Separate workspaces for transcription, language, and phonology

Arc now has separate Transcription, Language, and Phonology screens for working with the same sample. Each brings the information for that task into view, so reviewing the words does not require keeping grammatical labels and phonetic transcriptions on screen at the same time.

The same transcript, viewed for different tasks

Transcription focuses on what was said, who said it, and where utterances begin and end.

Language brings parts of speech and transcript markings alongside the words. Markings such as fillers, retraces, and errors are written beneath the words they describe.

Phonology pairs the words with target and produced pronunciations in IPA for comparison with the recording.

The Language and Phonology screens appear according to the measures selected for the sample. The work stays connected: corrections to the transcript carry through to the screens used for analysis.

Refreshing language annotations after corrections

Correcting a transcript can change the context used to identify parts of speech. The Language screen now includes an action to reassess those labels using the revised text and update the clause counts that depend on them.

Seeing how the draft was divided

For newly imported samples, an utterance’s details show why its boundary was placed and which speaker the automatic draft assigned to it. This provides context for reviewing the draft’s divisions and speaker assignments. The original speaker assignment remains available after a correction.