Papers › CleanMel: Mel-Spectrogram Enhancement for Improving Both Speech Quality and ASR
CleanMel: Mel-Spectrogram Enhancement for Improving Both Speech Quality and ASR
Nian Shao, Rui Zhou, Pengyu Wang, Xian Li, Ying Fang, Yujie Yang, Xiaofei Li
In this work, we propose CleanMel, a single-channel Mel-spectrogram denoising and dereverberation network for improving both speech quality and automatic speech recognition (ASR) performance. The proposed network takes as input the noisy and reverberant microphone recording and predicts the corresponding clean Mel-spectrogram. The enhanced Mel-spectrogram can be either transformed to speech waveform with a neural vocoder or directly used for ASR. The proposed network is composed of interleaved cross-band and narrow-band processing in the Mel-frequency domain, for learning the full-band spectral pattern and the narrow-band properties of signals, respectively. Compared to linear-frequency domain or time-domain speech enhancement, the key advantage of Mel-spectrogram enhancement is that Mel-frequency presents speech in a more compact way and thus is easier to learn, which will benefit both speech quality and ASR. Experimental results on four English and one Chinese datasets demonstrate a significant improvement in both speech quality and ASR performance achieved by the proposed model. Code and audio examples of our model are available online in https://audio.westlake.edu.cn/Research/CleanMel.html.
Code
Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.
Code Syntology ran Syntology
Not run by Syntology. Nothing on this page verifies that the listed code works.
Tasks
Results from the paper archive 2025-07-28
| Task | Dataset | Model | Metric | Value | Rank at snapshot | Leaderboard | Report |
|---|---|---|---|---|---|---|---|
| Automatic Speech Recognition (ASR) | RealMAN | CleanMel-L-mask | CER | 14.4 | #1 of 2 | Archive leaderboard | report |
| Speech Enhancement | RealMAN | CleanMel-L-map | DNSMOS | 3.82 | #1 of 2 | Archive leaderboard | report |
| Speech Enhancement | RealMAN | CleanMel-L-map | DNSMOS BAK | 4.03 | #1 of 2 | Archive leaderboard | report |
| Speech Enhancement | RealMAN | CleanMel-L-map | DNSMOS OVRL | 3.25 | #1 of 2 | Archive leaderboard | report |
| Speech Enhancement | RealMAN | CleanMel-L-map | DNSMOS SIG | 3.55 | #1 of 2 | Archive leaderboard | report |
| Speech Enhancement | RealMAN | CleanMel-L-map | PESQ-WB | 2.10 | #1 of 2 | Archive leaderboard | report |
Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.
Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections