Pith. sign in

REVIEW 3 major objections 4 minor 1 cited by

A Comprehensive Real-World Assessment of Audio Watermarking Algorithms: Will They Survive Neural Codecs?

T0 review · 3 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read Neural codecs beat all tested audio watermarks, even retrained ones

desk verdict Useful benchmark with a real DAC-negative result, but the 'even when trained' claim outruns the evidence. read the letter →

arxiv 2505.19663 v2 pith:SFRT4R6J submitted 2025-05-26 cs.SD cs.AIcs.CRcs.LGeess.AS

Yigitcan Özer , Woosung Choi , Joan Serrà , Mayank Kumar Singh , Wei-Hsiang Liao , Yuki Mitsufuji This is my paper · ORCID
classification cs.SDcs.AIcs.CRcs.LGeess.AS
keywords audiowatermarkingneuralcodecsrobustnessbenchmarkadversarialtrainingEncodecDescriptCodecimperceptibilityRAW-Bench
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

This paper introduces RAW-Bench, a standardized benchmark for testing deep-learning audio watermarking systems against real-world distortions. The authors evaluate four published watermarking methods under a pipeline of twenty attacks, including conventional and neural compression, noise, filtering, and time-domain modifications, on a new test set spanning speech, music, and environmental sounds. The central finding is that neural compression, specifically Encodec and Descript Audio Codec, is the most damaging attack, and that retraining the watermarking models with these attacks improves bitwise accuracy but never brings full-message accuracy to acceptable levels. In fact, none of the considered methods survives the Descript Audio Codec attack. If this holds, it means current deep-learning watermarking cannot protect audio that passes through modern neural codecs, which are increasingly the final stage of real-world audio pipelines.

What carries the argument

The load-bearing object is the Robust Audio Watermarking Benchmark (RAW-Bench), built from three components: a diverse test dataset of raw 44.1 kHz recordings across music, speech, and environmental sounds; an attack pipeline of twenty distortions organized into six categories (mixing, dynamics, filtering, low-level modifications, neural compression, and conventional compression) with loose and strict parameter settings; and a retraining protocol that feeds the strict attacks into the models with uniform per-category weighting. This machinery turns the question 'do watermarks survive neural codecs?' into measurable bitwise and full-message accuracies, and it is what allows the paper to attribute failures to specific attack types rather than to dataset quirks.

What would settle it

Retrain one of the tested models with Descript Audio Codec as the dominant attack (for example, applying DAC distortion to a large share of each batch, with hyperparameters tuned for this single attack) and evaluate on the RAW-Bench DA attack at strict settings. If full-message accuracy rises well above the reported near-zero values, the paper's claim that neural codecs present a fundamental limitation would be refuted; if it stays near zero, the claim is strengthened.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that neural codecs and deep-learning audio watermarking compete for the same imperceptible information, and the codecs currently win. Across four pre-trained watermarking models, the Descript Audio Codec at 44.1 kHz reduces full-message extraction accuracy to essentially zero in every case, and Encodec is nearly as destructive. Retraining two of the models on a pipeline that includes these codec distortions improves robustness on some attacks but leaves full-message accuracy near zero for both neural codecs. The paper concludes that this is not a tuning failure but a fundamental tension: a codec that succeeds at removing imperceptible components will remove watermarks that are designed to be imperceptible.

Load-bearing premise

The conclusion that retraining with neural-codec attacks is insufficient rests on one specific retraining setup—a balanced mix of many attacks, a proprietary music dataset, relaxed quality constraints for one of the models, and undisclosed training details—so a more focused or intensive codec-training regime could in principle overturn the claim.

Editorial extensions

If this is right

  • Audio that passes through Encodec or Descript Audio Codec will lose embedded watermarks from all four tested deep-learning methods, making current watermarking unreliable for distribution chains that use neural codecs.
  • Retraining with a broad attack pipeline improves robustness on some distortions but not on neural compression, reverb, or phase shift, so attack augmentation alone is not a sufficient fix.
  • The advantage one model gains from using Encodec's architecture does not transfer to a different neural codec, indicating that codec-specific robustness does not generalize.
  • The benchmark's strict and loose attack thresholds provide a common yardstick for future watermarking systems to report robustness at matched perceptual impact.

