Scenario index for OpenSVBench.
This page groups trial sets into interpretable speaker-verification scenarios. Each card shows the condition being tested, the linked trial sets, and the current ranking leader for that scenario.
High-level evaluation conditions
The board describes systems through scenarios such as aging robustness, short-duration robustness, speaking-style robustness, overlap, and distance mismatch.
- Scenarios: high-level evaluation conditions.
- Trial sets: ranked benchmark entries inside each scenario.
- Datasets: source metadata shown inside scenario and trial-set views.
Every scenario has its own local leaderboard
Each card shows the scenario summary, linked trial sets, and the current leader.
Accent / Dialect Robustness
How reliably the model tracks identity across accent and dialect mismatch.
Aging Robustness
How stable identity representations remain across age-derived and longitudinal recording gaps.
Channel / Device Robustness
How resilient the model is to device and channel mismatch.
Cross-Lingual Robustness
How well speaker identity survives enrollment-test language mismatch.
Distance Robustness
How much performance changes across meeting and domestic distance mismatch.
Genre-Shift Robustness
Whether performance holds when CN-Celeb enrollment and test speech come from different source genres.
In-The-Wild Robustness
How strong the model is on unconstrained celebrity and media speech across official CN-Celeb and VoxCeleb protocols.
Noise / Reverb Robustness
How reliably the model preserves identity when room acoustics, distractor noise, and microphone placement deviate from an easier in-corpus reference condition.
Overlap Robustness
How well speaker identity survives light, mid, and heavy overlap in both meeting and domestic recordings.
Short-Duration Robustness
How much performance holds up when speech evidence is limited by duration.
Speaking-Style Robustness
Whether the model can preserve identity across emotion-driven change, whispered speech, and noise-induced Lombard speaking style.