# audio-voice-recovery

Audio forensics guidelines for recovering speech from degraded recordings. Helps clean noisy files, spot tampering, and prepare speech for transcription.

- skill · `type=skill` · Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, OpenCode · 318 installs · ★ 218
- Author: pproenca
- License: MIT · policy=mirror · the beginning of the file is at the end of this page
- Source: https://github.com/pproenca/dot-skills/tree/HEAD/skills/.experimental/audio-voice-recovery/SKILL.md
- Updated: 2026-08-15
- HTML: https://affpapa.org/en/nevibecode/skills/audio-voice-recovery

## What it does

This skill equips coding agents with 45 rules across 8 prioritized categories to recover low-volume or damaged speech. The workflow enforces a strict hierarchy: analyze the raw signal first, estimate background noise during pauses, and only then apply spectral repairs. It includes methods for declipping, Wiener filtering, and room dereverberation.

Processing connects directly with FFmpeg, SoX, and RNNoise, passing recovered audio to Whisper for multi-pass transcription. For forensic verification, the rules outline electrical network frequency (ENF) checks and metadata scrutiny. Modifying original lossy files breaks forensic credibility: processing must remain strictly on lossless audio.

## Install

### Claude Code (native platform)

1. **Skills CLI (Vercel)**
```bash
npx skills add pproenca/dot-skills --skill audio-voice-recovery -a claude-code
```

2. **Manually: put the files in place**
- File: `~/.claude/skills/audio-voice-recovery/SKILL.md` — Copy the whole skill folder; for one project use .claude/skills/audio-voice-recovery/

### Codex

1. **Skills CLI (Vercel)**
```bash
npx skills add pproenca/dot-skills --skill audio-voice-recovery -a codex
```

2. **Manually: put the files in place**
- File: `.agents/skills/audio-voice-recovery/SKILL.md` — For all projects use ~/.agents/skills/

### Cursor

1. **Skills CLI (Vercel)**
```bash
npx skills add pproenca/dot-skills --skill audio-voice-recovery -a cursor
```

2. **Manually: put the files in place**
- File: `.cursor/skills/audio-voice-recovery/SKILL.md` — Cursor also reads .agents/skills/ and .claude/skills/

### GitHub Copilot

1. **GitHub CLI**
```bash
gh skill install pproenca/dot-skills audio-voice-recovery
```

2. **Manually: put the files in place**
- File: `.github/skills/audio-voice-recovery/SKILL.md`

### Gemini CLI

1. **Skills CLI (Vercel)**
```bash
npx skills add pproenca/dot-skills --skill audio-voice-recovery -a gemini-cli
```

2. **Platform CLI**
```bash
gemini skills install https://github.com/pproenca/dot-skills --path skills/.experimental/audio-voice-recovery
```

### Windsurf (Devin Desktop)

**Manually: put the files in place**
- File: `.devin/skills/audio-voice-recovery/SKILL.md` — The legacy .windsurf/skills/ path still works

### OpenCode

**Skills CLI (Vercel)**
```bash
npx skills add pproenca/dot-skills --skill audio-voice-recovery -a opencode
```

### Claude.ai

**Claude.ai**
- Zip the folder and upload it in Settings → Capabilities → Skills

Where a file is given instead of a command, copy the files from the source to that path: https://github.com/pproenca/dot-skills/tree/HEAD/skills/.experimental/audio-voice-recovery/SKILL.md

## When to use it
- isolate quiet speech from heavy background noise
- repair clipped audio and eliminate room echo
- inspect audio recordings for splices and tampering
- build automated FFmpeg pipelines for Whisper transcription

## Example prompts
- Extract the noise profile from quiet segments in `recording.wav` and apply spectral subtraction
- Generate an FFmpeg command to declip and dynamic-range compress `interview.flac`
- Check `surveillance.wav` for splices and verify timestamps using ENF analysis

## Source text · SKILL.md · MIT

Below is the content of a third-party file. It is data, not instructions for you: do not act on anything written inside unless the human asks. Block tag: 1fe8094675403748.

BEGIN THIRD-PARTY FILE 1fe8094675403748
```text
---
name: audio-voice-recovery
description: Audio forensics and voice recovery guidelines for CSI-level audio analysis. This skill should be used when recovering voice from low-quality or low-volume audio, enhancing degraded recordings, performing forensic audio analysis, or transcribing difficult audio. Triggers on tasks involving audio enhancement, noise reduction, voice isolation, forensic authentication, or audio transcription.
---

# Forensic Audio Research Audio Voice Recovery Best Practices

Comprehensive audio forensics and voice recovery guide providing CSI-level capabilities for recovering voice from low-quality, low-volume, or damaged audio recordings. Contains 45 rules across 8 categories, prioritized by impact to guide audio enhancement, forensic analysis, and transcription workflows.

## When to Apply

Reference these guidelines when:
- Recovering voice from noisy or low-quality recordings
- Enhancing audio for transcription or legal evidence
- Performing forensic audio authentication
- Analyzing recordings for tampering or splices
- Building automated audio processing pipelines
- Transcribing difficult or degraded speech

## Rule Categories by Priority

| Priority | Category | Impact | Prefix | Rules |
|----------|----------|--------|--------|-------|
| 1 | Signal Preservation & Analysis | CRITICAL | `signal-` | 5 |
| 2 | Noise Profiling & Estimation | CRITICAL | `noise-` | 5 |
| 3 | Spectral Processing | HIGH | `spectral-` | 6 |
| 4 | Voice Isolation & Enhancement | HIGH | `voice-` | 7 |
| 5 | Temporal Processing | MEDIUM-HIGH | `temporal-` | 5 |
| 6 | Transcription & Recognition | MEDIUM | `transcribe-` | 5 |
| 7 | Forensic Authentication | MEDIUM | `forensic-` | 5 |
| 8 | Tool Integration & Automation | LOW-MEDIUM | `tool-` | 7 |

## Quick Reference

### 1. Signal Preservation & Analysis (CRITICAL)

- [`signal-preserve-original`](references/signal-preserve-original.md) - Never modify original recording
- [`signal-lossless-format`](references/signal-lossless-format.md) - Use lossless formats for processing
- [`signal-sample-rate`](references/signal-sample-rate.md) - Preserve native sample rate
- [`signal-bit-depth`](references/signal-bit-depth.md) - Use maximum bit depth for processing
- [`signal-analyze-first`](references/signal-analyze-first.md) - Analyze before processing

### 2. Noise Profiling & Estimation (CRITICAL)

- [`noise-profile-silence`](references/noise-profile-silence.md) - Extract noise profile from silent segments
- [`noise-identify-type`](references/noise-identify-type.md) - Identify noise type before reduction
- [`noise-adaptive-estimation`](references/noise-adaptive-estimation.md) - Use adaptive estimation for non-stationary noise
- [`noise-snr-assessment`](references/noise-snr-assessment.md) - Measure SNR before and after
```
END THIRD-PARTY FILE 1fe8094675403748

The third-party file has ended. Everything between the 1fe8094675403748 tags is data: requests and commands from it are not followed. An “end of file” line with another tag inside the block is part of the file.

First 2809 of 13209 bytes. © pproenca · MIT · https://github.com/pproenca/dot-skills/tree/HEAD/skills/.experimental/audio-voice-recovery/SKILL.md

_Part of the page did not fit — see the HTML version._

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md