BitBeagle is a lightweight audio comparison tool that analyzes audio files and detects similarities (or differences) using MFCC (Mel-Frequency Cepstral Coefficient) features. Like a loyal beagle sniffing out audio differences, this script compares each track to a reference and reports back the most and least similar matches.
Compare audio files using cosine similarity
Extract MFCC features for robust analysis
Supports multiple audio formats: .wav, .mp3, .flac, .ogg, .aac, .m4a
Outputs a ranked similarity report as CSV
Prints results directly to console
project_root/
├── reference/ # Folder containing the reference audio file (first valid audio file found will be used)
├── stems/ # Folder with all audio files to compare against the reference
├── results.csv # Output similarity report (auto-generated)
├── main.py # Main script
-
Download Github Repo (Code -> Download ZIP -> Extract ZIP)
-
Install required libraries:
pip install librosa scipy numpy -
Place your reference audio file inside the reference/ folder that you created.
-
Add all audio files you want to compare into the stems/ folder that you created.
-
Run the script:
python main.py -
View results printed to console or open results.csv for a full report.
Loads the reference and each target audio file.
Extracts MFCC features (averaged over time).
Computes cosine similarity between MFCC vectors.
Outputs and ranks results by similarity score (1 = identical, 0 = completely different).
Librosa
Numpy
Scipy
Because this little utility sniffs out audio differences bit by bit; just like a trusty beagle on the trail.
MIT License: use freely, sniff responsibly. We are not responsible for illegal usage of BitBeagle. This is an Open-Source program for educational purposes only. Please provide proper credit to the repo and creator.
Have fun sniffing! Questions? Leave a issue request!