Reading between the lines

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

  • A test the paper leaves implicit: retraining with the neural codec as the sole attack, or with a curriculum that starts on mild codec settings and hardens, would separate 'the attack is fundamentally destructive' from 'the retraining recipe was too diluted'.
  • The same RAW-Bench pipeline could be pointed at latent-based watermarks for generative audio, which may survive codecs better because they are embedded in a space the codec already preserves; that would test whether the competition-for-imperceptible-information story extends beyond carrier-signal methods.
  • If codec makers and watermark designers iterate against each other on this benchmark, the specific result about Descript Audio Codec could become a moving target, but the general trade-off between imperceptible embedding and lossy compression is a structural constraint that would remain.
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

3 major / 4 minor

Summary. The paper introduces RAW-Bench, a standardized benchmark for deep learning-based audio watermarking, together with a 20-attack robustness pipeline and a diverse test set of raw 44.1 kHz recordings spanning speech, music, and environmental sounds. It evaluates four publicly available pre-trained watermarking models (AudioSeal, SilentCipher, Timbre, WavMark) under loose and strict attack settings, and additionally retrains AudioSeal and SilentCipher with the proposed attack pipeline. The main findings are that neural codecs (Encodec and Descript Audio Codec) are the most damaging distortions, that even retraining with such codec attacks leaves full-message accuracy at or near zero for Descript Audio Codec, and that attack-augmented training generally helps but does not fix all vulnerabilities.

Significance. If the results hold, the paper makes a valuable and timely contribution: it provides a reproducible benchmark, a public test dataset, an out-of-sample verification protocol, and the first systematic cross-model comparison that includes SilentCipher and retrained variants. The negative result on neural codecs is important for the audio watermarking community, and the 'compete for the same space' argument is a thought-provoking design principle. The release of the evaluation code (github.com/SonyResearch/raw_bench) and the careful construction of the test set from multiple public corpora are concrete strengths that support the paper's utility.

major comments (3)
  1. [Section 3 (Retraining) and Section 4 (Table 5, rows AS* and SC*)] The abstract and Section 4 claim that neural codecs pose the most significant challenge 'even when algorithms are trained with such compressions.' This claim rests on a single retraining configuration: uniform per-category attack weighting (so the two neural-codec attacks receive only a small fraction of training batches), a proprietary 1250-hour music dataset plus 40-hour VCTK and BBC subsets, and no reported epochs, optimizer, loss, or learning-rate schedule. The post-retraining DA numbers (AS* bitwise 0.60, full-message 0.00; SC* bitwise 0.42, full-message 0.00) are consistent with undertraining rather than a fundamental incompatibility. I request ablations that vary the fraction of codec-distorted examples, the codebook configuration, and the training budget, and that report the training hyperparameters, before the 'even when trained' conclusion is stated as a general limitation.
  2. [Table 5] All robustness results are reported as point estimates without confidence intervals, error bars, or significance tests. For a benchmark explicitly intended to enable systematic comparison, this is a limitation for the finer-grained ranking claims (for example, AS versus TI on many attack columns, or the small AS* versus AS improvements). I request at least standard deviations or confidence intervals across test segments or independent runs, or a statistical test for the main comparisons, so that the reader can distinguish meaningful differences from noise.
  3. [Table 1 vs. Table 5] Full-message accuracy is compared across models with different message lengths (AudioSeal 16 bits, SilentCipher 23.8 bits, Timbre 30 bits, WavMark 16 bits). Since the probability that every bit decodes correctly depends on message length even at equal per-bit accuracy, cross-model comparisons of full-message accuracy are biased. The capacity is approximately matched, but the message lengths are not; I recommend reporting per-bit equivalent accuracy or fixing the payload length whenever full-message accuracy is used for cross-model conclusions.
minor comments (4)
  1. [Section 4, 'Will Watermarks Survive Neural Codecs?'] The paragraph arguing that watermarking and neural codecs 'compete for the same space' is speculative; it is labeled as a belief, but it should be explicitly presented as a hypothesis for future work rather than a conclusion of the benchmark.
  2. [Table 2] The row for DA lists 'Descript Audio Codec [16] (at 44.1 kHz)', but the abbreviations EN and DA are not expanded in the table caption; please define 'EN' as Encodec and 'DA' as Descript Audio Codec in the caption.
  3. [Section 4, Robustness paragraph] The phrase 'the architecture of AS is based on EN' is imprecise; AudioSeal uses an Encodec-based architecture, and the sentence would benefit from stating that explicitly rather than using the abbreviation EN alone.
  4. [General] The paper would benefit from a figure summarizing the Table 5 results (for example, grouped bar charts per attack category), since the current dense table is hard to read and the main qualitative findings (e.g., the DA collapse) are less visually salient than they deserve.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: all conclusions rest on external test data, pre-trained weights, and measured metrics.

