ICASSP 2026 URGENT Challenge - Track 1: Universal Speech Enhancement
Organized by urgent - Current server time: Jan. 1, 2026, 12:28 p.m. UTC
First phase
Validation phase
Sept. 20, 2025, noon UTC
End
Competition Ends
Nov. 8, 2025, noon UTC
First phase
Validation phaseSept. 20, 2025, noon UTC
End
Competition EndsNov. 8, 2025, noon UTC
Validation phase
Start: Sept. 20, 2025, noon
Description: Please follow the instructions in https://urgent-challenge.github.io/urgent2026/track1/#submission to prepare the zip file for submission.
Non-blind test phase
Start: Oct. 14, 2025, noon
Description: Please follow the instructions in https://urgent-challenge.github.io/urgent2026/track1/#submission to prepare the zip file for submission.
Blind test phase
Start: Nov. 4, 2025, noon
Description: Please follow the instructions in https://urgent-challenge.github.io/urgent2026/#submission to prepare the zip file for submission.
Competition Ends
Nov. 8, 2025, noon
You must be logged in to participate in competitions.
Sign InStage-1: Final ranking on objective metrics (click to expand/collapse)
Download Table
| Results | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| # | User | Non-intrusive SE metrics | Intrusive SE metrics | Downstream-task-independent metrics | Downstream-task-dependent metrics | Overall ranking score | |||||||||
| DNSMOS | NISQA | UTMOS | SCOREQ | PESQ | ESTOI | POLQA | SpeechBERTScore | LPS | SpkSim | EmoSim | LID | CAcc (%) | 1 |
WR |
2.87 (5) | 3.49 (4) | 2.48 (5) | 3.52 (2) | 2.71 (1) | 0.87 (1) | 3.46 (1) | 0.88 (1) | 0.75 (3) | 0.78 (2) | 0.99 (1) | 0.87 (2) | 69.69 (1) | 2.12 | 2 |
GHW |
2.76 (11) | 3.40 (7) | 2.46 (6) | 3.41 (3) | 2.69 (2) | 0.86 (2) | 3.46 (1) | 0.87 (2) | 0.72 (6) | 0.77 (3) | 0.99 (1) | 0.86 (3) | 67.66 (5) | 3.85 | 3 |
subatomicseer |
2.71 (13) | 3.26 (10) | 2.14 (9) | 3.16 (7) | 2.60 (3) | 0.86 (2) | 3.42 (3) | 0.87 (2) | 0.76 (2) | 0.78 (2) | 0.99 (1) | 0.88 (1) | 69.46 (2) | 3.98 | 4 |
baird |
2.85 (6) | 3.42 (6) | 2.51 (4) | 3.52 (2) | 2.50 (8) | 0.86 (2) | 3.33 (5) | 0.87 (2) | 0.72 (6) | 0.77 (3) | 0.99 (1) | 0.85 (4) | 67.13 (7) | 4.31 | 5 |
RICK2000 |
2.63 (15) | 3.01 (13) | 2.07 (12) | 2.97 (10) | 2.54 (6) | 0.86 (2) | 3.43 (2) | 0.88 (1) | 0.74 (4) | 0.78 (2) | 0.99 (1) | 0.88 (1) | 69.69 (1) | 4.90 | 6 |
Ali-Universal-SE |
2.83 (8) | 3.17 (11) | 2.28 (7) | 3.18 (6) | 2.31 (10) | 0.82 (5) | 3.14 (9) | 0.87 (2) | 0.75 (3) | 0.78 (2) | 0.99 (1) | 0.88 (1) | 69.35 (3) | 5.06 | 7 |
camssly |
2.89 (3) | 3.40 (7) | 2.09 (11) | 3.10 (8) | 2.52 (7) | 0.83 (4) | 3.18 (7) | 0.86 (3) | 0.71 (7) | 0.74 (6) | 0.99 (1) | 0.87 (2) | 67.22 (6) | 5.50 | 8 |
Jyxiong |
2.87 (5) | 3.31 (9) | 2.09 (11) | 3.01 (9) | 2.55 (5) | 0.82 (5) | 3.25 (6) | 0.86 (3) | 0.73 (5) | 0.74 (6) | 0.99 (1) | 0.86 (3) | 66.23 (10) | 5.71 | 9 |
rc |
2.68 (14) | 2.81 (16) | 2.04 (13) | 2.80 (13) | 2.57 (4) | 0.87 (1) | 3.40 (4) | 0.88 (1) | 0.77 (1) | 0.79 (1) | 0.99 (1) | 0.85 (4) | 63.58 (16) | 5.88 | 10 |
inverseai |
2.88 (4) | 3.78 (2) | 2.85 (1) | 3.67 (1) | 2.15 (13) | 0.81 (6) | 2.94 (14) | 0.84 (5) | 0.69 (9) | 0.70 (8) | 0.99 (1) | 0.79 (7) | 64.30 (14) | 6.88 | 11 |
