BIGVSAN: Enhancing Gan-Based Neural Vocoders with Slicing Adversarial Network
The DOI matches its Crossref record; 39 of the 39 references it cites were found in a registry or an index.
Abstract
Generative adversarial network (GAN)-based vocoders have been intensively studied because they can synthesize high-fidelity audio waveforms faster than real-time. However, it has been reported that most GANs fail to obtain the optimal projection for discriminating between real and fake data in the feature space. In the literature, it has been demonstrated that slicing adversarial network (SAN), an improved GAN training framework that can find the optimal projection, is effective in the image generation task. In this paper, we investigate the effectiveness of SAN in the vocoding task. For this purpose, we propose a scheme to modify least-squares GAN, which most GAN-based vocoders adopt, so that their loss functions satisfy the requirements of SAN. Through our experiments, we demonstrate that SAN can improve the performance of GAN-based vocoders, including BigVGAN, with small modifications. Our code is available at https://github.com/sony/bigvsan.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“In recent years, research [44,45,46,47,48,49,50,51] has continued to refine these technologies, enabling synthesised speech to sound increasingly natural while also improving synthesis speed.”
From Audio Deepfake Detection: What Has Been Achieved and What Lies Ahead (2025) · cites this paper as [50]
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
A Style-Based Generator Architecture for Generative Adversarial Networks
Karras, T., Laine, S., & Aila, T. (2019) · 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2019.00453 · Cited by 11,714 · cited 1×
“GAN was first applied to image generation [23], and various GANs have achieved impressive results in not only image generation [27, 28] but also vocoding tasks.”
Method“Least squares generative adversarial networks,”
Mao, X., Li, Q., Xie, H., Lau, R. Y. K., Wang, Z., & Smolley, S. P. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.304 · Cited by 5,437 · cited 3×
“It enhances discriminators by making their last linear projections more discriminative with only small modifications to least-squares GAN [22].”
BackgroundMethodInfluentialWaveNet: a generative model for raw audio
Oord, A. V. D., Dieleman, S., Zen, H., Simonyan, K., Vinyals, O., Graves, A., Kalchbrenner, N., Senior, A., & Kavukcuoglu, K. (2016) · arXiv · doi:10.48550/arxiv.1609.03499 · Cited by 4,911 · cited 1×
“Deep generative models have demonstrated a particularly high performance in synthesizing waveforms, including autoregressive models [3, 4], generative adversarial network (GAN)-based models [5, 6, 7], flow-based models [8, 9, 10, 11, 12, 13], and diffusion-based models [14, 15, 16, 17, 18].”
BackgroundRix, A. W., Beerends, J. G., Hollier, M. P., & Hekstra, A. P. (2002) · 2001 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings (Cat. No.01CH37221) · doi:10.1109/icassp.2001.941023 · Cited by 3,149 · cited 1×
“…STFT (M-STFT) [6], which measures the spectral distance across multiple resolutions 2 , 2) perceptual evaluation of speech quality (PESQ) [34], which is a widely adopted automated assessment of voice quality 3 , 3) mel-cepstral distortion (MCD) [35] with dynamic time warping, which…”
MethodNatural TTS Synthesis by Conditioning Wavenet on MEL Spectrogram Predictions
Shen, J., Pang, R., Weiss, R. J., Schuster, M., Jaitly, N., Yang, Z., Chen, Z., Zhang, Y., Wang, Y., Skerrv-Ryan, R., et al. (2018) · 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp.2018.8461368 · Cited by 2,791
Arjovsky, M., Chintala, S., & Bottou, L. (2017) · arXiv · doi:10.48550/arxiv.1701.07875 · Cited by 2,225 · cited 1×
“Regarding the choices of R i : R → R ( i = 1 , 2 , 3) , there are several GAN variants that can be used [23, 29, 30].”
BackgroundSEGAN: Speech Enhancement Generative Adversarial Network
Pascual, S., Bonafonte, A., & Serrà, J. (2017) · Interspeech 2017 · doi:10.21437/interspeech.2017-1428 · Cited by 1,297 · cited 2×
“Indeed, most previous speech-related models adopted least-squares GAN because of its training stability [25, 26, 6, 7], and this is what motivates us to create a SAN counterpart of least-squares GAN.”
