Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

SingNet: Towards a Large-Scale, Diverse, and In-the-Wild Singing Voice Dataset

T0 review · 3 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read This paper claims that an open pipeline can turn internet songs and sample packs into about 3,000 hours of usable singing voice data, and that pre-training on this corpus improves lyric transcription, vocoders, and singing voice conversion.

desk verdict SingNet is the largest in-the-wild singing dataset with a genuinely useful pipeline, but a likely train/test overlap and unreleased resources weaken the empirical claims. read the letter →

arxiv 2505.09325 v1 pith:E6LQAXUS submitted 2025-05-14 cs.SD eess.AS

Yicheng Gu , Chaoren Wang , Junan Zhang , Xueyao Zhang , Zihao Fang , Haorui He , Zhizheng Wu This is my paper · ORCID
classification cs.SDeess.AS
keywords SingNetsingingvoicedatasetin-the-wildaudiodataprocessingpipelinesourceseparationautomaticlyrictranscriptionneuralvocoderconversion
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

SingNet tries to break the data bottleneck that has kept singing voice synthesis and conversion tied to small, expensive studio recordings. The authors build an automated, open-source pipeline that mines about 3,000 hours of singing voice from internet songs and sample packs, spanning many languages and styles. They then pre-train Wav2vec2, BigVGAN, and NSF-HiFiGAN models on this data and benchmark them on automatic lyric transcription, neural vocoding, and singing voice conversion. If the pipeline and benchmarks hold up, singing voice research gains a scalable, renewable data source and open checkpoints that do not require professional studio recording.

What carries the argument

The load-bearing mechanism is the open data-processing pipeline. It separates vocals from songs with the MDX23 model, runs a chain of restoration plugins to remove noise, reverb, clicks, and plosives, segments audio by lyric timestamps or voice activity detection, rejects multi-singer clips via multi-F0 detection, and keeps only clips scored above 3.0 by a fine-tuned MOS predictor. For sample packs, a semi-supervised captioning system trains an audio classifier that separates acapella and adlibs from elements and FX. Pre-trained checkpoints built on this data serve as the benchmark instruments for ALT, vocoder, and SVC experiments.

What would settle it

Run a near-duplicate audio search between the 6-hour SingNet-SS evaluation set and the 3,500-hour training mixture; if any exact or near-exact segments appear, the reported ALT/vocoder/SVC gains could come from memorization rather than from dataset scale. A second concrete check is a blind listening test in which experienced listeners mark artifacts on clips that scored just above 3.0.

Watch

Extended reading notes

Core claim

The central discovery claimed is that the singing-voice data scarcity bottleneck can be addressed without studio recording: a pipeline of source separation, audio restoration, segmentation, multi-F0 rejection, and MOS-based quality filtering yields about 3,000 hours of multilingual, multi-style singing voice data (2,629 hours from in-the-wild songs and 321 hours from sample packs). Pre-training on this mixture, optionally combined with existing singing datasets to form 3,500 hours, produces models that outperform or match existing open-sourced models on automatic lyric transcription, neural vocoder quality, and singing voice conversion benchmarks.

Load-bearing premise

The benchmark results assume the MOS quality gate (>3.0) and the separation/restoration chain leave no audible artifacts, and that the 6-hour in-the-wild evaluation set contains no actual training audio from the 3,500-hour mixture.

Editorial extensions

If this is right

  • Singing-voice self-supervised pre-training makes ALT fine-tuning possible without speech-based transfer learning: the SingNet-trained Wav2vec2 reaches 6.76% WER versus 7.79% for the transfer-learning approach.
  • Vocoders trained on 3,500 hours of singing voice clearly beat speech/general-sound-only baselines on in-the-wild audio (BigVGAN MOS 3.55 vs 3.13; NSF-HiFiGAN MOS 3.52 vs 3.39).
  • SVC subjective quality and similarity improve at the 3,500-hour scale (studio MOS 3.73 and SMOS 3.71; in-the-wild MOS 3.16 and SMOS 3.11), although objective F0 and intelligibility metrics are roughly flat across scales.
  • The dataset can be extended over time by running the released pipeline on more sources, making scale a renewable resource rather than a fixed studio corpus.

Reading between the lines

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

  • If SingNet's scale and cleanliness hold up, singing-voice synthesis could follow the speech path where web-scale data plus robust pre-training replaces studio-perfect corpora; the released checkpoints make this testable by any group.
  • The curious SVC pattern where 35 and 3,500 hours beat 350 hours suggests an intermediate-data regime in which encoder capacity lags diversity; a testable extension is to vary encoder size with data diversity to see whether the dip disappears.
  • A direct test of the MOS filter's role would be to train the vocoder and SVC on unfiltered versus filtered SingNet subsets; if gains vanish without filtering, the quality gate is the real driver rather than raw hours.
  • The sample-pack classifier could become a standalone tool for mining studio-quality acapella stems at scale, labeling newly released packs as dry or wet acapella and adlibs without human captions.
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 / 5 minor

Summary. SingNet proposes a pipeline for extracting large-scale in-the-wild singing voice data from internet songs and sample packs, yielding around 3,000 hours of multilingual, multi-style data. The authors pre-train Wav2vec2, BigVGAN, and NSF-HiFiGAN on a 3,500-hour singing mixture and report benchmarks for automatic lyric transcription, neural vocoding, and zero-shot singing voice conversion. They argue that their dataset and open-sourced checkpoints outperform existing smaller or speech/general-audio-based training sets, and that scaling up singing data improves downstream performance.

