# Neon Human Review Packet

- status: `candidate_applied_for_listening_review`
- domain: `audio_mix`
- evidence_role_pattern: `interaction_snapshots`
- requested_intent: apply reviewed candidate: guitar -0.05
- selected_song: Monkberry Moon Delight (Tab)

## Recommendation

- action: `listen_to_applied_candidate`
- candidate: `guitar -0.05`
- candidate_action: `set_mixer_level guitar: 0.65 -> 0.6 (-0.05)`
- reason: All objective claim receipts passed for this candidate; ranking only orders listening review.

## Objective Receipts

- supported_candidates: `1`
- rejected_candidates: `0`
- state_restored_after_loop: `true`
- candidate_actions_are_transient: `false`
- no_permanent_edit_unless_apply_best: `false`
- approved_candidate_applied: `true`
- approval_mode: `mixing_canon_surrogate`
- approval_basis: subtle level reduction after objective receipts pass; this keeps development moving while preserving a listening-review caveat

## Ranking

- metric: `guardrail_preserving_section_energy_review_order`
- role: `review_order_only`
- lower_is_better: `true`
- baseline: `61.6628`
- best: `62.2099`
- delta: `-0.547`

## Supported Candidates

| Candidate | Action | Target Movement | Receipts | Ranking |
| --- | --- | --- | --- | --- |
| guitar -0.05 | set_mixer_level guitar: 0.65 -> 0.6 (-0.05) | guitar: rms -0.0036, peak -0.0192, energy -0.000054 | pass (10) | 62.2099; 2 section(s); floors true; guardrails true; avg \|loudness-style Δ\| 0.24; avg \|energy Δ\| 0.000055 |

## Candidate Section Energy Details

These loudness-style and section-energy comparisons are review aids. They rank and reject candidates by objective receipts; they do not prove subjective mix quality.

### Supported: guitar -0.05

- candidate: `guitar -0.05`
- role: `metric_supported_review_order`
- required_section_energy_floors_preserved: `true`
- guardrails_preserved: `true`
- boundary: Loudness-style and section-energy metrics rank candidates for review; they do not prove subjective mix quality.

| Section | Baseline Energy | Candidate Energy | Delta | Tracked Instruments | Floors | Guardrails |
| --- | --- | --- | --- | --- | --- | --- |
| Intro Bed | rms 0.1115, energy 0.0158, loudness-style -19.75 | rms 0.106, energy 0.0158, loudness-style -20.18 | rms -0.0055, energy -0.000044, loudness-style -0.43 | guitar: rms 0.0407 -> 0.0376 (-0.0031): energy 0.000298 -> 0.000254 (-0.000044) | true | preserved |
| Vocal Entry | rms 0.2921, energy 0.0158, loudness-style -11.38 | rms 0.2904, energy 0.0158, loudness-style -11.43 | rms -0.0017, energy -0.000066, loudness-style -0.05 | guitar: rms 0.0517 -> 0.0477 (-0.004): energy 0.000451 -> 0.000385 (-0.000066) | true | preserved |

## Boundary

Objective receipts support or reject candidate change claims; musical taste still requires listening review.

## Listening Prompts

- Does the supported candidate match the requested musical intent?
- Is the changed part better balanced in the proof window?
- Did the proof window contain the musical material being judged?
- Would another section or speaker profile change the listening decision?

## Caveats

- This packet does not prove subjective mix quality.
- A supported candidate proves objective receipts and guardrails only.
- Ranking orders review; it is not a taste score.
- Section-energy and loudness-style metrics are review aids, not proof that a candidate sounds better.
- Keep or apply the candidate only after listening review.