Methodf-GAN: Training Generative Neural Samplers using Variational Divergence Minimization
Nowozin, S., Cseke, B., & Tomioka, R. (2016) · arXiv · doi:10.48550/arxiv.1606.00709 · Cited by 1,183 · cited 1×
“Regarding the choices of R i : R → R ( i = 1 , 2 , 3) , there are several GAN variants that can be used [23, 29, 30].”
Background“Waveglow: a flow-based generative network for speech synthesis,”
Prenger, R., Valle, R., & Catanzaro, B. (2019) · ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp.2019.8683143 · Cited by 1,004
Yamamoto, R., Song, E., & Kim, J.-M. (2020) · ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp40776.2020.9053795 · Cited by 827 · cited 6×
“Indeed, most previous models adopted least-squares GAN because of its training stability [25, 26, 6, 7], and this is what motivates us to create a SAN counterpart of least-squares GAN.”
BackgroundMethodInfluential
Research works printed but not linked: 3
- Ren, Fast-Speech 2: Fast and high-quality end-to-end text to speech
- 10.5555/2969033.2969125 · doi:10.5555/2969033.2969125
- Steinmetz, auraloss: Audio focused loss functions in PyTorch
Cited by
Audio Deepfake Detection: What Has Been Achieved and What Lies Ahead
Zhang, B., Cui, H., Nguyen, V., & Whitty, M. (2025) · Sensors · doi:10.3390/s25071989 · Cited by 72
Multi-Scale Sub-Band Constant-Q Transform Discriminator for High-Fidelity Vocoder
Gu, Y., Zhang, X., Xue, L., & Wu, Z. (2024) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp48485.2024.10448436 · Cited by 36
“Deep generative models including autoregressive-based [1, 2], flow-based [3, 4], GAN-based [5–11], and diffusion-based [12, 13] models have been successful for this task.”
BackgroundRey-Paredes, M., Pérez, C. J., & Mateos-Caballero, A. (2024) · IEEE Open Journal of the Computer Society · doi:10.1109/ojcs.2024.3504864 · Cited by 20
“BigVSAN model was trained in the work of [37] for 10 million steps on the LibriTTS dataset [42] at a sampling rate of 24 kHz taking a full 100-band mel-spectrogram as input.”
Comunità, M., Zhong, Z., Takahashi, A., Yang, S., Zhao, M., Saito, K., Ikemiya, Y., Shibuya, T., Takahashi, S., & Mitsufuji, Y. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2406.17672 · Cited by 13
“3.1, the standard Mel-spectrogram transform from vocoders [27,28] is utilized, which transforms a 10-second audio clip at sampling rate 22.05kHz into 848 frames with 80 Mel bins.”
MethodInfluentialSelf-citationAdvancements and Applications of Generative Adversarial Networks: A Comprehensive Review
Pawar, D. (2024) · International Journal for Research in Applied Science and Engineering Technology · doi:10.22214/ijraset.2024.62148 · Cited by 10
Metadata
- Authors
- Shibuya, T., Takida, Y., & Mitsufuji, Y.
- Affiliations
- Sony (Japan) (JP)
- Venue
- ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
- Publisher
- IEEE
- Published
- 2024-04-14
- Pages
- 10121-10125
- ISSN
- 1520-6149 (ISSN-L)
- OpenAlex
- W4392931580
- Cited by
- 14
- References
- 38
- Citations by year
- 2026: 3 · 2025: 3 · 2024: 8
- FWCI
- 4.13
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- IEEE International Conference on Acoustics Speech and Signal Processing (publishedVersion)
- Funders
- National Institute of Advanced Industrial Science and Technology
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Speech Recognition and Synthesis
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Generative Adversarial Networks and Image Synthesis
- Physical Sciences › Computer Science › Signal Processing › Speech and Audio Processing
- Keywords
- sound synthesis, adversarial training, loss function modification
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: BIGVSAN: Enhancing Gan-Based Neural Vocoders with Slicing Adversarial Network. https://app.cochat.ai/papers/10.1109/icassp48485.2024.10446121
- The paper
- Shibuya, T., Takida, Y., & Mitsufuji, Y. (2024). BIGVSAN: Enhancing Gan-Based Neural Vocoders with Slicing Adversarial Network. ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). https://doi.org/10.1109/icassp48485.2024.10446121
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.