Significance. If the dataset, pipeline, and checkpoints are released and the benchmark claims are validated with clean train/test separation, SingNet would be a valuable community resource: it is the largest in-the-wild singing dataset to date, the pipeline is an original contribution, and the pre-trained models provide useful baselines. The paper also makes a falsifiable scaling claim about singing voice data. However, the current experimental design does not yet support these claims, because the evaluation sets appear to overlap with the training mixture and the subjective scaling results are statistically inconclusive.

major comments (3)
  1. [Section 4.1.1 and Section 4.1.3] The evaluation sets are not shown to be disjoint from the training data. The 3,500-hour training mixture is described as including all existing singing datasets, which includes SingStyle111, and also includes SingNet-SS (Section 4.1.1). The Studio Recording evaluation is a random 1-hour sample from SingStyle111, and the In-the-wild evaluation is a 6-hour random sample from SingNet-SS. The paper never states that these samples were excluded from training, nor does it describe deduplication at the song, singer, or recording level. If even part of the test material appears in the training mixture, Tables 2-4 may reflect memorization rather than generalization. This is a load-bearing issue for the central claim that SingNet helps downstream tasks, and it also extends to the ALT and vocoder experiments. The authors must specify the exclusion/deduplication procedure and ideally rerun the experiments on strictly held-out material.
  2. [Table 4 and Section 4.5] The SVC scaling claim is not supported by the reported confidence intervals. In the Studio Recording setting, MOS is 3.72±0.17 for 35h and 3.73±0.19 for 3500h, and SMOS is 3.55±0.25 vs 3.71±0.27; in the In-the-wild setting, MOS is 3.16±0.16 for 3500h versus 3.01±0.17 for 350h, and SMOS is 3.11±0.25 vs 2.98±0.24. All pairs overlap within their 95% intervals. Moreover, the 350h condition is non-monotonic (worse than the 35h condition on several metrics), yet the text asserts that the 3500h system 'performs the best'. Stating that 'the MOS scores are within 95% CI' is not the same as showing a significant difference. The claim that scale alone drives SVC quality needs significance testing, more listeners/utterances, or a different evaluation design.
  3. [Data availability and licensing] The paper is a dataset paper but never states whether the SingNet audio, annotations, and pipeline code will be released, nor under what license. The only explicit release commitment is for checkpoints 'after the double-blind review period' (Section 1). Without a clear data-release plan, the dataset contribution cannot be used or independently verified, and the paper's central artifact remains inaccessible. Please add a data availability statement covering the audio, the sample-pack captions, the pipeline scripts, and the license/usage terms, including any copyright or redistribution restrictions on the crawled internet audio.
minor comments (5)
  1. [Section 1 (contributions)] In the contribution bullet list, 'in-the-wide' should be 'in-the-wild'.
  2. [Section 3.3 / Figure 4] The diversity claim is supported only by a qualitative PCA scatter plot. A quantitative measure of coverage or dispersion (e.g., estimated volume, entropy, or nearest-neighbor statistics) would strengthen the claim that SingNet is more acoustically and semantically diverse than the existing-dataset mixture.
  3. [Section 3.1.3] The MOS-based filtering threshold of 3.0 appears to be a hand-chosen value. A sentence justifying this choice, or a brief sensitivity analysis showing how downstream performance varies with the threshold, would make the pipeline more reproducible.
  4. [Table 3] In the Studio Recording setting for BigVGAN, adding SingingVoice-3500 improves MCD (1.777 to 1.520) but worsens F0RMSE (35.897 to 37.125). The text claims the singing-voice-trained system 'significantly outperformed' the Large-Compilation system; this should be qualified because the F0 metric moves in the opposite direction.
  5. [Section 4.5] The 'manual review' of synthesized samples is anecdotal and should be described as such, not used as primary evidence for the scaling claim. If quantitative support exists (e.g., error analysis on a subset), it should be reported systematically.

Circularity Check

2 steps flagged · score 6.0 of 10

Benchmark evaluations are drawn from the training mixture, so the claimed effectiveness of SingNet is not tested on held-out data.

  1. fitted input called prediction [Section 4.1.1 (Datasets) and Section 4.3 (ALT)]
    "We utilize all the existing singing voice datasets for training, as illustrated in Table. 1, resulting in a singing voice mixture of 3500 hours with various recording qualities, styles, singers, and languages. ... We use SingStyle111 as the test set."

    SingStyle111 is one of the 'existing singing voice datasets' in Table 1 and is therefore included in the 3500-hour SingingVoice-3500 mixture used to pre-train the Wav2vec2 model. The ALT evaluation then reports WER on SingStyle111. Since the pre-training data includes the test set, the model's lower WER compared with the LibriVox-only baseline reflects exposure to the test data, not a measured generalization benefit of large-scale singing data.

