AudiowaveAudioWave

Upload any audio file and separate it into individual stems — vocals, drums, bass, and other accompaniment. All processing is done entirely in your browser.

Click or drag an audio file here to get started

Free Online Source Separation — Extract Vocals, Drums, Bass & Accompaniment

Separate any song into individual stems — vocals, drums, bass, and other accompaniment — directly in your browser. Powered by FFmpeg.wasm, this free audio source separation tool runs entirely on your device with zero server uploads.

What Is Audio Source Separation?

Audio source separation is the process of isolating individual sound sources from a mixed audio recording, such as separating a complete song into its constituent parts — vocals, drums, bass, and other instrumentation. This technology is widely used in music production, remixing, karaoke creation, audio restoration, and content creation. Our online tool brings this powerful capability directly to your browser using frequency-based filtering and phase cancellation techniques via FFmpeg.wasm, enabling high-quality stem extraction without any specialized software or hardware.

How Our Source Separation Tool Works

Our source separation tool employs audio DSP techniques through FFmpeg to decompose your audio into four stems. First, the audio is converted to a high-quality WAV format. Then, using mid-side decomposition, the center channel (typically containing vocals and bass) and side channel (containing drums and other instruments) are separated. Each stem is further refined using frequency-specific filters: the vocal stem uses a bandpass filter from 250 Hz to 8000 Hz to capture the human voice range; the bass stem uses a low-pass filter at 200 Hz; the drum stem applies phase cancellation and low-pass filtering at 150 Hz; and the remaining other stem captures the side channel content above 200 Hz. Each extracted stem is available for preview and download.

Supported Audio Formats

Our source separation tool supports a wide range of audio input formats, making it compatible with virtually any music file you have:

MP3

The most widely used lossy audio format. Ideal for music distribution and general use with good compression and broad compatibility.

WAV

Uncompressed audio format that preserves full audio fidelity. Best for professional music production and high-quality stem extraction.

FLAC

Lossless compressed audio format that reduces file size while preserving original quality. Perfect for archiving and audiophile use.

AAC

Advanced Audio Codec, the standard for YouTube and modern video platforms. Offers better quality than MP3 at similar bitrates.

M4A

Apple's audio format based on AAC encoding. Commonly used in iTunes, Apple Music, and iOS recordings.

OGG

Free and open-source container format using Vorbis compression. Popular in gaming and open-source applications.

OPUS

Highly efficient modern codec with superior quality at low bitrates. Ideal for streaming and real-time communication.

AIFF

Apple's uncompressed audio format, similar to WAV. Used extensively in professional audio production on macOS.

AC3

Dolby Digital audio codec, commonly found in DVDs, Blu-rays, and digital broadcast content.

AU

Legacy audio format from Sun Microsystems / NeXT. Included for comprehensive format coverage.

How to Separate Audio Stems — Step by Step

  1. Upload your audio file by clicking or dragging it onto the upload area. Supported formats include MP3, WAV, FLAC, AAC, M4A, OGG, OPUS, AIFF, AC3, and AU.
  2. Click the "Start Separation" button to begin processing. The tool will analyze and extract four separate stems: vocals, drums, bass, and other accompaniment.
  3. Wait for the processing to complete — each stem is processed sequentially and you can track progress on each stem card in the grid.
  4. Preview each stem using the built-in audio player. Click the Preview button on any stem card to listen before downloading.
  5. Download individual stems as high-quality WAV files using the Download button on each stem card. Use them in your music production, remixes, or content creation projects.

Popular Use Cases for Source Separation

  • Music Remixing & Mashups — Isolate individual stems to create custom remixes, mashups, or live performance backing tracks with precise control over each instrument.
  • Karaoke & Vocal Removal — Extract or remove vocal tracks to create instrumental karaoke versions of any song for singing practice, performances, or entertainment.
  • Audio Education & Analysis — Study individual instrument parts for music education, transcription, ear training, and analyzing song arrangements and production techniques.
  • Content Creation — Extract background music or specific instrument tracks for podcasts, video productions, presentations, and social media content without licensing concerns.
  • Audio Restoration — Separate and isolate problematic elements in archival recordings, such as isolating vocals from noisy backgrounds for cleaning and enhancement.

Frequently Asked Questions About Source Separation

1. How accurate is the source separation?
Our tool uses frequency-based filtering and mid-side decomposition techniques to separate audio stems. While it provides good separation for many songs, especially those with clear frequency separation between elements, it may not achieve the same level of precision as AI-based models like Demucs for complex mixes. The vocal and bass stems (center channel) tend to be cleaner, while drum and other stems (side channel) work best on songs with strong stereo separation.
2. Is my audio file uploaded to any server?
No. Everything runs entirely in your browser using FFmpeg.wasm (FFmpeg compiled to WebAssembly). Your audio file never leaves your device. This makes the tool 100% private and secure — perfect for copyrighted music, unreleased content, or sensitive audio. No signup required, no server uploads, no data retention.
3. What audio formats can I use as input?
We support 10 major audio input formats: MP3, WAV, FLAC, AAC, M4A, OGG, OPUS, AIFF, AC3, and AU. All files are converted to WAV internally for processing. The output stems are available for preview as MP3 and download as WAV files to preserve maximum quality.
4. What is mid-side decomposition in audio separation?
Mid-side (M-S) decomposition is a stereo audio processing technique that separates a stereo signal into two components: the Mid channel (sum of left and right, containing center-panned sounds like vocals and bass) and the Side channel (difference between left and right, containing stereo-panned sounds like drums and background instruments). Our tool uses this technique as a foundation for separating the four stems, then applies frequency-specific filters to refine each stem.
5. What is the quality of the downloaded stems?
Downloaded stems are provided as WAV files encoded with 16-bit PCM at 44100 Hz sample rate, which is CD-quality audio. Preview files are MP3 at 128 kbps for quick streaming. The WAV format preserves full fidelity so you can use the stems in professional music production, DAWs (Digital Audio Workstations), and further processing without quality loss.
6. Can I use the separated stems for commercial projects?
The tool itself is free to use. However, the copyright of the separated stems belongs to the original rights holder of the source audio. You should ensure you have the necessary rights or licenses to use the separated stems for commercial purposes. The tool is a technical utility and does not grant any usage rights to the content you process.
7. Does this tool work on mobile devices?
Yes, the source separation tool works on any device with a modern browser, including smartphones and tablets (iOS Safari, Chrome for Android, Samsung Internet). All processing happens in your browser via WebAssembly, so no app installation is needed. The interface is fully responsive and optimized for touch input, though processing on mobile devices may take longer due to hardware limitations.

Free Online Source Separation — Extract Vocals, Drums, Bass & More

Our free online source separation tool lets you split any song into four individual stems — vocals, drums, bass, and other accompaniment — directly in your browser. Using FFmpeg.wasm for frequency-based filtering and mid-side decomposition, it provides a practical solution for music remixing, karaoke creation, audio analysis, content production, and restoration. With support for 10 audio input formats, stem preview, WAV download, zero server uploads, and no signup required, it is an essential tool for musicians, producers, educators, and content creators who need quick and private access to stem extraction technology.