FLAC vs WAV vs MP3 — what changes for transcription

For practical transcription accuracy, FLAC is identical to WAV: both are lossless, the engine sees identical waveforms after decoding. The difference is file size: a 1-hour stereo FLAC is typically 200–350 MB vs ~600 MB for the same recording as WAV — about half. Compared to MP3, FLAC carries the full uncompressed signal where MP3 has thrown away ~90% of the bits; the gap matters for noisy, low-volume, or unusually-spoken audio (heavy accents, technical jargon, overlapping voices) where every artefact-free decibel helps.

Where FLAC is the right input

Archival projects digitising tape, cassette, or vinyl into a permanent text+audio archive. Oral history projects capturing interviews that need to be both listened to (decades from now) and indexed (now). Audiobook publishers with master FLAC stems wanting an accurate transcript for ebook companion text or for accessibility captions. Music transcription of spoken-word tracks (poetry, monologue, comedy) where the artistic recording shouldn't be re-encoded.

Markdown structure for archival recordings

The structured Markdown output is especially valuable for archival use — a flat plain-text transcript becomes hard to navigate at hour-plus durations, while structured Markdown with section headings (## at topic shifts), speaker labels, and timestamps lets a researcher decades from now grep for a specific term and jump straight to the audio position. The Markdown is the index; the FLAC is the master.