Speech Enhancement On Demand
Metrics
PESQ (wb)
Para. (M)
Results
Performance results of various models on this benchmark
Comparison Table
| Model Name | PESQ (wb) | Para. (M) |
|---|---|---|
| monaural-speech-enhancement-with-complex | 3.43 | 0.86 |
| deepfilternet-perceptually-motivated-real | 3.17 | - |
| let-ssms-be-convnets-state-space-modeling | 3.25 | - |
| perceptual-contrast-stretching-on-target | 3.35 | - |
| metricgan-okd-multi-metric-optimization-of | 3.24 | 1.89 |
| an-analysis-of-the-variance-of-diffusion | 3.11 | - |
| deep-residual-dense-lattice-network-for | 2.93 | - |
| the-pesqetarian-on-the-relevance-of-goodhart | 3.82 | 30 |
| Model 9 | 3.61 | 2.04 |
| manner-multi-view-attention-network-for-noise | 3.21 | - |
| dense-tsnet-dense-connected-two-stage | 3.05 | 0.014 |
| an-investigation-of-incorporating-mamba-for | 3.69 | 2.25 |
| rose-a-recognition-oriented-speech | 3.01 | 36.98 |
| xlstm-senet-xlstm-for-single-channel-speech | 3.53 | 2.27 |
| scp-gan-self-correcting-discriminator | 3.52 | - |
| raw-speech-enhancement-with-deep-state-space | 3.27 | - |
| speech-enhancement-and-dereverberation-with | 2.93 | - |
| a-modulation-domain-loss-for-neural-network-1 | 2.82 | - |
| Model 19 | 3.63 | 2.04 |
| multi-view-attention-transfer-for-efficient | 3.12 | 1.38 |
| metricgan-okd-multi-metric-optimization-of | 3.12 | 0.82 |
| deep-residual-dense-lattice-network-for | 3.02 | - |
| Model 23 | 3.54 | - |
| metricgan-an-improved-version-of-metricgan | 3.15 | - |
| real-time-speech-enhancement-in-the-waveform | 3.07 | - |
| cmgan-conformer-based-metric-gan-for-monaural | 3.41 | - |
| d2net-a-denoising-and-dereverberation-network | 3.27 | - |
| metricgan-generative-adversarial-networks | 2.86 | - |
| deep-residual-dense-lattice-network-for | 2.94 | - |
| explicit-estimation-of-magnitude-and-phase | 3.60 | 2.26 |
| end-to-end-speech-enhancement-based-on | 2.7 | - |
| investigating-training-objectives-for | 3.70 | - |
| deep-residual-dense-lattice-network-for | 2.84 | - |
| mamba-seunet-mamba-unet-for-monaural-speech | 3.73 | 6.28 |
| improving-perceptual-quality-by-phone | 3.15 | - |
| real-time-speech-enhancement-in-the-waveform | 2.93 | - |
| fspen-an-ultra-lightweight-network-for-real | 2.97 | 0.079 |
| primek-net-multi-scale-spectral-learning-via | 3.61 | 1.41 |
| perceptual-loss-based-speech-denoising-with | 3.17 | - |
| boosting-self-supervised-embeddings-for | 3.20 | - |