speakerchange

Speakerchange is the transition of speaking responsibility from one participant to another within a conversation, broadcast, or other scripted dialogue. It marks the point at which the current speaker stops and the next speaker begins.

In discourse analysis, speakerchange is a central feature of turn-taking and is analyzed for timing, overlap, and politeness. In transcription practices, speakerchange is indicated by labeling the speaker and starting a new line or segment when the speaker changes.

In transcription, common conventions include labeling speakers with initials or names (e.g., A:, B:) and using line breaks to signal a new turn. Some annotation schemes insert a dedicated marker such as [Speaker change] or use inline tags in corpus metadata. In subtitling and captioning, a speaker change cue helps readers track who is speaking, sometimes accompanied by color coding or on-screen labels.

In media and broadcast contexts, speakerchange is important for clarity and timing, including news, interviews, and panel discussions. In dialog systems and research datasets, speakerchange or diarization marks help separate utterances by different speakers for analysis or training.

Automatic detection of speakerchange remains challenging due to overlaps, rapid turn-taking, and inaudible segments. Techniques include speaker diarization, prosodic features, and machine learning models, often combined with transcription alignment.

Related topics include turn-taking, dialogue act, diarization, and captioning.