SURE-Tagger prediction
ASR transcript
Interactive demo · curated audio
From speech evidence to useful tags.
Listen to recorded and synthetic clips, inspect the structured fields and descriptions, then combine tags to find comparable scenes.
10 audio clips
5 recorded + 5 synthetic
3 prediction sources
01
Listen
Choose a recording
Explore five recorded clips and five controlled synthetic scenes.
02
Representation
What the tags say
Typed fields make conditions queryable. A short description preserves context; unknown values stay unknown.
Evidence-linked description
View complete SURE-Tagger JSON
03
Compose conditions
Find a speech scene
Combine tag values to make a small, comparable subset. This preview does not compute ASR error rates.
04
Same audio
Compare descriptions
Read each system's output for the selected clip. Differences are visible without hiding the full records.