full rationale

The paper is an empirical benchmark, not a derivation. Every load-bearing result is an observed accuracy on a fixed, publicly described test pipeline: the robustness numbers in Table 5 are measurements of pre-trained or retrained public models under specified distortions, and the central claim that neural compression defeats watermarking follows from those measurements, not from any construction that defines the outcome in terms of the inputs. The retraining stage does include the attack pipeline, including neural codecs, and the evaluation also applies the attack pipeline; this is a standard adversarial-training protocol, not a circular one, because the training exposure is not the same operation as the test-time scoring and no parameter is fitted to the reported accuracy values. The possible weakness that the retraining protocol is underspecified (uniform category weighting, proprietary 1250-hour dataset, lowered SDR bound, unspecified epochs and optimizer) is a validity limitation about whether the 'even when trained with such compressions' claim generalizes; it is not a circularity, because the claim is not equivalent to its inputs by definition. The paper also cites prior work by the same authors, notably SilentCipher (ref. 9) and the Piano Concerto Dataset (ref. 29), but these are used as a baseline model and a data source whose performance is measured externally, not as unverified justification for the conclusion; SilentCipher performs worse than other methods on several attacks, so self-preference cannot explain the findings. The closing 'codecs remove imperceptible information' argument is explicitly speculative ('we believe'), and it is an interpretation of the measured results rather than a premise from which they are derived. No equation, fitted parameter, or self-citation chain reduces the paper's conclusions to its inputs, so no circular step is present.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The paper introduces no new physical or mathematical entities. Its conclusions depend on benchmark design choices (attack thresholds, message lengths, retraining configuration) which are listed as free parameters, and on domain assumptions about model representativeness, dataset disjointness, and the fairness of the retraining protocol.

free parameters (3)
  • Loose/strict thresholds for each attack = GN 40 dB, BN 35 dB, RV 6 dB, LP 6000 Hz, HP 250 Hz, EQ +/-0.375 dB, TS +/-0.05, TJ 0.20, GA +/-0.50, QN 12 bits, PS…
    Chosen by an internal listening test with five expert listeners (Section 3, Attack Pipeline); these thresholds define strict versus loose attack settings and therefore shape the entire robustness ranking.
  • Message bit-length per model = 16 (AS), 23.8 (SC), 30 (TI), 16 (WM) bits
    The authors set capacity to be comparable (around 5 bps), but the differing message lengths affect full-message accuracy, and no correction or discussion is provided (Table 1).
  • Retraining configuration details = Lowered SDR bound for SC*, uniform weighting per attack category, proprietary 1250 h music + 40 h VCTK + 40 h BBC…
    Section 3 (Retraining) and Section 4 note the SDR bound was lowered 'on purpose'; without the exact configuration, the retraining result is not reproducible.
assumptions (4)
  • domain assumption The four selected models (AudioSeal, SilentCipher, Timbre, WavMark) are representative of the current design space of deep-learning audio watermarking.
    Section 1 frames conclusions about watermarking algorithms generally, but only four publicly available models are evaluated; a different or newer model could behave differently, especially for neural codec robustness.
  • domain assumption The loose/strict thresholds set by a five-listener internal listening test reflect the real-world boundary between imperceptible and acceptable distortions.
    Section 3 (Attack Pipeline) defines the entire loose/strict split from this subjective test; the robustness ranking is conditional on these thresholds.
  • domain assumption The test datasets are disjoint from the training data of the evaluated models, as asserted by the authors.
    Section 3 (Test Dataset) states a verification was performed, but no procedure or artifact is shown; if overlap existed, the clean-condition and robustness numbers would be inflated.
  • domain assumption The retraining protocol (uniform category weighting, SpecAugment, proprietary data) is a fair test of whether adversarial training can confer robustness to neural codecs.
    Section 3 (Retraining) and Section 4 (Re-training) use this protocol to conclude that training with codec attacks is insufficient; the absence of ablations means a codec-specialized training scheme might succeed.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Comprehensive Real-World Assessment of Audio Watermarking Algorithms: Will They Survive Neural Codecs?." pith.science (2026). https://pith.science/paper/SFRT4R6J

