Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
102 tokens/sec
GPT-4o
59 tokens/sec
Gemini 2.5 Pro Pro
43 tokens/sec
o3 Pro
6 tokens/sec
GPT-4.1 Pro
50 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

SpecDiff-GAN: A Spectrally-Shaped Noise Diffusion GAN for Speech and Music Synthesis (2402.01753v1)

Published 30 Jan 2024 in cs.SD, cs.LG, eess.AS, and eess.SP

Abstract: Generative adversarial network (GAN) models can synthesize highquality audio signals while ensuring fast sample generation. However, they are difficult to train and are prone to several issues including mode collapse and divergence. In this paper, we introduce SpecDiff-GAN, a neural vocoder based on HiFi-GAN, which was initially devised for speech synthesis from mel spectrogram. In our model, the training stability is enhanced by means of a forward diffusion process which consists in injecting noise from a Gaussian distribution to both real and fake samples before inputting them to the discriminator. We further improve the model by exploiting a spectrally-shaped noise distribution with the aim to make the discriminator's task more challenging. We then show the merits of our proposed model for speech and music synthesis on several datasets. Our experiments confirm that our model compares favorably in audio quality and efficiency compared to several baselines.

PDF HTML Abstract
Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Bookmark Chat Bubble Oval Streamline Icon: https://streamlinehq.com Chat (Pro)
Definition Search Book Streamline Icon: https://streamlinehq.com
References (32)
  1. J. Shen, R. Pang, R. J. Weiss, M. Schuster, N. Jaitly, Z. Yang et al., “Natural TTS synthesis by conditioning WaveNet on mel spectrogram predictions,” in Proc. ICASSP, 2018.
  2. Y. Jia, R. J. Weiss, F. Biadsy, W. Macherey, M. Johnson, Z. Chen et al., “Direct speech-to-speech translation with a sequence-to-sequence model,” in Proc. Interspeech, 2019.
  3. B. Sisman, J. Yamagishi, S. King, and H. Li, “An overview of voice conversion and its challenges: From statistical modeling to deep learning,” IEEE/ACM Trans. Audio, Speech, Lang. Process., vol. 29, pp. 132–157, 2020.
  4. J. Engel, K. K. Agrawal, S. Chen, I. Gulrajani, C. Donahue, and A. Roberts, “GANSynth: Adversarial neural audio synthesis,” arXiv preprint arXiv:1902.08710, 2019.
  5. S. Ji, J. Luo, and X. Yang, “A comprehensive survey on deep music generation: Multi-level representations, algorithms, evaluations, and future directions,” arXiv preprint arXiv:2011.06801, 2020.
  6. D. Moffat, R. Selfridge, and J. Reiss, “Sound effect synthesis,” in Foundations in Sound Design for Interactive Media, M. Filimowicz, Ed.   Routledge, 2019.
  7. C. Schreck, D. Rohmer, D. L. James, S. Hahmann, and M.-P. Cani, “Real-time sound synthesis for paper material based on geometric analysis,” in Proc. ACM SIGGRAPH/Eurographics SCA, Jul. 2016.
  8. A. van den Oord, S. Dieleman, H. Zen, K. Simonyan, O. Vinyals, A. Graves et al., “WaveNet: A generative model for raw audio,” in Proc. ISCA SSW, 2016.
  9. R. Prenger, R. Valle, and B. Catanzaro, “Waveglow: A flow-based generative network for speech synthesis,” in Proc. ICASSP, May 2019.
  10. I. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair et al., “Generative adversarial nets,” in Proc. NeurIPS, 2014.
  11. J. Kong, J. Kim, and J. Bae, “HiFi-GAN: Generative adversarial networks for efficient and high fidelity speech synthesis,” in Proc. NeurIPS, 2020.
  12. S. gil Lee, W. Ping, B. Ginsburg, B. Catanzaro, and S. Yoon, “BigVGAN: A universal neural vocoder with large-scale training,” in Proc. ICLR, 2023.
