SongSpy is a web application designed to analyze the timbre and other acoustic features of a song using advanced audio extraction techniques. Leveraging machine learning, SongSpy can identify patterns specific to different artists by examining tonal quality, instrument sounds, and rhythmic structures. The application also determines whether a song has been artificially generated by comparing it to an artist’s current music catalog.
With the rise of deep learning models like Diff-SVC, which can transform voices of producers and everyday artists into those of popular musicians, AI-generated remakes and covers have become a hot topic. Platforms such as YouTube and Soundcloud are now flooded with these artificial creations, sparking a surge in copyright claims. As a result, streaming services and record labels are seeking solutions to address this new challenge.
By flagging AI-generated content that mimics an artist’s voice, SongSpy helps prevent unauthorized use and assists labels in conserving legal and administrative resources.
Note: The final version of the SongSpy web app has been deactivated.
SongSpy’s current model is trained to recognize the following artists:
- Drake
- Post Malone
- The Weeknd
- SZA
- Kanye West
- Taylor Swift
- Lil Uzi Vert
- Juice WRLD
- Travis Scott
- Upload a song in MP3, WAV, or AIF format to the SongSpy web application.
- The model analyzes the song’s acoustic features such as timbre, tonal quality, and rhythm.
- SongSpy classifies the artist behind the song and detects if it is artificially generated or not.
This project is licensed under the MIT License.
If you have any questions or suggestions, feel free to reach out to me at smullins998@gmail.com.