edi-voice-check

To preserve raw interview voice in manuscript drafts, run a reverse-direction check that extracts voice-worthy phrases from JSONL data and validates their presence in blueprint.md and manuscript sections.

Reverse-direction quality assurance for interview data is a validation process that ensures vivid phrases, tone, and expressions from raw JSONL sampling data survive into audience-ready blueprint and manuscript revisions.

You validate speaker voice travel by checking raw JSONL interview data against the blueprint and manuscript files, using a voice rubric to determine preservation and provide remediation guidance for missing phrases.

Yes, you need a voice rubric to perform data validation on manuscript sections, as it is required to determine voice preservation and provide specific remediation guidance for inserting missing phrases.

The best way to check if original interview phrasing was lost during editing is a reverse-direction spot-check that compares audience-ready language against raw JSONL content and suggests where to insert preserved phrases.

Interview voice-check validation requires access to raw sampling JSONL data, the blueprint.md file, manuscript sections like 03-plot, and a voice rubric to assess and remediate preservation.