Trial Set
3D-Speaker Dialect
Keeps the upstream cross-dialect trial file.
- Source dataset
- 3D-Speaker
- Trials
- 180000
- Protocol source
- Official protocol
- Source split
- 3D-Speaker dialect trial file
Source Dataset
About 3D-Speaker
3D-Speaker is a large speech corpus designed to separate speaker identity from recording device, distance, and dialect variation. It contains more than 10,000 speakers, with controlled recordings across multiple devices and distances and dialect speech for some speakers. The three OpenSVBench pages retain the corresponding official device, distance, and dialect trial files.
- Language
- Primarily Chinese, including dialect speech
- Speakers
- 10,000+
- Controlled variation
- Recording device, distance, and dialect
- Original use
- Speech representation disentanglement and evaluation
Direct Ranking
Systems on this trial set
23 reviewed systems currently have valid results.
| Rank | System | Submitted by | Trial-Set EER ↓ | Trial-Set minDCF ↓ |
|---|---|---|---|---|
| #1 |
Agora Voiceprint (ResNet293)
agora_vp
|
agora_vp | 8.6240 | 0.534607 |
| #2 |
iic/eres2net-large-3dspeaker
eres2net_large_3dspeaker
|
Public checkpoint | 12.0000 | 0.664867 |
| #3 |
W2V-BERT 2.0 + MHFA (VoxCeleb2 + VoxBlink2)
Wav2Vec2-BERT
|
pengjy | 14.7233 | 0.841560 |
| #4 |
AntSpeaker/MECT-B2
antspeaker-mect-b2-vb2-full-core
|
Public checkpoint | 14.9333 | 0.633973 |
| #5 |
wespeaker/w2vbert2
vb2-w2vbert
|
Public checkpoint | 16.2767 | 0.770180 |
| #6 |
palabraai/redimnet2-b6-vb2+vox2
redimnet2-b6-vb2vox2-lm-full-core
|
Public checkpoint | 17.1067 | 0.655653 |
| #7 |
Dual-LoRA W2V-BERT2-Adapter-MFA (TidyVoice)
duallora-w2vbert2-tidyvoice-full-core
|
shangguanqituan | 17.2133 | 0.781067 |
| #8 |
palabraai/redimnet2-b6
redimnet2
|
Public checkpoint | 17.2467 | 0.788147 |
| #9 |
wespeaker/res293-voxceleb
wespeaker_resnet293
|
Public checkpoint | 17.9400 | 0.795833 |
| #10 |
wespeaker/res152-voxceleb
wespeaker_resnet152
|
Public checkpoint | 17.9540 | 0.818293 |
| #11 |
iic/eres2net-en
eres_en
|
Public checkpoint | 18.6367 | 0.845573 |
| #12 |
wespeaker/samresnet100
vb2-sam100
|
Public checkpoint | 19.5813 | 0.801173 |
| #13 |
wespeaker/res34-voxceleb
wespeaker_resnet34_voxceleb
|
Public checkpoint | 19.7947 | 0.869053 |
| #14 |
iic/eres2netv2-zh
eresv2_zh
|
Public checkpoint | 20.9300 | 0.947533 |
| #15 |
speechbrain/ecapa
ecapa
|
Public checkpoint | 21.5767 | 0.926973 |
| #16 |
wespeaker/campplus-voxceleb
campplus_wespeaker
|
Public checkpoint | 22.1600 | 0.912873 |
| #17 |
wespeaker/ecapa1024-voxceleb
wespeaker_ecapa1024
|
Public checkpoint | 22.8587 | 0.928093 |
| #18 |
wespeaker/ecapa512-voxceleb
wespeaker_ecapa512
|
Public checkpoint | 25.2400 | 0.999327 |
| #19 |
wespeaker/res34-cnceleb
wespeaker_r34
|
Public checkpoint | 25.3313 | 0.992747 |
| #20 |
iic/campplus
campplus
|
Public checkpoint | 30.1540 | 0.998420 |
| #21 |
microsoft/wavlm-base-plus-sv
wavlm_base
|
Public checkpoint | 35.6433 | 1.000000 |
| #22 |
microsoft/unispeech-sat-base-plus-sv
unispeech_sat_base_plus
|
Public checkpoint | 38.2767 | 0.999867 |
| #23 |
speechbrain/xvector
xvector
|
Public checkpoint | 41.4233 | 0.999793 |