, both of which need strengthening.

major comments (4)
  1. [§4.1, Fig. 2] The central claim that Opus 1.5 at 9 kb/s with the proposed BWE 'meets the quality' of EVS at 9.6 kb/s and Opus 1.4 at 18 kb/s rests entirely on the statement that these conditions were 'statistically tied,' but the paper reports no number of listeners, no number of test items, no test statistic, no p-values, and no confidence intervals. In addition, the Figure 2 caption states 'p = 0.95,' which if read literally indicates a non-significant result and is at odds with the claim of significance. A non-significant difference is not evidence of equality unless the analysis is framed as an equivalence test with a pre-specified margin or a power analysis is provided. Please report the complete statistical details and either perform an equivalence/non-inferiority analysis or qualify the 'meets quality' conclusion accordingly.
  2. [§1 and Abstract] The complexity figure of ~140 MFLOPS (~70 MMACS) is asserted without any methodology or measurement description. It is not stated which operations are counted (feature encoder, GRU, AdaConv/AdaShape, upsampling filters, signal-path filtering), whether the number refers to the C implementation or to a framework, or how it relates to the 10 ms frame rate. Because 'lightweight' is a central contribution, please provide a reproducible complexity count (e.g., per-second MACs/FLOPs with a clear counting convention, or measured runtime on a specified platform) and use the same convention when comparing with [11] and [14].
  3. [§4.1] The title and introduction claim robustness to real-world variability, but the subjective evaluation is limited to clean anechoic EARS speech passed through a single codec configuration (Opus 1.5, decoder complexity 10). No condition with additive noise, reverberation, device microphones, or another codec is reported. The augmentation list in §3.1 shows the intended robustness, but it does not verify it. Please add at least objective evaluations on noisy/reverberant conditions (e.g., PESQ, DNSMOS, or a small subjective test) or revise the robustness claims and title to reflect the actual test scope.
  4. [§4.1 (EVS comparison)] The comparison to EVS is indirect: EVS at 9.6 kb/s is a superwideband codec (coding up to 14.4 kHz) and the tie with Opus 1.4 at 18 kb/s is also a single operating point. The conclusion that 'blind bandwidth extension can meet the quality of classical guided bandwidth extensions' therefore goes beyond the evidence, which supports only a statement about these specific codecs and bitrates. Please qualify the abstract and conclusion accordingly.
minor comments (6)
  1. [§4.1] Please report the actual number of EARS speakers and sentences used in the listening test; 'three random sentence pairs per speaker' is not sufficient to infer the test size.
  2. [Fig. 2] The figure would be much more informative with error bars or confidence intervals; currently only condition means are shown.
  3. [References] Reference [19] is cited for the claim about the critical impact of microphone channels, but [19] is Huber's Robust Statistics, which appears unrelated; a relevant reference on microphone variability in speech processing should be supplied.
  4. [Global] There are several typos: 'bandwith' (Introduction), 'agressively' (Section 2), 'covolution' (Section 3.2), and 'F our' in Reference [8].
  5. [§4.2] The statement that the second NonLin module 'could likely be omitted' is based only on informal listening; either report the informal test more concretely or mark it clearly as a hypothesis.
  6. [Fig. 1] The repeated labels '2↑ 2↑ 2↑' and '1.5↑ 1.5↑ 1.5↑' are not explained in the caption; please clarify their meaning.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the central claim is an empirical listening-test measurement, and the model's self-cited components are not used to derive the quality result.