@misc{pith2026250519663,
  author       = {Pith},
  title        = {Pith review of: A Comprehensive Real-World Assessment of Audio Watermarking Algorithms: Will They Survive Neural Codecs?},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SFRT4R6J}},
  note         = {Machine review of arXiv:2505.19663}
}
read the original abstract

We introduce the Robust Audio Watermarking Benchmark (RAW-Bench), a benchmark for evaluating deep learning-based audio watermarking methods with standardized and systematic comparisons. To simulate real-world usage, we introduce a comprehensive audio attack pipeline with various distortions such as compression, background noise, and reverberation, along with a diverse test dataset including speech, environmental sounds, and music recordings. Evaluating four existing watermarking methods on RAW-bench reveals two main insights: (i) neural compression techniques pose the most significant challenge, even when algorithms are trained with such compressions; and (ii) training with audio attacks generally improves robustness, although it is insufficient in some cases. Furthermore, we find that specific distortions, such as polarity inversion, time stretching, or reverb, seriously affect certain methods. The evaluation framework is accessible at github.com/SonyResearch/raw_bench.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Who Gets Flagged? The Pluralistic Evaluation Gap in AI Content Watermarking

    cs.CY 2026-04 conditional novelty 6.0 of 10

    Major watermarking benchmarks omit cross-lingual, cultural, and demographic reporting, creating a pluralistic evaluation gap that current governance mandates ignore.

Reference graph

Works this paper leans on

40 extracted references · 37 canonical work pages · cited by 1 Pith paper

  1. [1]

    However, such advancements have also introduced chal- lenges in content authenticity and copyright protection [5, 6]

    Introduction Recent advances in audio-based applications have enabled seamless content sharing, improved creative workflows, and facilitated the widespread adoption of generative AI models [1–4]. However, such advancements have also introduced chal- lenges in content authenticity and copyright protection [5, 6]. To address these challenges, audio watermar...

    work page

  2. [2]

    Related Work To the best of our knowledge, the only study that compares deep learning-based audio watermarking models is AudioMark- Bench [20]. AudioMarkBench is a benchmarking framework that evaluates the robustness of three audio watermarking mod- els (AudioSeal, Timbre, and WavMark), using their publicly available pre-trained weights, on a subset of sp...

    arXiv 2025

  3. [3]

    It includes clas- sical and popular music, speech, and environmental sounds, which account for a wide range of real-world use cases

    Methodology Test Dataset — To evaluate watermarking algorithms in var- ious domains, we create a comprehensive test dataset using open-source collections from various sources. It includes clas- sical and popular music, speech, and environmental sounds, which account for a wide range of real-world use cases. To maintain a high fidelity, all audio recording...

    work page

  4. [4]

    Results and Discussion Imperceptibility — As a first step, we evaluate the considered models in clean (distortion-free) conditions, focusing on over- all perceptual quality and detection accuracy (Table 3). Among all pre-trained models, SC consistently outperforms others in perceptual quality, achieving the highest SI-SNR and lowest MCD, indicating minima...

    work page

  5. [5]

    We de- signed a comprehensive audio attack pipeline that simulates real-world distortions, and introduced a diverse test dataset comprising multiple audio domains

    Conclusion We introduced a systematic evaluation framework for deep learning-based audio watermarking algorithms, addressing im- portant gaps in robustness analysis and benchmarking. We de- signed a comprehensive audio attack pipeline that simulates real-world distortions, and introduced a diverse test dataset comprising multiple audio domains. By studyin...

    work page

  6. [6]