  2. fitted input called prediction [Section 4.1.1 (Datasets) and Sections 4.4-4.5 (vocoder and SVC experiments)]
    "We utilize all the existing singing voice datasets for training, as illustrated in Table. 1, resulting in a singing voice mixture of 3500 hours with various recording qualities, styles, singers, and languages. Two datasets are used for evaluation; we randomly sample 1 hour and 6 hours of multilingual, multi-singer, and multi-style audio from SingStyle111 and SingNet-SS to form the Studio Recording and In-the-wild evaluation sets, respectively."

    Both SingStyle111 and SingNet-SS are part of the 'all existing singing voice datasets' that compose the training mixture. The vocoder and SVC models trained on SingingVoice-3500 are therefore evaluated on audio drawn from their own training data. The reported improvements on the In-the-wild setting (Tables 3 and 4) are not independent predictions; they are measurements on training-set samples, or at least on the same source datasets without a stated exclusion.

full rationale

The paper contains no mathematical derivation whose conclusion is equivalent to its premises, and the only self-citation (Emilia, which shares four authors) is used merely as inspiration and is not load-bearing. The significant circularity is experimental: Section 4.1.1 states that the training mixture is formed from 'all the existing singing voice datasets' (Table 1 includes SingStyle111 and SingNet-SS), and then the same section defines the Studio Recording and In-the-wild evaluation sets as random samples of SingStyle111 and SingNet-SS. No exclusion of the evaluation samples from the training mixture is described. Consequently, the ALT, vocoder, and SVC benchmark results that are used to demonstrate the effectiveness of SingNet are computed on data that were part of the training set, making the reported gains substantially forced by construction rather than evidence of generalization. This warrants a score of 6 under the rubric: one or more 'predictions' reduce, by the paper's own data-split description, to a fit on the evaluation set.

Assumptions & free parameters 2 free parameters · 2 assumptions · 0 invented entities

The central claims rest mainly on domain assumptions about the reliability of source separation, DSP restoration, and MOS-based filtering. There are no new physical entities or formal axioms; the free parameters are data-filtering thresholds that affect the dataset's final composition.

free parameters (2)
  • MOS filtering threshold = 3.0
    Hand-chosen threshold on the fine-tuned singing voice scorer to keep audio segments during filtering (Section 3.1.3). It directly determines the final dataset content and size.
  • Minimum VAD segment length = 0.5 s
    Clips shorter than 0.5 seconds are discarded after VAD segmentation in the SingNet-SP pipeline (Section 3.1.2). Affects the number and length of training utterances.
assumptions (2)
  • domain assumption Source separation and audio restoration tools can recover clean singing vocals from in-the-wild songs and sample packs.
    Section 3.1.1 assumes MDX23 and the VST chain produce usable dry vocals; no independent quality measurement is reported beyond the MOS filter.
  • domain assumption The fine-tuned MOS predictor trained on SingMOS is a valid proxy for singing voice quality when filtering in-the-wild data.
    Section 3.1.3 uses a threshold of 3.0 on this scorer to decide what data to keep, but the scorer's calibration on heavily processed internet audio is not verified.

how reviews work

0 comments
Cite this review

Pith. "Pith review of SingNet: Towards a Large-Scale, Diverse, and In-the-Wild Singing Voice Dataset." pith.science (2026). https://pith.science/paper/E6LQAXUS

@misc{pith2026250509325,
  author       = {Pith},
  title        = {Pith review of: SingNet: Towards a Large-Scale, Diverse, and In-the-Wild Singing Voice Dataset},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/E6LQAXUS}},
  note         = {Machine review of arXiv:2505.09325}
}
read the original abstract

The lack of a publicly-available large-scale and diverse dataset has long been a significant bottleneck for singing voice applications like Singing Voice Synthesis (SVS) and Singing Voice Conversion (SVC). To tackle this problem, we present SingNet, an extensive, diverse, and in-the-wild singing voice dataset. Specifically, we propose a data processing pipeline to extract ready-to-use training data from sample packs and songs on the internet, forming 3000 hours of singing voices in various languages and styles. Furthermore, to facilitate the use and demonstrate the effectiveness of SingNet, we pre-train and open-source various state-of-the-art (SOTA) models on Wav2vec2, BigVGAN, and NSF-HiFiGAN based on our collected singing voice data. We also conduct benchmark experiments on Automatic Lyric Transcription (ALT), Neural Vocoder, and Singing Voice Conversion (SVC). Audio demos are available at: https://singnet-dataset.github.io/.

Figures

Figures reproduced from arXiv: 2505.09325 by the authors.