full rationale

No circular step is present. The paper's central claim—that the proposed BWE model significantly improves perceived quality of Opus 1.5 at 6–12 kb/s and statistically ties EVS at 9.6 kb/s and Opus 1.4 at 18 kb/s—is an empirical measurement from an external P.808 DCR listening test on the EARS corpus, which was not used in training. The training losses (envelope matching, spectral fine structure, lowband L2, and adversarial) optimize signal fidelity and plausibility, not the DMOS outcome, so the evaluation is not forced by construction. The self-citations to AdaConv [16], AdaShape/NOLACE [17], and pitch features [18] are reuse of previously published components and are not invoked as evidence for the quality claim; the quality claim rests on the listening test and external anchors (EVS, EnCodec, Opus 1.4). The model inspection in §4.2 is a post-hoc decomposition of module contributions, not a derivation of the headline result. The absence of reported confidence intervals for the 'statistically tied' comparisons is a statistical-reporting concern, not a circularity.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The central claim is empirical; the model's parameters are fitted to TTS data, but the paper also introduces several hand-chosen design choices (nonlinearity, loss weights, augmentation rates) that are not independently justified and could affect the results. No new physical entities are introduced.

free parameters (5)
  • Nonlinearity f(x)=x sin(log|x|)
    Hand-designed in Eq. (1) to preserve scale and induce similar distortion regardless of input scale; not fitted.
  • Regression loss weights (Lenv, Lspec, Ltdlp) = 1/13, 2/13, 10/13
    Weights in Eq. (4) are set by hand; no ablation is reported.
  • Data augmentation rates (EQ, noise, RIR, DC offset) = 40%, 20%, 20%, 10%
    Section 3.1; chosen to increase robustness, rates and order are arbitrary.
  • Training lowpass cutoff range = 7.5-8 kHz with varying slope
    Section 3.1; defines the simulated wideband input, affecting the gap the model must fill.
  • Adversarial training regularization weight = Lreg = 0.6 Lpre
    Section 3.2; hand-chosen to balance regression and adversarial losses.
assumptions (4)
  • domain assumption Highband (8-20 kHz) speech can be reconstructed from wideband (0-8 kHz) features with sufficient plausibility for perceptual quality.
    The whole BWE task relies on this ill-posed inverse problem; the model is trained to produce plausible, not exact, highbands.
  • domain assumption The 72-dimensional feature set (32 ERB log-magnitudes plus 40 complex phase differences) is sufficient for the model to infer highband structure.
    Section 2 states these features provide spectral envelope, pitch, and voicing cues; no ablation verifies each component is necessary.
  • standard math The libopus 16-to-48 kHz upsampler's IIR/FIR structure introduces only 13 samples of delay and can be approximated by FIR filters in training.
    Section 2; the delay compensation is used in training and the 0.27 ms lookahead claim depends on it.
  • domain assumption P.808 DCR listening test results provide a valid measure of perceived quality for these codec conditions.
    Section 4.1 relies on the open-source P.808 implementation; the paper assumes the standard methodology is appropriate for comparing BWE conditions.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A lightweight and robust method for blind wideband-to-fullband extension of speech." pith.science (2026). https://pith.science/paper/4YBWSGD7

@misc{pith2026241211392,
  author       = {Pith},
  title        = {Pith review of: A lightweight and robust method for blind wideband-to-fullband extension of speech},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4YBWSGD7}},
  note         = {Machine review of arXiv:2412.11392}
}
read the original abstract

Reducing the bandwidth of speech is common practice in resource constrained environments like low-bandwidth speech transmission or low-complexity vocoding. We propose a lightweight and robust method for extending the bandwidth of wideband speech signals that is inspired by classical methods developed in the speech coding context. The resulting model has just ~370K parameters and a complexity of ~140 MFLOPS (or ~70 MMACS). With a frame size of 10 ms and a lookahead of only 0.27 ms, the model is well-suited for use with common wideband speech codecs. We evaluate the model's robustness by pairing it with the Opus SILK speech codec (1.5 release) and verify in a P.808 DCR listening test that it significantly improves quality from 6 to 12 kb/s. We also demonstrate that Opus 1.5 together with the proposed bandwidth extension at 9 kb/s meets the quality of 3GPP EVS at 9.6 kb/s and that of Opus 1.4 at 18 kb/s showing that the blind bandwidth extension can meet the quality of classical guided bandwidth extensions thus providing a way for backward-compatible quality improvement.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

