> For the complete documentation index, see [llms.txt](https://musicwave.gitbook.io/docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://musicwave.gitbook.io/docs/tools-documentation/stem-splitter.md).

# Stem Splitter Guide

The Stem Splitter separates any song into individual instrument tracks (stems): vocals, drums, bass, and other instruments. This guide covers how to use it for remixes, karaoke, sampling, and music production.

## What are stems?

A stem is an individual audio track within a song. Modern productions are built from layered stems:

* **Vocals** — lead and backing vocal tracks
* **Drums** — kick, snare, hi-hats, percussion
* **Bass** — bass guitar or synth bass
* **Other instruments** — guitars, keys, synths, strings, etc.

Original recordings have all stems separated in the studio. Once mixed and mastered, they're combined into a single audio file. The Stem Splitter reverses this process using AI.

## Why split stems?

Separated stems open up creative and practical possibilities:

* **Karaoke** — remove vocals to sing along
* **Acapella** — extract vocals only for remixing
* **Remixing** — rebuild songs with different elements
* **Sampling** — sample specific instruments cleanly
* **Practice** — isolate parts to learn or play along
* **Mashups** — combine vocals from one song with instrumentals from another
* **Cover songs** — use original instrumentals with new vocals
* **DJ tools** — create acapellas and instrumentals for live mixing

## How to use the Stem Splitter

1. Go to [musicwave.ai](https://www.musicwave.ai)
2. Click **Stem Splitter**
3. Upload your audio file
4. Choose splitting mode (2-stem, 4-stem, or 5-stem)
5. Click **Split**
6. Wait for processing (typically 30-90 seconds)
7. Download individual stems

### Splitting modes

| Mode   | Output stems                      | Use case                  |
| ------ | --------------------------------- | ------------------------- |
| 2-stem | Vocals + Instrumental             | Karaoke, vocal extraction |
| 4-stem | Vocals, Drums, Bass, Other        | Remixing, mashups         |
| 5-stem | Vocals, Drums, Bass, Piano, Other | Detailed production       |

Choose based on what you need. More stems = more processing time but more flexibility.

## Supported file formats

* MP3 (any bitrate)
* WAV (any sample rate)
* FLAC
* M4A / AAC
* OGG

## Output quality

The Stem Splitter outputs:

* **Format:** WAV (lossless)
* **Sample rate:** matches input
* **Bit depth:** 16 or 24-bit

Output quality depends on input quality. Higher quality input = cleaner stems.

## Use cases in detail

### Karaoke and singing practice

Use 2-stem mode to remove vocals from your favorite songs. Sing along to the instrumental version. Great for:

* Vocal practice and warm-ups
* Recording your own cover versions
* Hosting karaoke nights
* Auditions and performances

### Mashups

Combine vocals from one song with instrumentals from another:

1. Split Song A → keep vocals
2. Split Song B → keep instrumental
3. Match BPM (use the [BPM Finder](/docs/tools-documentation/bpm-key-finder.md))
4. Layer the tracks in a DAW
5. Adjust pitch if keys don't match

### Sampling

Extract clean drum loops, bass lines, or melodic phrases:

1. Split the source song
2. Open the desired stem in your DAW
3. Cut and loop the section you want
4. Use as a sample in new productions

This produces much cleaner samples than recording from a full mix.

### Acapella and instrumental versions

* **Acapella version** — vocals only, no instruments
* **Instrumental version** — instruments only, no vocals

These have many uses:

* DJ tools for live remixing
* Sync licensing alternatives
* Educational use (study vocal performance)
* Tribute tracks and karaoke

### Music production

For producers, stem splitting enables:

* Studying how songs are arranged
* Learning specific drum patterns or bass lines
* Replacing parts of existing songs
* Creating beat-only versions for rappers
* Building remixes from official releases

### Cover songs

Combine the Stem Splitter with [Voice Models](/docs/cover-song-tutorials/voice-models.md):

1. Split a song to get the instrumental
2. Generate a new vocal using a different voice model
3. Mix the new vocal over the original instrumental
4. Result: a unique AI cover

## Quality tips

### Use high-quality source files

* Prefer WAV or FLAC over MP3
* Higher bitrate MP3s split better than low bitrate
* Avoid streaming rips (compressed audio quality varies)

### Some songs split better than others

Songs that split well:

* Modern productions with clean separation
* Songs with distinct instrumentation
* Studio recordings with good engineering

Songs that split less cleanly:

* Live recordings with crowd noise
* Heavily processed or distorted music
* Dense orchestral arrangements
* Old recordings with limited fidelity
* Songs where vocals are heavily layered with instruments

### Check for artifacts

After splitting, listen for:

* **"Bleed"** — instruments leaking into other stems
* **Artifacts** — strange noises in extracted stems
* **Phase issues** — when stems combined sound off

These are normal limitations of AI separation. Quality is impressive but not perfect.

## Common issues and solutions

### "Vocals still audible in instrumental"

Some songs have vocals so blended they can't be fully removed. Try a higher quality source file, or use 5-stem mode for cleaner separation.

### "Drums sound thin"

Drums separation works best on songs with prominent drums. Acoustic or sparse arrangements may produce thinner drum stems.

### "Bass is mixed with other instruments"

If bass and other instruments occupy similar frequencies, they can blend. Try 5-stem mode for better isolation.

### "Audio sounds robotic or weird"

This typically happens with low-quality input or songs with heavy effects. Try a different version of the song or a cleaner source.

## Output file management

The Stem Splitter typically delivers a ZIP file containing all stems. Each stem is named clearly:

* `vocals.wav`
* `drums.wav`
* `bass.wav`
* `other.wav` (or `piano.wav` and `other.wav` in 5-stem mode)

Organize these in your DAW project for easy access.

## Free vs. paid tier

* **Free tier** — limited splits per day, standard quality
* **Pro tier** — unlimited splits, higher quality processing

## Legal considerations

Splitting stems doesn't grant you rights to use them commercially. The original song is still copyrighted. Stem splitting is fine for:

* Personal practice and learning
* Educational analysis
* Use with proper licenses (cover licenses, sync licenses)
* Public domain music
* Your own recordings

Check copyright before using split stems in published content.

## Related tools

* [BPM & Key Finder](/docs/tools-documentation/bpm-key-finder.md) — for matching tempos in mashups
* [Voice Models](/docs/cover-song-tutorials/voice-models.md) — apply new vocals to instrumentals
* [Audio Quality Optimization](/docs/cover-song-tutorials/audio-quality.md) — post-processing tips

[Try MusicWave's Stem Splitter free →](https://www.musicwave.ai)


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://musicwave.gitbook.io/docs/tools-documentation/stem-splitter.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.