    Fastspeech 2: Fast and high-quality end-to-end text to speech,

    Y . Ren, C. Hu, X. Tan, T. Qin, S. Zhao, Z. Zhao, and T.-Y . Liu, “Fastspeech 2: Fast and high-quality end-to-end text to speech,” in Proc. of the Int. Conf. on Learn. Represent. (ICLR) , Vienna, Austria, 2021

    work page 2021

  7. [7]

    V oiceBox: Text-guided multilingual universal speech gen- eration at scale,

    M. Le, A. Vyas, B. Shi, B. Karrer, L. Sari, R. Moritz, M. Williamson, V . Manohar, Y . Adi, J. Mahadeokar, and W.-N. Hsu, “V oiceBox: Text-guided multilingual universal speech gen- eration at scale,” in Adv. in Neural Inf. Process. Syst. (NeurIPS) , New Orleans, LA, 2023, pp. 14 005–14 034

    work page 2023

  8. [8]

    Simple and controllable music gen- eration,

    J. Copet, F. Kreuk, I. Gat, T. Remez, D. Kant, G. Synnaeve, Y . Adi, and A. D ´efossez, “Simple and controllable music gen- eration,” in Adv. in Neural Inf. Process. Syst. (NeurIPS), New Or- leans, LA, USA, 2023

    work page 2023

Show all 40 references
  1. [9]

    Music ControlNet: Multiple time-varying controls for music genera- tion,

    S.-L. Wu, C. Donahue, S. Watanabe, and N. J. Bryan, “Music ControlNet: Multiple time-varying controls for music genera- tion,” IEEE/ACM Trans. on Audio, Speech, and Lang. Process. , vol. 32, pp. 2692–2703, 2024

    2024

  2. [10]

    A comparative study on recent neu- ral spoofing countermeasures for synthetic speech detection,

    X. Wang and J. Yamagishi, “A comparative study on recent neu- ral spoofing countermeasures for synthetic speech detection,” in Proc. of the Annual Conf. of the Int. Speech Commun. Assoc. (In- terspeech), Brno, Czech Republic, 2021, pp. 4259–4263

    2021

  3. [11]

    Towards assessing data replication in music generation with mu- sic similarity metrics on raw audio,

    R. Batlle-Roca, W.-H. Liao, X. Serra, Y . Mitsufuji, and E. G´omez, “Towards assessing data replication in music generation with mu- sic similarity metrics on raw audio,” in Proc. of the Int. Soc. for Music Inf. Retriev. Conf. (ISMIR), San Francisco, CA, USA, 2024, pp. 1004–1011

    2024

  4. [12]

    Twenty years of digital audio watermarking – A comprehensive review,

    G. Hua, J. Huang, Y . Q. Shi, J. Goh, and V . L. Thing, “Twenty years of digital audio watermarking – A comprehensive review,” Signal Process., vol. 128, pp. 222–242, 2016

    2016

  5. [13]

    Proactive detection of voice cloning with localized watermarking,

    R. S. Roman, P. Fernandez, H. Elsahar, A. D ´efossez, T. Furon, and T. Tran, “Proactive detection of voice cloning with localized watermarking,” inProc. of the Int. Conf. on Mach. Learn. (ICML), Vienna, Austria, 2024

    2024

  6. [14]

    SilentCi- pher: Deep audio watermarking,

    M. K. Singh, N. Takahashi, W. Liao, and Y . Mitsufuji, “SilentCi- pher: Deep audio watermarking,” in Proc. of the Annual Conf. of the Int. Speech Commun. Assoc. (Interspeech) , Kos Island, Greece, 2024, pp. 2235–2239

    2024

  7. [15]

    De- tecting voice cloning attacks via Timbre Watermarking,

    C. Liu, J. Zhang, T. Zhang, X. Yang, W. Zhang, and N. Yu, “De- tecting voice cloning attacks via Timbre Watermarking,” inNetw. and Distrib. Syst.Secur. Symp., Vancouver, Canada, 2024

    2024

  8. [16]

    Wavmark: Watermarking for audio generation,

    G. Chen, Y . Wu, S. Liu, T. Liu, X. Du, and F. Wei, “Wavmark: Watermarking for audio generation,” 2023

    2023

  9. [17]