32 extracted references · 26 canonical work pages

  1. [11]

    Low- complexity streaming speech super-resolution,

    E. Soltanmohammadi, P. Smaragdis, and M. Goodwin, “Low- complexity streaming speech super-resolution,” in IEEE 2023 Workshop on Machine Learning for Signal Processing (MLSP) ,

    work page 2023

  2. [14]

    Low-Complexity Real-Time Neural Network for Blind Bandwidth Extension of Wideband Speech,

    E. G ´omez, M. H. Vali, and T. B ¨ackstr¨om, “Low-Complexity Real-Time Neural Network for Blind Bandwidth Extension of Wideband Speech,” in 2023 31st European Signal Processing Conference (EUSIPCO) , 2023, pp. 31–35

    work page 2023

  3. [1]

    G.711 : Pulse code modulation (PCM) of voice frequencies,

    “G.711 : Pulse code modulation (PCM) of voice frequencies,” ITU-T, Tech. Rep

    work page

  4. [2]

    G.722.2 : Wideband coding of speech at around 16 kbit/s using Adaptive Multi-Rate Wideband (AMR-WB),

    “G.722.2 : Wideband coding of speech at around 16 kbit/s using Adaptive Multi-Rate Wideband (AMR-WB),” ITU-T, Tech. Rep

    work page

  5. [3]

    V oice coding with opus,

    K. V os, K. Sørensen, S. Jensen, and J.-M. Valin, “V oice coding with opus,” 135th Audio Engineering Society Convention 2013 , pp. 722–731, 01 2013

    work page 2013

  6. [4]

    LPCNet: Improving Neural Speech Synthesis Through Linear Prediction

    J. Valin and J. Skoglund, “LPCNet: Improving Neural Speech Synthesis Through Linear Prediction,” CoRR, vol. abs/1810.11846, 2018. [Online]. Available: http://arxiv.org/abs/1810.11846

    work page Pith review arXiv 2018

  7. [5]

    High-frequency regeneration in speech coding systems,

    J. Makhoul and M. Berouti, “High-frequency regeneration in speech coding systems,” in ICASSP ’79. IEEE International Conference on Acoustics, Speech, and Signal Processing , vol. 4, 1979, pp. 428–431

    work page 1979

  8. [6]

    Statistical recovery of wideband speech from narrowband speech,

    Y . M. Cheng, D. O’Shaughnessy, and P. Mermelstein, “Statistical recovery of wideband speech from narrowband speech,” IEEE Transactions on Speech and Audio Processing , vol. 2, no. 4, pp. 544–548, 1994

    work page 1994

Show all 32 references
  1. [7]

    Signal restoration of broad band speech using nonlinear processing,

    H. Yasukawa, “Signal restoration of broad band speech using nonlinear processing,” in 1996 8th European Signal Processing Conference (EUSIPCO 1996) , 1996, pp. 1–4

    1996

  2. [8]

    Restoration of wide band signal from telephone speech using linear prediction error processing,

    ——, “Restoration of wide band signal from telephone speech using linear prediction error processing,” in Proceeding of F ourth International Conference on Spoken Language Processing. ICSLP ’96 , vol. 2, 1996, pp. 901–904 vol.2

    1996

  3. [9]

    Super-wideband bandwidth extension for speech in the 3GPP EVS codec,

    V . Atti, V . Krishnan, D. Dewasurendra, V . Chebiyyam, S. Subasingha, D. J. Sinder, V . Rajendran, I. Varga, J. Gibbs, L. Miao, V . Grancharov, and H. Pobloth, “Super-wideband bandwidth extension for speech in the 3GPP EVS codec,” in 2015 IEEE International Conference on Acou...

