Loading…

Module 3 Overview | DATASPHERES AI - Dataspheres AI

What to Expect This module covers transcription for field recordings: speaker diarization for interviews and focus groups, language hints, the low-resource...

What to Expect This module covers transcription for field recordings: speaker diarization for interviews and focus groups, language hints, the low-resource-language workflow, and a triage procedure that catches defective transcripts before they reach coding. Module Learning Objectives By the end of this module, you will be able to: Select the correct transcription mode (group, single, low-resource) for a given recording (CLO 3) Apply language hints and forced re-transcription to recover usable text (CLO 3) Triage transcript quality before analysis and flag machine drafts for verification (CLO 3) To-Do List Read: Diarized, multilingual transcription Read: Quality discipline - ASR output as a first draft Complete: Module 3 quiz Your learning in this module is evaluated by the module quiz (pass mark 60%). Quizzes are untimed and autograded. In the sample study Fourteen recordings are transcribed into the datasphere's Library - about 78,000 timestamped words. Open any transcript and notice what ASR gets wrong: garbled names, positional speaker labels in the debate. The transcript is a first draft you review - the sample study says so on every document header.