Bump Audio Analysis

Audio analysis tool to detect bump/knock sounds in long WAV recordings.

Setup

1. Create and activate virtual environment:

python3 -m venv venv
source venv/bin/activate

2. Install dependencies:

pip install -r requirements.txt

Usage

Basic usage:

python bump_detector.py your_audio.wav

Options:

• --threshold 2.5 - Adjust sensitivity (lower = more sensitive, default: 3.0)
• --output results.json - Custom output file path
• --quiet - Minimal console output

Output

The script generates:

• Console output with timestamps of detected events
• JSON file with detailed analysis results

Example output:

Analysis Results:
Duration: 02:30:45.123
Total detections: 23
Background level: 0.012345
Detection threshold: 0.037035

Detected events:
  1. 00:05:23.456
  2. 00:12:34.789
  ...

How It Works

The detector uses two complementary methods:

1. Onset Detection - Identifies sudden spectral changes using librosa
2. Amplitude Thresholding - Finds peaks above background noise level

Results are combined and filtered to minimize false positives while capturing actual bump/knock events.