New Script: Demucs Stem Splitter for Vegas Pro (AI Audio Separation)

kjetil wrote on 8/17/2025, 11:57 AM

Hi everyone,

I’ve been working on a script for Vegas Pro that integrates Demucs to separate audio tracks into stems (vocals, drums, bass, guitar, piano, etc.) and then automatically imports them as separate audio tracks inside Vegas.

I’m now sharing my current build in case anyone wants to test it, play around with it.

What’s included

demucs_cli.exe (required AI engine, ~5.6 GB unpacked – zipped download provided)
Vegas Pro .dll script file
Configuration .ini file
Demo screenshot (JPG)
Demo video (MP4)
install.txt with simple installation instructions

How to use

Download and unzip the package (link below).
Follow the steps in install.txt to copy the files into your Vegas Pro script folder.
Run the script on an audio event – Demucs will process it and bring the stems back into your Vegas project as new tracks.

Download

👉 Google Drive link to package: https://drive.google.com/drive/folders/1Iw3w6y_2MVOLo5zN6YYIYvdL6W2xq_GX?usp=sharing

(Please note: the AI engine is large – around 5.6 GB when unpacked.)

Demo

Here’s a short preview video showing how it works:

Here is a screenshot of the menu showing options

Important

This is an early build and provided as-is.
Tested in Vegas Pro 21/22 (should also work in other recent versions).
Performance depends on your system (GPU recommended).
No warranties – use at your own risk.

Thanks,
Kjetil

Comments

kjetil wrote on 8/18/2025, 9:13 AM

Update: I have zipped both Demucs Stem Splitter.dll and Demucs Stem Splitter.ini into Script Menu.zip to avoid warnings some users experienced when downloading them separately. The old separate files have been removed — please use the zip file instead.

Thanks,
Kjetil

snibchi1 wrote on 8/19/2025, 5:01 AM

Hmm... Four posts in nine years. Why should I trust your code?

JimmyH wrote on 8/19/2025, 5:04 AM

Hmm... Four posts in nine years. Why should I trust your code?

Rude

RogerS wrote on 8/19/2025, 9:19 AM

Very cool concept to integrate this into VEGAS.

JanGunnar-Martinsen wrote on 8/19/2025, 2:00 PM

I have tried it and it works perfect. I work with live music videos , and sometimes vocals are too low or drums too loud.. it is so easy to fix.. Thank you Kjetil, it is just what I needed.. 🙂👍

john_dennis wrote on 8/19/2025, 10:34 PM

@kjetil The shade tree script user was able to get it done without reading the Install.txt file closely. Rather, he read the error messages that were thrown when things weren't done correctly. With no CUDA available, he used CPU on a two-track live club recording from 1976. In the 1990's the open-reel tapes were transferred to digital and CDs made for the musicians on the event. Many times in the process he wished he had an eight track recorder, but that was not the case in the 1970s for a location recording.

He promised to spend more time with the output stems as time permits.

He and I thank you for the opportunity to create something new from something old.

kjetil wrote on 8/19/2025, 10:52 PM

Thank you all for the kind words and feedback! I really appreciate it, and it’s great to hear the script is useful in different workflows. 🙂👍

RogerS wrote on 8/19/2025, 11:38 PM

Do you think this would be useful in stripping out background music in an interview situation?

JanGunnar-Martinsen wrote on 8/19/2025, 11:44 PM

Do you think this would be useful in stripping out background music in an interview situation?

Absolutely. I have tried that also, and had to put back a bit noise to make it real..

john_dennis wrote on 8/20/2025, 6:00 AM

I created an icon for the script to make it stand out. Link: Demucs Stem Splitter Icon

kjetil wrote on 8/20/2025, 10:53 AM

A big thank you to @john_dennis for creating an icon for the script! I’ve updated the Script Menu.zip package to include it. Really appreciate the contribution! 🙂👍

aaronstanley75 wrote on 8/20/2025, 3:49 PM

I did everything normally, and it shows me this when I want to run the script, please help me

john_dennis wrote on 8/20/2025, 5:28 PM

@aaronstanley75 Save your project with a new unique name to a folder on your desktop, Documents folder or another drive that you have write access to. You’ll do fine.

aaronstanley75 wrote on 8/21/2025, 3:33 AM

Thank you for your reply. It worked on my side, but I find it surprising that it needs to be saved in order to use it.