    Maskmark: Robust neural watermarking for real and synthetic speech,

    P. O’Reilly, Z. Jin, J. Su, and B. Pardo, “Maskmark: Robust neural watermarking for real and synthetic speech,” in Proc. of the IEEE Int. Conf. on Acoust., Speech, and Signal Process. (ICASSP) , Seoul, South Korea, 2024, pp. 4650–4654

    2024

  10. [18]

    Latent watermarking of audio generative models,

    R. S. Roman, P. Fernandez, A. Deleforge, Y . Adi, and R. Ser- izel, “Latent watermarking of audio generative models,” in Proc. of the IEEE Int. Conf. on Acoust., Speech, and Signal Process. (ICASSP), Hyderabad, India, 2025

    2025

  11. [19]

    Survey of robust and imperceptible watermarking,

    N. Agarwal, A. K. Singh, and P. K. Singh, “Survey of robust and imperceptible watermarking,” Multimed. Tools Appl., vol. 78, no. 7, pp. 8603–8633, 2019

    2019

  12. [20]

    High fidelity neural audio compression,

    A. D ´efossez, J. Copet, G. Synnaeve, and Y . Adi, “High fidelity neural audio compression,” Trans. on Mach. Learn. Research (TMLR), 2023

    2023

  13. [21]

    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,” in Adv. in Neural Inf. Process. Syst. (NeurIPS) , New Orleans, LA, USA, 2023

    2023

  14. [22]

    Audio codec augmentation for ro- bust collaborative watermarking of speech synthesis,

    L. Juvela and X. Wang, “Audio codec augmentation for ro- bust collaborative watermarking of speech synthesis,” in Proc. of the IEEE Int. Conf. on Acoust., Speech, and Signal Process. (ICASSP), Hyderabad, India, 2025

    2025

  15. [23]

    The diverse environments multi-channel acoustic noise database (DEMAND): A database of multichannel environmental noise recordings,

    J. Thiemann, N. Ito, and E. Vincent, “The diverse environments multi-channel acoustic noise database (DEMAND): A database of multichannel environmental noise recordings,” Proc. of Meet. on Acoust., vol. 19, no. 1, p. 035081, 2013

    2013

  16. [24]

    A binaural room impulse re- sponse database for the evaluation of dereverberation algorithms,

    M. Jeub, M. Sch ¨afer, and P. Vary, “A binaural room impulse re- sponse database for the evaluation of dereverberation algorithms,” in Proc. of Int. Conf. on Digital Signal Process. (DSP), Santorini, Greece, 2009, pp. 1–4

    2009

  17. [25]

    AudioMark- Bench: Benchmarking robustness of audio watermarking,

    H. Liu, M. Guo, Z. Jiang, L. Wang, and N. Z. Gong, “AudioMark- Bench: Benchmarking robustness of audio watermarking,” inAdv. in Neural Inf. Process. Syst. (NeurIPS), Vancouver, Canada, 2024

    2024

  18. [26]

    Soundprism: An online system for score- informed source separation of music audio,

    Z. Duan and B. Pardo, “Soundprism: An online system for score- informed source separation of music audio,” IEEE Journal of Se- lected Topics in Signal Process. , vol. 5, no. 6, pp. 1205–1215, 2011

    2011

  19. [27]

    Clotho: An audio captioning dataset,

    K. Drossos, S. Lipping, and T. Virtanen, “Clotho: An audio captioning dataset,” in Proc. of the IEEE Int. Conf. on Acoust., Speech, and Signal Process. (ICASSP), 2020, pp. 736–740

    2020

  20. [28]

    G. J. Mysore, “Can we automatically transform speech recorded on common consumer devices in real-world environments into professional production quality speech? – A dataset, insights, and challenges,” IEEE Signal Process. Lett., vol. 22, no. 8, pp. 1006– 1010, 2015

    2015

  21. [29]

    Freisch¨utz Dig- ital: Demos of audio-related contributions,

    T. Pr ¨atzlich, M. M¨uller, B. W. Bohl, and J. Veit, “Freisch¨utz Dig- ital: Demos of audio-related contributions,” in Demos and Late Breaking News of the Int. Soc. for Music Inf. Retriev. Conf. (IS- MIR), M´alaga, Spain, 2015

