Personalized Speech Enhancement on DNS Track 1: Headset 5 (test)
4.75SIG ScoreUniSE + GSPO
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| UniSE + GSPOPost-training=GSPO2026.06 | 4.75 | 4.76 | 4.63 | |
| GenSE + GSPOPost-training=GSPO2026.06 | 4.65 | 4.61 | 4.45 | |
| UniSE2026.06 | 4.52 | 4.43 | 4.21 | |
| UniFlowDDPM2026.06 | 4.24 | 3.99 | 3.73 | |
| LLaSE-G12026.06 | 4.21 | 3.99 | 3.72 | |
| UniFlowFM2026.06 | 4.2 | 4.01 | 3.7 | |
| UniFlowMF2026.06 | 4.18 | 3.99 | 3.67 | |
| NoisyReference Speaker Embedding=-, On-the-fly Speaker Embedding=-2026.01 | 4.15 | 2.37 | 2.71 | |
| Noisy2026.06 | 4.15 | 2.37 | 2.71 | |
| GenSE2026.06 | 4.13 | 3.61 | 3.41 | |
| TEA-PSEVersion=3.02026.06 | 4.12 | 4.05 | 3.65 | |
| pDeepFilterNet2 (oracle similarity)Reference Speaker Embedding=ECAPA, On-the-fly Speaker Embedding=ECAPA, scaling factor alpha=2, Params (M)=17.67, MACs (G)=1.752026.01 | 3.85 | 3.33 | 3.08 | |
| E3NetReference Speaker Embedding=ECAPA, On-the-fly Speaker Embedding=-, Params (M)=6.62, MACs (G)=11.082026.01 | 3.82 | 3.47 | 3.1 | |
| pDeepFilterNet2 (oracle similarity w/o scaling)Reference Speaker Embedding=ECAPA, On-the-fly Speaker Embedding=ECAPA, scaling factor alpha=1, Params (M)=17.67, MACs (G)=1.752026.01 | 3.82 | 3.43 | 3.11 | |
| pDeepFilterNet2 (light similarity)Reference Speaker Embedding=ECAPA, On-the-fly Speaker Embedding=TinyECAPA, scaling factor alpha=6, Params (M)=2.38, MACs (G)=0.372026.01 | 3.81 | 3.49 | 3.13 | |
| NAPSE2026.06 | 3.81 | 3.99 | 3.38 | |
| pDeepFilterNet2Reference Speaker Embedding=ECAPA, On-the-fly Speaker Embedding=-, Params (M)=2.23, MACs (G)=0.332026.01 | 3.77 | 3.47 | 3.08 | |
| pDCCRNReference Speaker Embedding=ECAPA, On-the-fly Speaker Embedding=-, Params (M)=4.50, MACs (G)=5.592026.01 | 3.76 | 3.29 | 2.99 | |
| pDeepFilterNet2 (light similarity w/o scaling)Reference Speaker Embedding=ECAPA, On-the-fly Speaker Embedding=TinyECAPA, scaling factor alpha=1, Params (M)=2.38, MACs (G)=0.372026.01 | 3.67 | 3.6 | 3.06 |