Protocol Index
Trial sets
Trial sets are the basic scored units behind the scenario and global summaries. Each page exposes the source, protocol, and direct system ranking.
26 trial sets
11 scenarios
Fixed-Core Protocols
Trial-set index
| Trial Set | Source Dataset | Protocol | Scenario Groups | Current Leader |
|---|---|---|---|---|
|
TidyVoiceX2-ASV
tidyvoice
|
TidyVoiceX2-ASV |
Official eval-key subset + constructed completion
TidyVoiceX2-ASV eval A/U
|
Cross-lingual
|
wespeaker/w2vbert2 |
|
HI-MIA
hi_mia
|
HI-MIA |
Self-constructed protocol
Public test split
|
Short-duration
|
iic/eres2net-large-3dspeaker |
|
GSC Short
gsc_short
|
Google Speech Commands |
Self-constructed protocol
Official testing split
|
Short-duration
|
iic/eres2netv2-zh |
|
GLOBE
globe
|
GLOBE |
Self-constructed protocol
GLOBE_V2 canonical HF train+test shards
|
Accent/dialect
|
wespeaker/w2vbert2 |
|
CN-Celeb
cnceleb
|
CN-Celeb |
Official protocol
Official evaluation protocol
|
In-the-wild
|
iic/eres2netv2-zh |
|
CN-Celeb Genre
cnceleb_genre
|
CN-Celeb |
Self-constructed supplement
CN-Celeb prepared evaluation pool
|
Genre shift
|
iic/eres2netv2-zh |
|
CN-Celeb Short
cnceleb_short
|
CN-Celeb |
Self-constructed supplement
CN-Celeb prepared evaluation pool
|
Short-duration
|
wespeaker/res34-cnceleb |
|
VoxCeleb1-O
voxceleb_o
|
VoxCeleb1 |
Official protocol
VoxCeleb1-O
|
In-the-wild
|
wespeaker/w2vbert2 |
|
VoxCeleb1-E
voxceleb_e
|
VoxCeleb1 |
Official protocol
VoxCeleb1-E
|
In-the-wild
|
wespeaker/w2vbert2 |
|
VoxCeleb1-H
voxceleb_h
|
VoxCeleb1 |
Official protocol
VoxCeleb1-H
|
In-the-wild
|
wespeaker/w2vbert2 |
|
VoxCeleb Short
voxceleb_short
|
VoxCeleb1 |
Self-constructed supplement
VoxCeleb1 evaluation audio
|
Short-duration
|
palabraai/redimnet2-b6 |
|
3D-Speaker Device
speaker3d_device
|
3D-Speaker |
Official protocol
3D-Speaker device trial file
|
Channel/device
|
iic/eres2net-large-3dspeaker |
|
3D-Speaker Distance
speaker3d_distance
|
3D-Speaker |
Official protocol
3D-Speaker distance trial file
|
Distance
|
iic/eres2net-large-3dspeaker |
|
3D-Speaker Dialect
speaker3d_dialect
|
3D-Speaker |
Official protocol
3D-Speaker dialect trial file
|
Accent/dialect
|
iic/eres2net-large-3dspeaker |
|
FFSVC 2022 Cross-Channel
ffsvc2022_cross_channel
|
FFSVC 2022 |
Official protocol subset
FFSVC 2022 public dev trials
|
Channel/device
|
iic/eres2net-large-3dspeaker |
|
FFSVC 2022 Cross-Domain
ffsvc2022_cross_domain
|
FFSVC 2022 |
Official protocol subset
FFSVC 2022 public dev trials
|
Distance
|
iic/eres2net-large-3dspeaker |
|
Whisper40 Whisper
whisper40_whisper
|
Whisper40 |
Self-constructed protocol
Provided Whisper40 archive
|
Speaking style
|
wespeaker/w2vbert2 |
|
Lombard Grid Lombard
lombardgrid_lombard
|
Lombard Grid |
Self-constructed protocol
Public Lombard Grid audio release
|
Speaking style
|
wespeaker/samresnet100 |
|
VOiCES Noise/Reverb
voices_noise_reverb
|
VOiCES |
Self-constructed protocol subset
VOiCES devkit test split
|
Noise/reverb
|
wespeaker/w2vbert2 |
|
CHiME-6 Domestic Far-Field
chime6_domestic_farfield
|
CHiME-6 |
Self-constructed protocol
CHiME-6 dev split
|
Distance
|
wespeaker/w2vbert2 |
|
CHiME-6 Overlap
chime6_overlap
|
CHiME-6 |
Self-constructed protocol
CHiME-6 dev split
|
Overlap
|
wespeaker/w2vbert2 |
|
ESD
esd
|
ESD |
Self-constructed protocol
Public ESD recordings
|
Speaking style
|
wespeaker/w2vbert2 |
|
AliMeeting Near/Far
alimeeting_near_far
|
AliMeeting |
Self-constructed protocol
Public AliMeeting test-set segments
|
Distance
|
iic/eres2net-large-3dspeaker |
|
AliMeeting Overlap
alimeeting_overlap
|
AliMeeting |
Self-constructed protocol
Public AliMeeting test-set segments
|
Overlap
|
iic/eres2net-large-3dspeaker |
|
VoxKnesset
voxknesset
|
VoxKnesset |
Self-constructed protocol
Official VoxKnesset test split
|
Aging
|
wespeaker/samresnet100 |
|
VoxPopuli Aging
voxpopuli_cross_time
|
VoxPopuli |
Self-constructed protocol
VoxPopuli English train split
|
Aging
|
wespeaker/w2vbert2 |