    2015

  4. [10]

    Adversarial Training for Speech Super-Resolution,

    S. E. Eskimez, K. Koishida, and Z. Duan, “Adversarial Training for Speech Super-Resolution,” IEEE Journal of Selected Topics in Signal Processing, vol. 13, no. 2, pp. 347–358, 2019

    2019

  5. [12]

    Neural V ocoder is All You Need for Speech Super-resolution,

    H. Liu, W. Y . Choi, X. Liu, Q. Kong, Q. Tian, and D. Wang, “Neural V ocoder is All You Need for Speech Super-resolution,” ArXiv, vol. abs/2203.14941, 2022. [Online]. Available: https://api.semanticscholar. org/CorpusID:247778792

    2022 arXiv

  6. [13]

    AERO: Audio Super Resolution in the Spectral Domain,

    M. Mandel, O. Tal, and Y . Adi, “AERO: Audio Super Resolution in the Spectral Domain,” in ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2023, pp. 1–5

    2023

  7. [15]

    TS 26.453: Codec for Enhanced V oice Services (EVS),

    3GPP, “TS 26.453: Codec for Enhanced V oice Services (EVS),” ETSI, Tech. Rep

  8. [16]

    Lace: A Light-Weight, Causal Model for Enhancing Coded Speech Through Adaptive Convolutions,

    J. B ¨uthe, J.-M. Valin, and A. Mustafa, “Lace: A Light-Weight, Causal Model for Enhancing Coded Speech Through Adaptive Convolutions,” in 2023 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics (WASPAA) , 2023, pp. 1–5

    2023

  9. [17]

    NOLACE: Improving Low-Complexity Speech Codec Enhancement Through Adaptive Temporal Shaping,

    J. B ¨uthe, A. Mustafa, J.-M. Valin, K. Helwani, and M. M. Goodwin, “NOLACE: Improving Low-Complexity Speech Codec Enhancement Through Adaptive Temporal Shaping,” in ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2024, pp. 476–480

    2024

  10. [18]

    Noise-Robust DSP-Assisted Neural Pitch Estimation With Very Low Complexity,

    K. Subramani, J.-M. Valin, J. B ¨uthe, P. Smaragdis, and M. Goodwin, “Noise-Robust DSP-Assisted Neural Pitch Estimation With Very Low Complexity,” in ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2024, pp. 11 851– 11 855

    2024

  11. [19]

    P. J. Huber, Robust statistics . John Wiley & Sons, 2004, vol. 523

    2004

  12. [20]

    Open-source Multi-speaker Corpora of the English Accents in the British Isles,

    I. Demirsahin, O. Kjartansson, A. Gutkin, and C. Rivera, “Open-source Multi-speaker Corpora of the English Accents in the British Isles,” in Proceedings of The 12th Language Resources and Evaluation Conference (LREC). Marseille, France: European Language Resources Association ...

    2020

  13. [21]

    Open-Source High Quality Speech Datasets for Basque, Catalan and Galician,

    O. Kjartansson, A. Gutkin, A. Butryna, I. Demirsahin, and C. Rivera, “Open-Source High Quality Speech Datasets for Basque, Catalan and Galician,” in Proceedings of the 1st Joint Workshop on Spoken Language Technologies for Under-resourced languages (SLTU) and Collaboration and...

    2020

  14. [22]

    A Step-by-Step Process for Building TTS V oices Using Open Source Data and Framework for Bangla, Javanese, Khmer, Nepali, Sinhala, and Sundanese,

    K. Sodimana, K. Pipatsrisawat, L. Ha, M. Jansche, O. Kjartansson, P. D. Silva, and S. Sarin, “A Step-by-Step Process for Building TTS V oices Using Open Source Data and Framework for Bangla, Javanese, Khmer, Nepali, Sinhala, and Sundanese,” in Proc. The 6th Intl. Workshop on S...

