Five voices each read you one sentence. After a 30-second pause, each reads a different sentence — and you name who's speaking. Chance is 20%.
No-fibs — what this lab is
The "strangers" are five synthetic text-to-speech voices (river, honey, sunny, storm, spark) — not real people, and their "identities" are just voice timbres. Voice-identity recognition across different spoken content is a genuine memory paradigm, but this is a self-recorded drill with a 1-item-per-voice design — chance is 20%, so read your score against that bar. The clips were pre-generated once (Honey-class TTS), so a run costs zero credits.