Demucs: Online AI Music Source Separation (Demucs v4)

Demucs is an AI-powered music source separation tool that allows users to split mixed songs into individual stems, such as vocals, drums, and bass. Developed from an open-source model, Demucs v4 operates on cloud GPUs, making it accessible without the need for local installations. Users can upload songs in various formats and receive cleanly separated audio tracks in a matter of minutes.

1. AI Music Source Separation

   Demucs v4 utilizes advanced AI technology to separate a mixed song into distinct stems, including vocals, drums, bass, and other instruments. This feature enhances the ability to create remixes, karaoke tracks, and more.

2. Multiple Output Models

   Users can choose from different models based on their needs:

   • Standard Model: 2 or 4 stems for most songs.

   • Best Quality Model: 2 or 4 stems with enhanced audio quality.

   • Six Stems Model: Includes guitar and piano, ideal for detailed music production.

3. Quick Processing Time

   The tool processes a four-minute song in approximately 20 to 32 seconds, depending on the selected model, making it efficient for users needing rapid results.

4. User-Friendly Interface

   Demucs Online provides a straightforward interface where users can easily upload songs, select models, and preview separated stems before downloading.

5. Privacy Protection

   Uploaded songs are processed for separation and automatically deleted after 24 hours, ensuring user privacy and data security.

6. Trial Credits

   New users receive 15 one-time trial credits upon signing up, allowing them to test the service without any payment information required.

1. Visit the Demucs Online website and log in or create an account.
2. Upload your song in a supported format (MP3, WAV, FLAC, M4A, OGG).
3. Choose the desired Demucs model based on the number of stems you need.
4. Preview the separated stems using the built-in player.
5. Download the stems individually or as a ZIP file.

What is Demucs?

Demucs is an open-source AI model designed for music source separation, capable of isolating vocals, drums, bass, and other components from a mixed audio track.

How does Demucs work?

Demucs analyzes a mixed song using a hybrid transformer model, processing it as both a raw waveform and a spectrogram to achieve high-quality separation.

Is Demucs Online free?

Yes, users can start for free with 15 trial credits, which cover approximately 15 minutes of audio processing.

How long does it take to process a song?

Typically, processing a four-minute song takes about 20 seconds on the default model, with a short warm-up time if the service has been idle.

Can I use the separated stems commercially?

While the Demucs code is MIT-licensed, the rights to the original recording remain with the owner. Users should check licensing terms for commercial use.

Price data is not explicitly provided, but users can start with 15 free trial credits. After that, one credit covers one minute of audio processing.

1. Demucs excels in providing high-quality music source separation, making it a valuable tool for musicians and content creators.
2. The user interface is intuitive, allowing for easy navigation and quick processing of audio files.
3. However, the reliance on cloud processing may be a limitation for users with slow internet connections or those who prefer local software solutions.
4. While the tool is effective, users may still need a basic understanding of music production to fully utilize the separated stems in their projects.
5. Overall, Demucs Online is a powerful resource for anyone looking to manipulate audio tracks, though it may not replace traditional music production methods for all users.