    2015

  22. [30]

    Guitarset: A dataset for guitar transcription

    Q. Xi, R. Bittner, J. Pauwels, X. Ye, and J. P. Bello, “Guitarset: A dataset for guitar transcription.” inProc. of the Int. Soc. for Music Inf. Retriev. Conf. (ISMIR), Paris, France, 2018, pp. 453–460

    2018

  23. [31]

    jaCappella corpus: A Japanese a cappella vo- cal ensemble corpus,

    T. Nakamura, S. Takamichi, N. Tanji, S. Fukayama, and H. Saruwatari, “jaCappella corpus: A Japanese a cappella vo- cal ensemble corpus,” in Proc. of the IEEE Int. Conf. on Acoust., Speech, and Signal Process. (ICASSP), 2023

    2023

  24. [32]

    Enabling factor- ized piano music modeling and generation with the MAESTRO dataset,

    C. Hawthorne, A. Stasyuk, A. Roberts, I. Simon, C. A. Huang, S. Dieleman, E. Elsen, J. H. Engel, and D. Eck, “Enabling factor- ized piano music modeling and generation with the MAESTRO dataset,” in Proc. of the Int. Conf. on Learn. Represent. (ICLR) , New Orleans, LA, USA, 2019

    2019

  25. [33]

    MoisesDB: A dataset for source separation beyond 4-stems,

    I. Pereira, F. Ara ´ujo, F. Korzeniowski, and R. V ogl, “MoisesDB: A dataset for source separation beyond 4-stems,” in Proc. of the Int. Soc. for Music Inf. Retriev. Conf. (ISMIR), Milan, Italy, 2023, pp. 619–626

    2023

  26. [34]

    Piano Concerto Dataset (PCD): A multitrack dataset of piano concertos,

    Y . ¨Ozer, S. Schw ¨ar, V . Arifi-M¨uller, J. Lawrence, E. Sen, and M. M¨uller, “Piano Concerto Dataset (PCD): A multitrack dataset of piano concertos,” Trans. of the Int. Soc. for Music Inf. Retriev. (TISMIR), vol. 6, no. 1, pp. 75–88, 2023

    2023

  27. [35]

    SpecAugment: A simple data augmen- tation method for automatic speech recognition,

    D. S. Park, W. Chan, Y . Zhang, C.-C. Chiu, B. Zoph, E. D. Cubuk, and Q. V . Le, “SpecAugment: A simple data augmen- tation method for automatic speech recognition,” in Proc. of the Annual Conf. of the Int. Speech Commun. Assoc. (Interspeech) , Graz, Austria, 2019, pp. 2613–2617

    2019

  28. [36]

    CSTR VCTK Cor- pus: English multi-speaker corpus for CSTR voice cloning toolkit (version 0.92),

    C. Veaux, J. Yamagishi, and K. MacDonald, “CSTR VCTK Cor- pus: English multi-speaker corpus for CSTR voice cloning toolkit (version 0.92),” 2019

    2019

  29. [37]

    BBC sound effects library,

    B. B. Corporation, “BBC sound effects library,” 1991, ac- cessed: 2025-02-07. [Online]. Available: https://sound-effects. bbcrewind.co.uk/

    1991

  30. [38]

    Conv-TasNet: Surpassing ideal time–frequency magnitude masking for speech separation,

    Y . Luo and N. Mesgarani, “Conv-TasNet: Surpassing ideal time–frequency magnitude masking for speech separation,” IEEE/ACM Trans. on Audio, Speech, and Lang Process., vol. 27, no. 8, pp. 1256–1266, 2019

    2019

  31. [39]

    Mel-cepstral distance measure for objective speech quality assessment,

    R. F. Kubichek, “Mel-cepstral distance measure for objective speech quality assessment,” in Proc. of IEEE Pacific Rim Conf. on Commun. Comp. and Signal Process. , vol. 1, 1993, pp. 125– 128

    1993

  32. [40]

    ViSQOL: an objective speech quality model,

    A. Hines, J. Skoglund, A. C. Kokaram, and N. Harte, “ViSQOL: an objective speech quality model,” EURASIP Journal on Audio, Speech, and Music Process., vol. 2015, pp. 1–18, 2015

    2015

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.