Figure 1
Figure 1. An overview of the SingNet data processing pipeline. It processes in-the-wild songs and [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Duration statistics (hours) of SingNet by language and style sorted by the data scales. “MIS” [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. BPM and Pitch statistics (occurrences) of SingNet. The pitch is illustrated as MIDI notes [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Comparison of acoustic and semantic diversities between SingNet and the mixture of [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: An overview of the audio annotation website. The sample packs used in annotation, the [PITH_FULL_IMAGE:figures/full_fig_p015_5.png]
Figure 6
Figure 6. Figure 6: An overview of the MOS test [PITH_FULL_IMAGE:figures/full_fig_p016_6.png]
Figure 7
Figure 7. Figure 7: An overview of the SMOS test. We adapted some ideas from the MUSHRA test to ensure the effectiveness of our subjective evaluation, as illustrated in [PITH_FULL_IMAGE:figures/full_fig_p016_7.png]

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. Aliasing-Free Neural Audio Synthesis

    cs.SD 2025-12 conditional novelty 6.0 of 10

    Pupu-Vocoder and Pupu-Codec use a closed-form anti-aliased SnakeBeta activation and resampling-based upsampling to reduce aliasing and improve singing, music, and audio synthesis.

Reference graph

Works this paper leans on

69 extracted references · 58 canonical work pages · cited by 1 Pith paper

  1. [1]

    Gpt-4 technical repor t

    Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. Gpt-4 technical repor t. arXiv preprint arXiv:2303.08774, 2023

    arXiv 2023

  2. [2]

    wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations

    Alexei Baevski, Yuhao Zhou, Abdelrahman Mohamed, and Michael Auli. wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations . In NeurIPS , 2020

    work page 2020

  3. [3]

    Automatic identification of emotional cues in Chinese opera singing

    Dawn AA Black, Ma Li, and Mi Tian. Automatic identification of emotional cues in Chinese opera singing . ICMPC , 2014

    work page 2014

  4. [4]

    WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing

    Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Xiangzhan Yu, and Furu Wei. WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing . JSTSP , 2022

    work page 2022

  5. [5]

    LDM-SVC: Latent Diffusion Model Based Zero-Shot Any-to-Any Singing Voice Conversion with Singer Guidance

    Shihao Chen, Yu Gu, Jie Zhang, Na Li, Rilin Chen, Liping Chen, and Lirong Dai. LDM-SVC: Latent Diffusion Model Based Zero-Shot Any-to-Any Singing Voice Conversion with Singer Guidance . CoRR , abs/2406.05325, 2024

    arXiv 2024

  6. [6]

    Children’s song dataset for singing voice research

    Soonbeom Choi, Wonil Kim, Saebyul Park, Sangeon Yong, and Juhan Nam. Children’s song dataset for singing voice research . In ISMIR , 2020

    work page 2020

  7. [7]

    w2v-BERT: Combining Contrastive Learning and Masked Language Modeling for Self-Supervised Speech Pre-Training

    Yu - An Chung, Yu Zhang, Wei Han, Chung - Cheng Chiu, James Qin, Ruoming Pang, and Yonghui Wu. w2v-BERT: Combining Contrastive Learning and Masked Language Modeling for Self-Supervised Speech Pre-Training . In ASRU , 2021

    work page 2021

  8. [8]

    Generalization Ability of MOS Prediction Networks

    Erica Cooper, Wen - Chin Huang, Tomoki Toda, and Junichi Yamagishi. Generalization Ability of MOS Prediction Networks . In ICASSP , 2022

    work page 2022

Show all 69 references
  1. [9]

    Multiple F0 Estimation in Vocal Ensembles using Convolutional Neura l networks

    Helena Cuesta, Brian McFee, and Emilia G \' o mez. Multiple F0 Estimation in Vocal Ensembles using Convolutional Neura l networks. In ISMIR , 2020

    2020

  2. [10]

    Dannenberg

    Shuqi Dai, Yuxuan Wu, Siqi Chen, Roy Huang, and Roger B. Dannenberg. SingStyle111: A Multilingual Singing Dataset With Style Transfer . In ISMIR , 2023

    2023

  3. [11]

    The NUS sung and spoken lyrics corpus: A quantitative comparison of singing and speech

    Zhiyan Duan, Haotian Fang, Bo Li, Khe Chai Sim, and Ye Wang. The NUS sung and spoken lyrics corpus: A quantitative comparison of singing and speech . In APSIPA , 2013

    2013

  4. [12]

    Fabbro, S

    G. Fabbro, S. Uhlich, C.-H. Lai, W. Choi, M. Martínez-Ramírez, W. Liao, Gadelha I., G. Ramos, E. Hsu, H. Rodrigues, F.-R. Stöter, A. Défossez, Y. Luo, J. Yu, D. Chakraborty, S. Mohanty, R. Solovyev, A. Stempkovskiy, T. Habruseva, N. Goswami, T. Harada, M. Kim, J. H. Lee, Y. Do...