  13. N. Kodali, J. Hays, J. Abernethy, and Z. Kira, “On convergence and stability of GANs,” arXiv preprint arXiv:1705.07215, 2017.
  14. J. Ho, A. Jain, and P. Abbeel, “Denoising diffusion probabilistic models,” in Proc. NeurIPS, 2020.
  15. N. Chen, Y. Zhang, H. Zen, R. J. Weiss, M. Norouzi, and W. Chan, “WaveGrad: Estimating gradients for waveform generation,” in Proc. ICLR, 2021.
  16. S. gil Lee, H. Kim, C. Shin, X. Tan, C. Liu, Q. Meng et al., “PriorGrad: Improving conditional denoising diffusion models with data-dependent adaptive prior,” in Proc. ICLR, 2022.
  17. Z. Wang, H. Zheng, P. He, W. Chen, and M. Zhou, “Diffusion-GAN: Training GANs with diffusion,” in Proc. ICLR, 2023.
  18. Y. Koizumi, H. Zen, K. Yatabe, N. Chen, and M. Bacchiani, “SpecGrad: Diffusion probabilistic model based neural vocoder with adaptive noise spectral shaping,” in Proc. Interspeech, 2022.
  19. X. Wu, “Enhancing unsupervised speech recognition with diffusion GANs,” in Proc. ICASSP, Jun. 2023.
  20. W. Jang, D. C. Y. Lim, J. Yoon, B. Kim, and J. Kim, “UnivNet: A neural vocoder with multi-resolution spectrogram discriminators for high-fidelity waveform generation,” in Proc. Interspeech, 2021.
  21. C. Wang, C. Zeng, and X. He, “HiFi-WaveGAN: Generative adversarial network with auxiliary spectrogram-phase loss for high-fidelity singing voice generation,” arXiv preprint arXiv:2210.12740, 2022.
  22. T. Karras, M. Aittala, J. Hellsten, S. Laine, J. Lehtinen, and T. Aila, “Training generative adversarial networks with limited data,” in Proc. NeurIPS, 2020.
  23. K. Ito and L. Johnson, “The LJ speech dataset,” https://keithito.com/LJ-Speech-Dataset/, 2017.
  24. J. Yamagishi, C. Veaux, and K. MacDonald, “CSTR VCTK Corpus: English multi-speaker corpus for CSTR voice cloning toolkit (version 0.92),” University of Edinburgh. The Centre for Speech Technology Research (CSTR), Tech. Rep., 2019.
  25. V. Emiya, N. Bertin, B. David, and R. Badeau, “MAPS - a piano database for multipitch estimation and automatic transcription of music,” INRIA, Tech. Rep., Jul. 2010.
  26. O. Gillet and G. Richard, “ENST-Drums: An extensive audio-visual database for drum signals processing,” in Proc. ISMIR, 2006.
  27. A. Rix, J. Beerends, M. Hollier, and A. Hekstra, “Perceptual evaluation of speech quality (PESQ) - a new method for speech quality assessment of telephone networks and codecs,” in Proc. ICASSP, vol. 2, 2001, pp. 749–752.
  28. C. H. Taal, R. C. Hendriks, R. Heusdens, and J. Jensen, “A short-time objective intelligibility measure for time-frequency weighted noisy speech,” in Proc. ICASSP, 2010, pp. 4214–4217.
  29. W. A. Jassim, J. Skoglund, M. Chinen, and A. Hines, “Warp-Q: Quality prediction for generative neural speech codecs,” Proc. ICASSP, pp. 401–405, 2021.
  30. K. Kilgour, M. Zuluaga, D. Roblek, and M. Sharifi, “Fréchet audio distance: A reference-free metric for evaluating music enhancement algorithms,” in Proc. Interspeech, 2019.
  31. S. Hershey, S. Chaudhuri, D. P. W. Ellis, J. F. Gemmeke, A. Jansen, R. C. Moore et al., “CNN architectures for large-scale audio classification,” in Proc. ICASSP, 2017.
  32. L. Ziyin, T. Hartwig, and M. Ueda, “Neural networks fail to learn periodic functions and how to fix it,” in Proc. NeurIPS, 2020.
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (5)
  1. Teysir Baoueb (4 papers)
  2. Haocheng Liu (3 papers)
  3. Mathieu Fontaine (15 papers)
  4. Jonathan Le Roux (82 papers)
  5. Gael Richard (14 papers)
Citations (1)
View on Semantic Scholar