russell@unturf.com f25731ca08 Add make whisper-refs — re-transcribe cloned-voices/*.wav with whisper-large-v3 ... F5-TTS cloning quality depends on ref_text matching the prosody of ref_audio (commas, periods, casing). Previous ref_texts were LibriSpeech ground-truth labels: ALL CAPS, no punctuation — wrong signal for a flow-matching TTS conditioned on text. Whisper hears what F5 will hear. - scripts/whisper_refs.py — transcribe all wavs, rewrite voice_to_speaker.default.yaml + cloned-voices/voices_metadata.json in place. Also writes cloned-voices/whisper_refs.json sidecar. - Makefile: whisper-refs target. Idempotent, rerun whenever cloned-voices/ changes. Run on a GPU host (4090/3090). ~30s for 40 short clips on a 4090.		2026-05-24 11:56:15 -04:00
..
add_voice.py	Organize repository: create scripts/ and docs/ directories	2025-11-09 09:25:21 -05:00
download_diverse_voices.py	Add idempotent voice registry system for permanent speaker-to-name assignments	2026-01-27 13:59:39 -05:00
download_diverse_voices.sh	Add 20 diverse voice samples for Qwen3-TTS	2026-01-26 16:50:54 -05:00
download_samples.sh	Organize repository: create scripts/ and docs/ directories	2025-11-09 09:25:21 -05:00
download_voice_samples.sh	Use LJ Speech sample for voice cloning (Alibaba Cloud URL blocked)	2026-01-26 16:28:39 -05:00
fetch_voices.py	Add 20 diverse voice samples for Qwen3-TTS	2026-01-26 16:50:54 -05:00
say.py	Organize repository: create scripts/ and docs/ directories	2025-11-09 09:25:21 -05:00
test_voices.sh	Organize repository: create scripts/ and docs/ directories	2025-11-09 09:25:21 -05:00
whisper_refs.py	Add make whisper-refs — re-transcribe cloned-voices/*.wav with whisper-large-v3	2026-05-24 11:56:15 -04:00