    2023

  5. [13]

    An Investigation of Time-Frequency Representation Discriminators for High-Fidelity Vocoder

    Yicheng Gu, Xueyao Zhang, Liumeng Xue, Haizhou Li, and Zhizheng Wu. An Investigation of Time-Frequency Representation Discriminators for High-Fidelity Vocoder . CoRR , abs/2404.17161, 2024 a

    2024 arXiv

  6. [14]

    Multi-scale sub-band constant-q transform discriminator for high-fidelity vocoder

    Yicheng Gu, Xueyao Zhang, Liumeng Xue, and Zhizheng Wu. Multi-scale sub-band constant-q transform discriminator for high-fidelity vocoder. In ICASSP 2024-2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp.\ 10616--10620. IEEE, 2024 b

    2024

  7. [15]

    SingAug: Data Augmentation for Singing Voice Synthesis with Cycle-consistent Training Strategy

    Shuai Guo, Jiatong Shi, Tao Qian, Shinji Watanabe, and Qin Jin. SingAug: Data Augmentation for Singing Voice Synthesis with Cycle-consistent Training Strategy . In Interspeech, 2022

    2022

  8. [16]

    Emilia: An Extensive, Multilingual, and Diverse Speech Dataset for Large-Scale Speech Generation

    Haorui He, Zengqiang Shang, Chaoren Wang, Xuyuan Li, Yicheng Gu, Hua Hua, Liwei Liu, Chen Yang, Jiaqi Li, Peiyang Shi, Yuancheng Wang, Kai Chen, Pengyuan Zhang, and Zhizheng Wu. Emilia: An Extensive, Multilingual, and Diverse Speech Dataset for Large-Scale Speech Generation . ...

    2024

  9. [17]

    On the Improvement of Singing Voice Separation for Monaural Recordings Using the MIR-1K Dataset

    Chao - Ling Hsu and Jyh - Shing Roger Jang. On the Improvement of Singing Voice Separation for Monaural Recordings Using the MIR-1K Dataset . TASLP , 2010

    2010

  10. [18]

    HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units

    Wei - Ning Hsu, Benjamin Bolte, Yao - Hung Hubert Tsai, Kushal Lakhotia, Ruslan Salakhutdinov, and Abdelrahman Mohamed. HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units . TASLP , 2021

    2021

  11. [19]

    Multi-Singer: Fast Multi-Singer Singing Voice Vocoder With A Large-Scale Corpus

    Rongjie Huang, Feiyang Chen, Yi Ren, Jinglin Liu, Chenye Cui, and Zhou Zhao. Multi-Singer: Fast Multi-Singer Singing Voice Vocoder With A Large-Scale Corpus . In ACM MM , 2021

    2021

  12. [20]

    The Singing Voice Conversion Challenge 2023

    Wen-Chin Huang, Lester Phillip Violeta, Songxiang Liu, Jiatong Shi, Yusuke Yasuda, and Tomoki Toda. The Singing Voice Conversion Challenge 2023 . arXiv, abs/2306.14422, 2023

    2023 arXiv

  13. [21]

    Jadoul, Bill Thompson, and Bart de Boer

    Y. Jadoul, Bill Thompson, and Bart de Boer. Introducing Parselmouth: A Python interface to Praat . J. Phonetics, 71: 0 1--15, 2018

    2018

  14. [22]

    Mega-TTS 2: Boosting Prompting Mechanisms for Zero-Shot Speech Synthesis

    Ziyue Jiang, Jinglin Liu, Yi Ren, Jinzheng He, Zhenhui Ye, Shengpeng Ji, Qian Yang, Chen Zhang, Pengfei Wei, Chunfeng Wang, Xiang Yin, Zejun Ma, and Zhou Zhao. Mega-TTS 2: Boosting Prompting Mechanisms for Zero-Shot Speech Synthesis . In ICLR , 2024

    2024

  15. [23]

    NaturalSpeech 3: Zero-Shot Speech Synthesis with Factorized Codec and Diffusion Models

    Zeqian Ju, Yuancheng Wang, Kai Shen, Xu Tan, Detai Xin, Dongchao Yang, Eric Liu, Yichong Leng, Kaitao Song, Siliang Tang, Zhizheng Wu, Tao Qin, Xiangyang Li, Wei Ye, Shikun Zhang, Jiang Bian, Lei He, Jinyu Li, and Sheng Zhao. NaturalSpeech 3: Zero-Shot Speech Synthesis with Fa...

    2024

  16. [24]

    Efficient Neural Audio Synthesis

    Nal Kalchbrenner, Erich Elsen, Karen Simonyan, Seb Noury, Norman Casagrande, Edward Lockhart, Florian Stimberg, A \" a ron van den Oord, Sander Dieleman, and Koray Kavukcuoglu. Efficient Neural Audio Synthesis . In ICML , pp.\ 2415--2424, 2018

    2018

  17. [25]

    STRAIGHT, exploitation of the other aspect of VOCODER: Perceptually isomorphic decomposition of speech sounds

    Hideki Kawahara. STRAIGHT, exploitation of the other aspect of VOCODER: Perceptually isomorphic decomposition of speech sounds . AST, pp.\ 349--353, 2006

    2006

  18. [26]

    Librivox: Free public domain audiobooks

    Jodi Kearns. Librivox: Free public domain audiobooks . Reference Reviews , 2014

    2014

  19. [27]

    PJS: phoneme-balanced Japanese singing-voice corpus

    Junya Koguchi, Shinnosuke Takamichi, and Masanori Morise. PJS: phoneme-balanced Japanese singing-voice corpus . In APSIPA , 2020

