Salvato in:
| Autori principali: | , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2509.19495 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866908556410748928 |
|---|---|
| author | Chhaglani, Bhawana Gao, Yang Richter, Julius Li, Xilin Zadissa, Syavosh Pruthi, Tarun Lovitt, Andrew |
| author_facet | Chhaglani, Bhawana Gao, Yang Richter, Julius Li, Xilin Zadissa, Syavosh Pruthi, Tarun Lovitt, Andrew |
| contents | Diffusion-based speech enhancement (SE) achieves natural-sounding speech and strong generalization, yet suffers from key limitations like generative artifacts and high inference latency. In this work, we systematically study artifact prediction and reduction in diffusion-based SE. We show that variance in speech embeddings can be used to predict phonetic errors during inference. Building on these findings, we propose an ensemble inference method guided by semantic consistency across multiple diffusion runs. This technique reduces WER by 15% in low-SNR conditions, effectively improving phonetic accuracy and semantic plausibility. Finally, we analyze the effect of the number of diffusion steps, showing that adaptive diffusion steps balance artifact suppression and latency. Our findings highlight semantic priors as a powerful tool to guide generative SE toward artifact-free outputs. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_19495 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | ArtiFree: Detecting and Reducing Generative Artifacts in Diffusion-based Speech Enhancement Chhaglani, Bhawana Gao, Yang Richter, Julius Li, Xilin Zadissa, Syavosh Pruthi, Tarun Lovitt, Andrew Sound Artificial Intelligence Diffusion-based speech enhancement (SE) achieves natural-sounding speech and strong generalization, yet suffers from key limitations like generative artifacts and high inference latency. In this work, we systematically study artifact prediction and reduction in diffusion-based SE. We show that variance in speech embeddings can be used to predict phonetic errors during inference. Building on these findings, we propose an ensemble inference method guided by semantic consistency across multiple diffusion runs. This technique reduces WER by 15% in low-SNR conditions, effectively improving phonetic accuracy and semantic plausibility. Finally, we analyze the effect of the number of diffusion steps, showing that adaptive diffusion steps balance artifact suppression and latency. Our findings highlight semantic priors as a powerful tool to guide generative SE toward artifact-free outputs. |
| title | ArtiFree: Detecting and Reducing Generative Artifacts in Diffusion-based Speech Enhancement |
| topic | Sound Artificial Intelligence |
| url | https://arxiv.org/abs/2509.19495 |