Diarization

The process of automatically determining "who spoke when" in an audio or video recording by segmenting the audio stream and assigning each segment to a specific speaker. In accessibility contexts, diarization is critical for deaf and hard of hearing users who rely on captions or transcripts, as knowing which participant said what significantly reduces cognitive load during meetings and teleconferences. Effective diarization combined with speaker labelling in captions provides functional equivalency closer to what hearing participants experience when they can distinguish speakers by voice.