Research map / Image generation / GANs
GAN vocoders and audio synthesis
Synthesize speech and audio waveforms with adversarial networks.
13 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2018–2024) · an approach in GANs, keyword cluster “synthesis · latent · face”
State History Tried Preprints Who Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
3 papers in 2023–24 and 0 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 5.2% to 0.0%.
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2018 · MuseGAN: Multi-track Sequential Generative Adversarial Networks for Symbolic Music Generation and Accompaniment (AAAI 2018 · 441 citations)
- 2019 · MelGAN: Generative Adversarial Networks for Conditional Waveform Synthesis (NeurIPS 2019 · 602 citations)
- 2020 · HiFi-GAN: Generative Adversarial Networks for Efficient and High Fidelity Speech Synthesis (NeurIPS 2020 · 731 citations)
- 2021 · CardioGAN: Attentive Generative Adversarial Network with Dual Discriminators for Synthesis of ECG from PPG (AAAI 2021 · 87 citations)
- 2022 · Chunked Autoregressive GAN for Conditional Waveform Synthesis (ICLR 2022 · 6 citations)
- 2023 · BigVGAN: A Universal Neural Vocoder with Large-Scale Training (ICLR 2023 · 46 citations)
- 2024 · FINALLY: fast and universal speech enhancement with studio-like quality (NeurIPS 2024 · 6 citations)
Earliest papers here
- MuseGAN: Multi-track Sequential Generative Adversarial Networks for Symbolic Music Generation and Accompaniment (AAAI 2018 · 441 citations)
- MelGAN: Generative Adversarial Networks for Conditional Waveform Synthesis (NeurIPS 2019 · 602 citations)
- GANSynth: Adversarial Neural Audio Synthesis (ICLR 2019 · 277 citations)
What has been tried
13 papers here report 136 measured changes (a component added, removed or swapped, compared with a baseline): helped 68%, no clear effect 6%, hurt 25%, mixed 1%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Use LMOS regression loss rather than no regression loss | VoxCeleb (real data) · MOS | 3.71 ± 0.08 -> 4.20 ± 0.05 | Helped | FINALLY: fast and universal speech enhancement with studio-like quality (NeurIPS 2024 · 6 citations) |
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 3 papers; from OpenAlex and the papers' affiliations).
- Samsung · 2 papers
Use it
Download BibTeX of 13 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Large-scale GAN latent analysis (108 papers)
- GAN-based image manipulation (90 papers)
- Conditional GANs (25 papers)
Elsewhere in the topic
- Stabilized GAN training (144 papers)
- Generative adversarial network theory (62 papers)
- GAN latent optimization and evaluation (52 papers)
- GANs for text generation (21 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).