A prototype connecting the recording, transcript, and report

Arc’s first prototype explores how the work of speech and language sampling can fit together: listening to a recording, reviewing its transcript, analyzing the sample, and preparing a report. A single demonstration recording and automatic transcript provide the material for developing that workflow.

Reviewing the transcript with the recording

Words in the transcript are linked to their places in the recording, making it possible to listen closely while checking the text. Word timings can be adjusted, utterances split or combined, and speaker assignments corrected.

The prototype also supports marking words as fillers, retraces, and errors, keeping these annotations attached to the transcript used for analysis.

Language and speech analysis

The prototype connects the transcript to SUGAR language measures and percentage of consonants correct (PCC). For phonological analysis, target and produced pronunciations appear beneath the words so they can be compared with the recording.

The demonstration sample provides a way to explore how transcription corrections and annotations contribute to the analysis within the same workflow.

Assembling a report

A report brings together sample information, language measures, phonological analysis, and transcript markings. Sections can be included or left out, with space for the SLP’s own notes. The report can be copied into a document or printed.

At this stage, the prototype works with its built-in demonstration sample. It establishes the workflow that will support importing and saving separate samples in the application.