audio-voice-recovery

Enhances degraded audio and transcribes speech using FFmpeg, SoX, Whisper, and RNNoise.

193|17|Updated Jan 16, 2026

Search Tags:#transcription #audio #enhancement #forensics #noise reduction #voice recovery

Ready to run

npx skills add https://github.com/pproenca/dot-skills --skill audio-voice-recovery

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: audio-voice-recovery
Source: https://github.com/pproenca/dot-skills/tree/main/skills/.experimental/audio-voice-recovery
Command: npx skills add https://github.com/pproenca/dot-skills --skill audio-voice-recovery

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires librosa, soundfile, scipy, numpy, noisereduce, openai-whisper, praat-parselmouth, torch, torchaudio, pydub, demucs, exiftool, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill addresses the challenge of understanding speech in low-quality, noisy, or damaged audio recordings, making critical information accessible.

Core Features & Use Cases

Noise Reduction: Apply advanced algorithms to remove hiss, hum, and background noise.
Voice Enhancement: Boost speech frequencies and improve clarity for better intelligibility.
Forensic Analysis: Detect tampering, verify timestamps, and extract speaker characteristics.
Use Case: Recovering crucial dialogue from a surveillance recording with significant background traffic noise and intermittent clipping.

Quick Start

Use the audio-voice-recovery skill to enhance the attached audio file 'evidence.wav' for clearer speech.

Frequently Asked Questions about audio-voice-recovery

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I enhance degraded audio recordings to make speech clearer?▼

To enhance degraded audio recordings, you can apply advanced noise reduction, spectral processing, and voice frequency boosting. This process removes hiss and background noise while improving speech intelligibility for difficult transcriptions.

Can I transcribe difficult audio with heavy background noise?▼

Yes, you can transcribe difficult audio by combining noise reduction algorithms with OpenAI Whisper. This workflow minimizes background interference first, allowing the transcription engine to accurately extract speech from low-quality recordings.

What is the best way to perform forensic audio analysis on a surveillance recording?▼

Forensic audio analysis involves detecting tampering, verifying timestamps, and extracting speaker characteristics. It uses spectral processing and 45 specialized rules across 8 categories to authenticate recordings while prioritizing signal preservation.

Does this audio voice recovery process work with standard formats like WAV?▼

Yes, the audio voice recovery process works with standard formats like WAV using FFmpeg, SoX, and Python libraries. It handles various input files for comprehensive analysis, automated workflows, and noise profiling.

How do I reduce background traffic noise from a surveillance recording?▼

To reduce background traffic noise, apply targeted noise profiling and temporal processing algorithms. These tools isolate and remove intermittent clipping and steady background hums while preserving the original voice signal.

What are the limitations of noise reduction for damaged audio?▼

Noise reduction for damaged audio is limited by severe signal degradation and intermittent clipping. While spectral processing and voice enhancement improve intelligibility, they cannot fully restore audio signals that lack sufficient original voice data.