    2018 doi

  15. [23]

    Crowdsourcing Latin American Spanish for Low-Resource Text-to- Speech,

    A. Guevara-Rukoz, I. Demirsahin, F. He, S.-H. C. Chu, S. Sarin, K. Pipatsrisawat, A. Gutkin, A. Butryna, and O. Kjartansson, “Crowdsourcing Latin American Spanish for Low-Resource Text-to- Speech,” in Proceedings of The 12th Language Resources and Evaluation Conference (LREC)....

    2020

  16. [24]

    Open-source Multi-speaker Speech Corpora for Building Gujarati, Kannada, Malayalam, Marathi, Tamil and Telugu Speech Synthesis Systems,

    F. He, S.-H. C. Chu, O. Kjartansson, C. Rivera, A. Katanova, A. Gutkin, I. Demirsahin, C. Johny, M. Jansche, S. Sarin, and K. Pipatsrisawat, “Open-source Multi-speaker Speech Corpora for Building Gujarati, Kannada, Malayalam, Marathi, Tamil and Telugu Speech Synthesis Systems,...

    2020

  17. [25]

    Burmese Speech Corpus, Finite-State Text Normalization and Pronunciation Grammars with an Application to Text-to-Speech,

    Y . M. Oo, T. Wattanavekin, C. Li, P. De Silva, S. Sarin, K. Pipatsrisawat, M. Jansche, O. Kjartansson, and A. Gutkin, “Burmese Speech Corpus, Finite-State Text Normalization and Pronunciation Grammars with an Application to Text-to-Speech,” in Proceedings of The 12th Language...

    2020

  18. [26]

    Rapid development of TTS corpora for four South African languages,

    D. van Niekerk, C. van Heerden, M. Davel, N. Kleynhans, O. Kjartansson, M. Jansche, and L. Ha, “Rapid development of TTS corpora for four South African languages,” in Proc. Interspeech 2017 , Stockholm, Sweden, Aug. 2017, pp. 2178–2182. [Online]. Available: http://dx.doi.org/1...

    2017 doi

  19. [27]

    Developing an Open-Source Corpus of Yoruba Speech,

    A. Gutkin, I. Demir s ¸ahin, O. Kjartansson, C. Rivera, and K. T ´ub`o.s´un, “Developing an Open-Source Corpus of Yoruba Speech,” in Proceedings of Interspeech 2020 . Shanghai, China: International Speech and Communication Association (ISCA), October 2020, pp. 404–408. [Online...

    2020 doi

  20. [28]

    Hi-Fi Multi- Speaker English TTS Dataset,

    E. Bakhturina, V . Lavrukhin, B. Ginsburg, and Y . Zhang, “Hi-Fi Multi- Speaker English TTS Dataset,” in Proc. Interspeech 2021 , 2021, pp. 2776–2780

    2021

  21. [29]

    An Open source Implementation of ITU-T Recommendation P.808 with Validation,

    B. Naderi and R. Cutler, “An Open source Implementation of ITU-T Recommendation P.808 with Validation,” in Proc. INTERSPEECH, 2020

    2020

  22. [30]

    EARS: An Anechoic Fullband Speech Dataset Benchmarked for Speech Enhancement and Dereverberation,

    J. Richter, Y .-C. Wu, S. Krenn, S. Welker, B. Lay, S. Watanabe, A. Richard, and T. Gerkmann, “EARS: An Anechoic Fullband Speech Dataset Benchmarked for Speech Enhancement and Dereverberation,” in Proc. Interspeech 2024 , 2024, pp. 4874–4877

    2024

  23. [31]

    High fidelity neural audio compression,

    A. D ´efossez, J. Copet, G. Synnaeve, and Y . Adi, “High fidelity neural audio compression,” ArXiv, vol. abs/2210.13438, 2022. [Online]. Available: https://api.semanticscholar.org/CorpusID:253097788

    2022 arXiv

  24. [2023]

    Available: https://www.amazon.science/publications/low- complexity-streaming-speech-super-resolution

    [Online]. Available: https://www.amazon.science/publications/low- complexity-streaming-speech-super-resolution

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.