Benchmark Scenario

Speaking-Style Robustness

Whether the model can preserve identity across emotion-driven change, whispered speech, and noise-induced Lombard speaking style.

Speaking style Speaking style shift 3 trial sets

Back To Scenario Index Open Global Leaderboard

Current Leader

wespeaker/w2vbert2

18 ranked

2.7895

scenario EER

0.157223 scenario minDCF

Scenario Coverage

3 trial sets

ESD Whisper40 Whisper Lombard Grid Lombard

Scenario Leaders

Top systems on this scenario

The current reviewed submissions for this scenario.

wespeaker/w2vbert2

vb2-w2vbert

2.7895

scenario EER

0.157223 scenario minDCF

wespeaker/res152-voxceleb

wespeaker_resnet152

4.1329

scenario EER

0.268357 scenario minDCF

wespeaker/samresnet100

vb2-sam100

4.3042

scenario EER

0.238095 scenario minDCF

Scenario Ranking

Full ranking for Speaking-Style Robustness

A scenario row first averages linked trial sets within each source dataset, then averages the source means and ranks systems by the resulting EER.

This Scenario All Scenarios Global Board

Rank	System	Rank Signal	Scenario EER	Scenario minDCF	Evidence Rank / EER by trial-set order: ESD · Whisper40 Whisper · Lombard Grid Lombard
#1	wespeaker/w2vbert2 vb2-w2vbert	Leader	2.7895	0.157223	#1 2.5000 #1 5.8125 #2 0.0559
#2	wespeaker/res152-voxceleb wespeaker_resnet152	Top 3	4.1329	0.268357	#3 3.0742 #3 9.2500 #3 0.0745
#3	wespeaker/samresnet100 vb2-sam100	Top 3	4.3042	0.238095	#2 2.7503 #6 10.1250 #1 0.0373
#4	wespeaker/res293-voxceleb wespeaker_resnet293	Competitive	4.3256	0.263660	#5 3.2714 #4 9.5562 #6 0.1490
#5	palabraai/redimnet2-b6 redimnet2	Competitive	4.4502	0.291976	#4 3.0850 #7 10.1875 #4 0.0782
#6	wespeaker/res34-voxceleb wespeaker_resnet34_voxceleb	Competitive	5.0359	0.303656	#6 3.5211 #9 11.4375 #5 0.1490
#7	iic/eres2net-en eres_en	Competitive	5.4878	0.330691	#7 4.3650 #10 11.8750 #8 0.2235
#8	iic/eres2netv2-zh eresv2_zh	Competitive	5.9688	0.361679	#12 7.5779 #2 9.1250 #12 1.2034
#9	wespeaker/campplus-voxceleb campplus_wespeaker	Competitive	6.3574	0.358438	#10 4.6625 #11 14.0000 #11 0.4098
#10	wespeaker/ecapa1024-voxceleb wespeaker_ecapa1024	Needs work	6.5691	0.361235	#9 4.6018 #12 14.8000 #10 0.3055
#11	wespeaker/res34-cnceleb wespeaker_r34	Needs work	7.0743	0.410279	#13 8.0200 #5 9.8125 #16 3.3905
#12	wespeaker/ecapa512-voxceleb wespeaker_ecapa512	Needs work	7.1128	0.367138	#8 4.4900 #15 16.6250 #7 0.2235
#13	speechbrain/ecapa ecapa	Needs work	7.2261	0.422165	#11 5.7925 #14 15.6250 #9 0.2608
#14	iic/campplus campplus	Needs work	7.5693	0.424138	#15 9.6571 #8 11.2625 #14 1.7884
#15	iic/eres2net-large-3dspeaker eres2net_large_3dspeaker	Needs work	8.6857	0.465867	#14 9.4684 #13 14.8750 #13 1.7139
#16	speechbrain/xvector xvector	Needs work	15.6681	0.597160	#17 13.0389 #16 32.1250 #15 1.8405
#17	microsoft/wavlm-base-plus-sv wavlm_base	Needs work	18.0734	0.805194	#18 13.3674 #17 32.6563 #18 8.1967
#18	microsoft/unispeech-sat-base-plus-sv unispeech_sat_base_plus	Needs work	18.1898	0.805772	#16 12.8825 #18 33.6875 #17 7.9993