ai_se |
2.84 (7) | 3.34 (8) | 2.27 (8) | 3.25 (5) | 2.40 (9) | 0.83 (4) | 3.15 (8) | 0.84 (5) | 0.68 (10) | 0.75 (5) | 0.99 (1) | 0.84 (5) | 63.96 (15) | 7.00 | 12 |
abid |
3.08 (1) | 4.18 (1) | 2.73 (2) | 3.52 (2) | 1.91 (17) | 0.78 (8) | 2.72 (17) | 0.82 (6) | 0.70 (8) | 0.71 (7) | 0.99 (1) | 0.78 (8) | 65.87 (12) | 7.38 | 13 |
Baseline_BSRNN |
2.63 (15) | 2.77 (17) | 2.01 (15) | 2.79 (14) | 2.31 (10) | 0.84 (3) | 3.10 (10) | 0.87 (2) | 0.73 (5) | 0.76 (4) | 0.99 (1) | 0.85 (4) | 67.04 (8) | 7.67 | 14 |
4Speech |
2.63 (15) | 2.73 (18) | 1.99 (17) | 2.80 (13) | 2.29 (12) | 0.84 (3) | 3.08 (12) | 0.87 (2) | 0.73 (5) | 0.76 (4) | 0.99 (1) | 0.87 (2) | 67.90 (4) | 7.75 | 15 |
xueke |
2.54 (18) | 2.72 (19) | 1.98 (18) | 2.79 (14) | 2.31 (10) | 0.84 (3) | 3.09 (11) | 0.87 (2) | 0.73 (5) | 0.76 (4) | 0.99 (1) | 0.88 (1) | 66.21 (11) | 8.25 | 16 |
seeee |
2.62 (16) | 2.77 (17) | 2.01 (15) | 2.78 (15) | 2.30 (11) | 0.83 (4) | 3.09 (11) | 0.87 (2) | 0.72 (6) | 0.74 (6) | 0.99 (1) | 0.85 (4) | 66.55 (9) | 8.35 | 17 |
Yuhan_Wei |
2.62 (16) | 2.82 (15) | 2.00 (16) | 2.74 (16) | 2.30 (11) | 0.83 (4) | 2.96 (13) | 0.87 (2) | 0.72 (6) | 0.71 (7) | 0.99 (1) | 0.84 (5) | 65.70 (13) | 8.90 | 18 |
SPEAR |
2.74 (12) | 3.46 (5) | 2.13 (10) | 2.88 (12) | 1.93 (16) | 0.80 (7) | 2.66 (18) | 0.85 (4) | 0.65 (11) | 0.67 (10) | 0.99 (1) | 0.80 (6) | 59.32 (22) | 10.17 | 19 |
sigma |
3.03 (2) | 3.77 (3) | 2.69 (3) | 3.34 (4) | 1.43 (20) | 0.30 (12) | 2.03 (20) | 0.75 (7) | 0.59 (14) | 0.40 (13) | 0.98 (2) | 0.78 (8) | 59.78 (21) | 10.46 | 20 |
miku8023 |
2.79 (9) | 2.94 (14) | 1.90 (20) | 2.79 (14) | 2.09 (14) | 0.80 (7) | 2.82 (16) | 0.84 (5) | 0.68 (10) | 0.69 (9) | 0.99 (1) | 0.86 (3) | 62.78 (19) | 10.52 | 21 |
XMU-SPEECH |
2.53 (19) | 2.73 (18) | 1.93 (19) | 2.71 (17) | 2.15 (13) | 0.82 (5) | 2.90 (15) | 0.86 (3) | 0.70 (8) | 0.71 (7) | 0.99 (1) | 0.85 (4) | 63.05 (18) | 10.56 | 22 |
Baseline_BSRNN-Flow |
2.77 (10) | 3.09 (12) | 2.02 (14) | 2.92 (11) | 1.97 (15) | 0.80 (7) | 2.82 (16) | 0.82 (6) | 0.63 (13) | 0.70 (8) | 0.99 (1) | 0.79 (7) | 58.33 (23) | 10.92 | 23 |
Benjamin_Glidden |
2.57 (17) | 2.69 (20) | 1.86 (21) | 2.65 (18) | 1.89 (18) | 0.77 (9) | 2.53 (19) | 0.82 (6) | 0.64 (12) | 0.67 (10) | 0.99 (1) | 0.84 (5) | 61.17 (20) | 13.08 | 24 |
Salute |
2.35 (20) | 2.44 (21) | 1.70 (22) | 2.23 (19) | 1.50 (19) | 0.63 (10) | 1.95 (21) | 0.72 (8) | 0.49 (15) | 0.52 (11) | 0.99 (1) | 0.75 (9) | 51.23 (24) | 14.98 | 25 |
noisy |
1.43 (21) | 1.42 (22) | 1.35 (23) | 1.75 (20) | 1.17 (21) | 0.54 (11) | 1.55 (22) | 0.65 (9) | 0.40 (16) | 0.51 (12) | 0.97 (3) | 0.85 (4) | 63.50 (17) | 15.25 |
During the submission stage, the real-time leaderboard used an earlier version of the evaluation_metrics/calculate_phoneme_similarity.py script to compute the LPS scores. For the final results, we fixed a bug in the LPS evaluation and updated the scores accordingly. While the corrected scores differ slightly from those shown on the previous leaderboard, the overall ranking order remains unchanged.