    2020

  20. [28]

    DiffWave: A Versatile Diffusion Model for Audio Synthesis

    Zhifeng Kong, Wei Ping, Jiaji Huang, Kexin Zhao, and Bryan Catanzaro. DiffWave: A Versatile Diffusion Model for Audio Synthesis . In ICLR , 2021

    2021

  21. [29]

    Mel-cepstral distance measure for objective speech quality assessment

    Robert Kubichek. Mel-cepstral distance measure for objective speech quality assessment. In Proceedings of IEEE Pacific Rim Conference on Communications Computers and Signal Processing, volume 1, pp.\ 125--128. IEEE, 1993

    1993

  22. [30]

    BigVGAN: A Universal Neural Vocoder with Large-Scale Training

    Sang - gil Lee, Wei Ping, Boris Ginsburg, Bryan Catanzaro, and Sungroh Yoon. BigVGAN: A Universal Neural Vocoder with Large-Scale Training . In ICLR , 2023

    2023

  23. [31]

    Dannenberg, Ruibo Liu, Wenhu Chen, Gus Xia, Yemin Shi, Wenhao Huang, Zili Wang, Yike Guo, and Jie Fu

    Yizhi Li, Ruibin Yuan, Ge Zhang, Yinghao Ma, Xingran Chen, Hanzhi Yin, Chenghao Xiao, Chenghua Lin, Anton Ragni, Emmanouil Benetos, Norbert Gyenge, Roger B. Dannenberg, Ruibo Liu, Wenhu Chen, Gus Xia, Yemin Shi, Wenhao Huang, Zili Wang, Yike Guo, and Jie Fu. MERT: Acoustic Mus...

    2024

  24. [32]

    DiffSinger: Singing Voice Synthesis via Shallow Diffusion Mechanism

    Jinglin Liu, Chengxi Li, Yi Ren, Feiyang Chen, and Zhou Zhao. DiffSinger: Singing Voice Synthesis via Shallow Diffusion Mechanism . In AAAI , 2022 a

    2022

  25. [33]

    Learning the Beauty in Songs: Neural Singing Voice Beautifier

    Jinglin Liu, Chengxi Li, Yi Ren, Zhiying Zhu, and Zhou Zhao. Learning the Beauty in Songs: Neural Singing Voice Beautifier . In ACL , 2022 b

    2022

  26. [34]

    DiffSVC: A Diffusion Probabilistic Model for Singing Voice Conversion

    Songxiang Liu, Yuewen Cao, Dan Su, and Helen Meng. DiffSVC: A Diffusion Probabilistic Model for Singing Voice Conversion . In ASRU , 2021

    2021

  27. [35]

    Decoupled Weight Decay Regularization

    Ilya Loshchilov and Frank Hutter. Decoupled Weight Decay Regularization . In ICLR, 2019

    2019

  28. [36]

    Brian McFee, Colin Raffel, Dawen Liang, Daniel P. W. Ellis, Matt McVicar, Eric Battenberg, and Oriol Nieto. librosa: Audio and Music Signal Analysis in Python . In SciPy, 2015

    2015

  29. [37]

    WORLD: a vocoder-based high-quality speech synthesis system for real-time applications

    Masanori Morise, Fumiya Yokomori, and Kenji Ozawa. WORLD: a vocoder-based high-quality speech synthesis system for real-time applications . IEICE Trans Inf Syst, 99 0 (7): 0 1877--1884, 2016

    2016

  30. [38]

    Tohoku kiritan singing database: A singing database for statistical parametric singing synthesis using japanese pop songs

    Itsuki Ogawa and Masanori Morise. Tohoku kiritan singing database: A singing database for statistical parametric singing synthesis using japanese pop songs . Acoustical Science and Technology, 2021

    2021

  31. [39]

    Audio Slicer , 2022

    Openvpi. Audio Slicer , 2022. URL https://github.com/openvpi/audio-slicer

    2022

  32. [40]

    DiffSinger Community Vocoders , 2024

    Openvpi. DiffSinger Community Vocoders , 2024. URL https://github.com/openvpi/vocoders

    2024

  33. [41]

    Transfer Learning of wav2vec 2.0 for Automatic Lyric Transcription

    Longshen Ou, Xiangming Gu, and Ye Wang. Transfer Learning of wav2vec 2.0 for Automatic Lyric Transcription . In ISMIR , 2022

    2022

  34. [42]

    Librispeech: An ASR corpus based on public domain audio books

    Vassil Panayotov, Guoguo Chen, Daniel Povey, and Sanjeev Khudanpur. Librispeech: An ASR corpus based on public domain audio books . In ICASSP , 2015

    2015

  35. [43]

    Waveglow: A Flow-based Generative Network for Speech Synthesis

    Ryan Prenger, Rafael Valle, and Bryan Catanzaro. Waveglow: A Flow-based Generative Network for Speech Synthesis . In ICASSP , pp.\ 3617--3621, 2019

    2019

  36. [44]