Can this also work with other software than Vegas Pro? If so, could you tell me where I can download it so that it works with other programs?

kjetil wrote on 8/21/2025, 4:54 AM

Glad to hear it worked on your side! The reason the project needs to be saved first is that the separated audio files from Demucs are written into the same folder as the Vegas project. That was the idea behind integrating it as a Vegas script.

As for other software: the script itself is designed only for Vegas Pro. However, the demucs_cli.exe tool included in the package is just a command-line program and can be used independently of Vegas with any DAW or editor that accepts standard audio files.

If you are interested in alternatives, there are also other tools like Demucs-GUI and Steinberg SpectraLayers that provide similar stem separation outside of Vegas.

This script is mainly shared here as a hobby project for Vegas users, so support is limited — but I hope this info points you in the right direction.

JimmyH wrote on 8/28/2025, 1:13 AM

It works very well!

Howard-Vigorita wrote on 9/1/2025, 1:39 PM

Nice. Will try soon. I've been using SpectralLayers for this sort of thing but your Vegas integration is compelling.

Manuel-CernaSnchez wrote on 10/4/2025, 2:56 PM

Hola a todos, hablo español, disculpenme por no poder hablar su idioma, traté de seguir las instrucciones de instalación para poder usar el script, pero en el punto 3 me pierdo, nose que hacer y la verdad no quiero molestarlos con estas cosas que me pasan, pero si alguien pudiera compartir un vídeo tutorial de cómo hacer la instalación correctamente paso a paso, se lo agradecería mucho.

john_dennis wrote on 10/4/2025, 8:46 PM

@Manuel-CernaSnchez

Put these three files into Documents / Vegas Scripts Menu.

In Note Pad, edit Demucs Stem Splitter.ini

... to show the path to the demucs_cli folder. (It could be anywhere that you have access rights.)

Set Device = cpu (until you figure out how to identify your GPU)*

In a Vegas Pro project, select an audio track.

Run the script:

After the script runs, you'll be greeted with all of the separate tracks on the timeline.

*I don't have an AMD or nVidia card so I use CPU only.

paul-marshall wrote on 10/5/2025, 6:03 AM

Brilliant! I use AI noise reduction quite a lot usually to reduce background traffic noise and wind. Thanks for sharing.

Sr_Garcia wrote on 10/7/2025, 3:36 AM

It works perfectly and is VERY easy to use.

Thank you for sharing your knowledge and your work with us all.

Manuel-CernaSnchez wrote on 10/9/2025, 5:25 PM

@Manuel-CernaSnchez

Coloque estos tres archivos en Documentos / Menú de scripts de Vegas.

En el Bloc de notas, edite Demucs Stem Splitter.ini

... para mostrar la ruta de acceso al demucs_cli carpeta. (Podría ser en cualquier lugar donde tenga derechos de acceso.)

Establecer dispositivo = CPU (hasta que descubras cómo identificar tu GPU)*

En un proyecto de Vegas Pro, seleccione una pista de audio.

Ejecutar el script:

Una vez que se ejecute el guión, se le recibirán todas las pistas separadas en la línea de tiempo.

*No tengo una tarjeta AMD o nVidia, así que solo uso CPU.

Muchas gracias por la ayuda, pude lograrlo al fin, aunque es un poco complicado, pero al fin pude instalarlo y todo bien la verdad

Manuel-CernaSnchez wrote on 10/10/2025, 9:36 AM

Hola con todos, yo había tenido un problema con instalar el script, no lo entendía bien, al final @john_dennis pudo ayudarme, muchas gracias. Pero un amigo de mi comunidad de edición hizo un ejecutable .exe no es virus, pueden revisarlo, sólo que realiza toda la instalación automaticamente, se los dejo ahí, saludos.
https://drive.google.com/drive/folders/1R3jrpYgpTPXCcRTmyFkhS5bqTqzs6S5z?usp=sharing

JimmyH wrote on 10/16/2025, 8:55 AM

@kjetil Just an update on my experience using this one

The paper from INRIA says best results are had when the audio file is up-sampled prior to Demucs and then the output down-sampled to match the original. Is that correct? I cannot tell with the stuff I've tested although maybe it is better and I'm just not able to hear. Also, it says that best results are had when no audio fx are applied prior to Demucs although I think this is not always true as I have had some improved results particularly with bass guitar stems. Sometimes I render after applying Resonant Filter and then do the splitting on that new .wav. Final note is that the 4-stem mode is usually the best for the music I've tested but the 1-stem-plus-accompaniment always matches or exceeds the splitting quality. 6-stem mode is great but output sometimes includes sharp peaks which I try to compress.