Clarify auto-tune opt-in behavior

[johnmschoonover]

Summary

• note in the CLI docs that auto-tune remains opt-in and must be enabled explicitly
• update the CLI help text to indicate the flag is disabled by default

Testing

• not run (documentation-only update)

---

https://chatgpt.com/codex/tasks/task_e_68eed5092c248333b915aaac7a016bd3

[chatgpt-codex-connector]

💡 Codex Review

DnD_Session_Transcribe/src/dnd_session_transcribe/features/asr.py

Lines 55 to 58 in 62cd39b

	segs_iter, info = fw.transcribe(
	audio_path,
	language="en",
	beam_size=asr_cfg.beam_size,

Guard beam_size from None before calling faster-whisper

Auto-tune heuristics set beam_size to None for noisy or mixed audio so sampling can be used, but run_asr forwards that value directly to WhisperModel.transcribe (beam_size=asr_cfg.beam_size). The faster-whisper API’s signature requires an int and defaults to 5; passing None raises a TypeError when auto-tune runs in apply mode. Ensure the call supplies a valid integer (e.g., omit the parameter or coerce to 1) when sampling is desired.

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting

• @codex fix this CI failure
• @codex address that feedback