    Cox, Mark Hasegawa - Johnson, and Shiyu Chang

    Kaizhi Qian, Yang Zhang, Heting Gao, Junrui Ni, Cheng - I Lai, David D. Cox, Mark Hasegawa - Johnson, and Shiyu Chang. ContentVec: An Improved Self-Supervised Speech Representation by Disentangling Speakers . In ICML , 2022

    2022

  37. [45]

    Robust Speech Recognition via Large-Scale Weak Supervision

    Alec Radford, Jong Wook Kim, Tao Xu, Greg Brockman, Christine McLeavey, and Ilya Sutskever. Robust Speech Recognition via Large-Scale Weak Supervision . In ICML , 2023

    2023

  38. [46]

    Deepsinger: Singing voice synthesis with data mined from the web

    Yi Ren, Xu Tan, Tao Qin, Jian Luan, Zhou Zhao, and Tie - Yan Liu. Deepsinger: Singing voice synthesis with data mined from the web. In ACM SIGKDD , 2020

    2020

  39. [47]

    NHSS: A speech and singing parallel database

    Bidisha Sharma, Xiaoxue Gao, Karthika Vijayan, Xiaohai Tian, and Haizhou Li. NHSS: A speech and singing parallel database . Speech Commun. , 2021

    2021

  40. [48]

    NaturalSpeech 2: Latent Diffusion Models are Natural and Zero-Shot Speech and Singing Synthesizers

    Kai Shen, Zeqian Ju, Xu Tan, Eric Liu, Yichong Leng, Lei He, Tao Qin, Sheng Zhao, and Jiang Bian. NaturalSpeech 2: Latent Diffusion Models are Natural and Zero-Shot Speech and Singing Synthesizers . In ICLR , 2024

    2024

  41. [49]

    Muskits: an End-to-end Music Processing Toolkit for Singing Voice Synthesis

    Jiatong Shi, Shuai Guo, Tao Qian, Tomoki Hayashi, Yuning Wu, Fangzheng Xu, Xuankai Chang, Huazhe Li, Peter Wu, Shinji Watanabe, and Qin Jin. Muskits: an End-to-end Music Processing Toolkit for Singing Voice Synthesis . In Interspeech, 2022

    2022

  42. [50]

    Singing Voice Data Scaling-up: An Introduction to ACE-Opencpop and KiSing-v2

    Jiatong Shi, Yueqian Lin, Xinyi Bai, Keyi Zhang, Yuning Wu, Yuxun Tang, Yifeng Yu, Qin Jin, and Shinji Watanabe. Singing Voice Data Scaling-up: An Introduction to ACE-Opencpop and KiSing-v2 . CoRR , abs/2401.17619, 2024

    2024 arXiv

  43. [51]

    Benchmarks and leaderboards for sound demixing tasks, 2023

    Roman Solovyev, Alexander Stempkovskiy, and Tatiana Habruseva. Benchmarks and leaderboards for sound demixing tasks, 2023

    2023

  44. [52]

    HiFi-GAN: High-Fidelity Denoising and Dereverberation Based on Speech Deep Features in Adversarial Networks

    Jiaqi Su, Zeyu Jin, and Adam Finkelstein. HiFi-GAN: High-Fidelity Denoising and Dereverberation Based on Speech Deep Features in Adversarial Networks . In INTERSPEECH , pp.\ 4506--4510, 2020

    2020

  45. [53]

    SingMOS: An extensive Open-Source Singing Voice Dataset for MOS Prediction

    Yuxun Tang, Jiatong Shi, Yuning Wu, and Qin Jin. SingMOS: An extensive Open-Source Singing Voice Dataset for MOS Prediction . CoRR , abs/2406.10911, 2024

    2024 arXiv

  46. [54]

    Senior, and Koray Kavukcuoglu

    A \" a ron van den Oord, Sander Dieleman, Heiga Zen, Karen Simonyan, Oriol Vinyals, Alex Graves, Nal Kalchbrenner, Andrew W. Senior, and Koray Kavukcuoglu. WaveNet: A Generative Model for Raw Audio . In SSW , pp.\ 125. ISCA , 2016

    2016

  47. [55]

    Opencpop: A High-Quality Open Source Chinese Popular Song Corpus for Singing Voice Synthesis

    Yu Wang, Xinsheng Wang, Pengcheng Zhu, Jie Wu, Hanzhao Li, Heyang Xue, Yongmao Zhang, Lei Xie, and Mengxiao Bi. Opencpop: A High-Quality Open Source Chinese Popular Song Corpus for Singing Voice Synthesis . In Interspeech, 2022

    2022

  48. [56]

    SaMoye: Zero-shot Singing Voice Conversion Based on Feature Disentanglement and Synthesis

    Zihao Wang, Le Ma, Yan Liu, and Kejun Zhang. SaMoye: Zero-shot Singing Voice Conversion Based on Feature Disentanglement and Synthesis . CoRR , abs/2407.07728, 2024

    2024 arXiv

  49. [57]

    VocalSet: A Singing Voice Dataset

    Julia Wilkins, Prem Seetharaman, Alison Wahl, and Bryan Pardo. VocalSet: A Singing Voice Dataset . In ISMIR , 2018

    2018

  50. [58]

