List your company

Semantic Representation Evaluation on ARCH (test)

48.96RAVDESSSpeechTokenizer
24.041630.510836.9843.4492Feb 27, 2025Apr 18, 2025Jun 7, 2025Jul 27, 2025Sep 15, 2025Nov 4, 2025Dec 24, 2025

Evaluation results

MethodLinks
SpeechTokenizerToken Rate=50, Q=8, Bitrate [kbps]=42025.1248.9697.2535.9155.6637.930.2750.99
SACodecToken Rate=75, Q=1, Bitrate [kbps]=0.752025.1242.6588.4532.8158.1233.8533.3148.09
DACToken Rate=75, Q=2, Bitrate [kbps]=1.52025.1242.3668.430.2358.5928.6531.6643.32
FACodecToken Rate=80, Q=1, Bitrate [kbps]=0.82025.1242.3182.9727.6653.6436.6533.1146.06
UniCodecTPS=752025.0240.2870.9429.5559.292634.17—
DACToken Rate=75, Q=8, Bitrate [kbps]=62025.1238.8972.9533.3159.4824.430.7743.3
UniCodecTPS=75, semantic stage=without2025.0236.8169.8428.0954.0520.830.21—
EncodecToken Rate=75, Q=8, Bitrate [kbps]=62025.1235.0749.1330.9742.2620.1526.7534.06
WavTokenizerToken Rate=75, Q=1, Bitrate [kbps]=0.92025.1234.3862.9226.8953.2223.428.1538.16
Wavtokenizer (speech)TPS=75, domain=speech2025.0232.5569.57—————
EncodecToken Rate=75, Q=2, Bitrate [kbps]=1.52025.1227.7861.3529.1850.0126.9530.7937.68
EncodecTPS=1502025.0227.4336.491932.4516.9926.3—
DACTPS=1002025.022562.8725.0251.3720.6529.91—
Wavtokenizer (music&audio)TPS=75, domain=music&audio2025.02——28.3557.6425.535.63—