    HuggingFace's Transformers: State-of-the-art Natural Language Processing

    Thomas Wolf, Lysandre Debut, Victor Sanh, Julien Chaumond, Clement Delangue, Anthony Moi, Pierric Cistac, Tim Rault, R \' e mi Louf, Morgan Funtowicz, and Jamie Brew. HuggingFace's Transformers: State-of-the-art Natural Language Processing . CoRR , abs/1910.03771, 2019

    1910 arXiv

  51. [59]

    DDSP-based Singing Vocoders: A New Subtractive-based Synthesizer and A Comprehensive Evaluation

    Da - Yi Wu, Wen - Yi Hsiao, Fu - Rong Yang, Oscar Friedman, Warren Jackson, Scott Bruzenak, Yi - Wen Liu, and Yi - Hsuan Yang. DDSP-based Singing Vocoders: A New Subtractive-based Synthesizer and A Comprehensive Evaluation . In ISMIR, pp.\ 76--83, 2022

    2022

  52. [60]

    M4Singer: A Multi-Style, Multi-Singer and Musical Score Provided Mandarin Singing Corpus

    Lichao Zhang, Ruiqi Li, Shoutong Wang, Liqun Deng, Jinglin Liu, Yi Ren, Jinzheng He, Rongjie Huang, Jieming Zhu, Xiao Chen, and Zhou Zhao. M4Singer: A Multi-Style, Multi-Singer and Musical Score Provided Mandarin Singing Corpus . In NeurIPS , 2022

    2022

  53. [61]

    Leveraging Content-based Features from Multiple Acoustic Models for Singing Voice Conversion

    Xueyao Zhang, Yicheng Gu, Haopeng Chen, Zihao Fang, Lexiao Zou, Liumeng Xue, and Zhizheng Wu. Leveraging Content-based Features from Multiple Acoustic Models for Singing Voice Conversion . CoRR , abs/2310.11160, 2023

    2023 arXiv

  54. [62]

    Amphion: An Open-Source Audio, Music and Speech Generation Toolkit

    Xueyao Zhang, Liumeng Xue, Yicheng Gu, Yuancheng Wang, Jiaqi Li, Haorui He, Chaoren Wang, Ting Song, Xi Chen, Zihao Fang, Haopeng Chen, Junan Zhang, Tze Ying Tang, Lexiao Zou, Mingxuan Wang, Jun Han, Kai Chen, Haizhou Li, and Zhizheng Wu. Amphion: An Open-Source Audio, Music a...

    2024

  55. [63]

    SinTechSVS: A Singing Technique Controllable Singing Voice Synthesis System

    Junchuan Zhao, Low Qi Hong Chetwin, and Ye Wang. SinTechSVS: A Singing Technique Controllable Singing Voice Synthesis System . TASLP , 2024

    2024

  56. [64]

    FT-GAN: Fine-Grained Tune Modeling for Chinese Opera Synthesis

    Meizhen Zheng, Peng Bai, Xiaodong Shi, Xun Zhou, and Yiting Yan. FT-GAN: Fine-Grained Tune Modeling for Chinese Opera Synthesis . In AAAI , 2024

    2024

  57. [65]

    Dannenberg, Jie Fu, Chenghua Lin, Emmanouil Benetos, Wenhu Chen, Wei Xue, and Yike Guo

    Le Zhuo, Ruibin Yuan, Jiahao Pan, Yinghao Ma, Yizhi Li, Ge Zhang, Si Liu, Roger B. Dannenberg, Jie Fu, Chenghua Lin, Emmanouil Benetos, Wenhu Chen, Wei Xue, and Yike Guo. LyricWhiz: Robust Multilingual Zero-Shot Lyrics Transcription by Whispering to ChatGPT . In ISMIR , 2023

    2023

  58. [66]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

  59. [67]

    @esa (Ref

    \@ifxundefined[1] #1\@undefined \@firstoftwo \@secondoftwo \@ifnum[1] #1 \@firstoftwo \@secondoftwo \@ifx[1] #1 \@firstoftwo \@secondoftwo [2] @ #1 \@temptokena #2 #1 @ \@temptokena \@ifclassloaded agu2001 natbib The agu2001 class already includes natbib coding, so you should ...

  60. [68]

    \@lbibitem[] @bibitem@first@sw\@secondoftwo \@lbibitem[#1]#2 \@extra@b@citeb \@ifundefined br@#2\@extra@b@citeb \@namedef br@#2 \@nameuse br@#2\@extra@b@citeb \@ifundefined b@#2\@extra@b@citeb @num @parse #2 @tmp #1 NAT@b@open@#2 NAT@b@shut@#2 \@ifnum @merge>\@ne @bibitem@firs...

  61. [69]

    @open @close @open @close and [1] URL: #1 \@ifundefined chapter * \@mkboth \@ifxundefined @sectionbib * \@mkboth * \@mkboth\@gobbletwo \@ifclassloaded amsart * \@ifclassloaded amsbook * \@ifxundefined @heading @heading NAT@ctr thebibliography [1] @ \@biblabel @NAT@ctr \@bibset...

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.