Skip to content

Commit

Permalink
more updates to make documentation and batch happen
Browse files Browse the repository at this point in the history
  • Loading branch information
bjaan committed Aug 2, 2023
1 parent b6efa60 commit ecb3928
Show file tree
Hide file tree
Showing 4 changed files with 58 additions and 4 deletions.
1 change: 1 addition & 0 deletions .gitignore
Original file line number Diff line number Diff line change
@@ -1,3 +1,4 @@

demucs/dist/
models/htdemucs_ft/
dist
49 changes: 46 additions & 3 deletions README.md
Original file line number Diff line number Diff line change
@@ -1,3 +1,44 @@
# DeVoX-Isolate - remove vocals and voice-overs from any audio file

Introducing DeVoX-Isolate: Your Ultimate Audio Tool for Seamless Vocal and Voice-over Removal!

Are you tired of struggling to remove unwanted vocals and voice-overs from your audio files? Do you wish to create instrumental versions of your favorite songs or podcasts effortlessly? Look no further! DeVoX-Isolate is here to revolutionize your audio editing experience.

DeVoX-Isolate is a cutting-edge software designed to remove vocals and voice-overs from any audio file with unparalleled precision and ease. Developed by a team of audio experts and powered by advanced AI algorithms, this tool offers an innovative and efficient way to achieve professional-grade audio manipulation.

Key Features:

* Seamless Vocal Removal: DeVoX-Isolate utilizes sophisticated deep learning technology to detect and isolate vocals and voice-overs from your audio tracks. By extracting vocals from background music, podcasts, or recordings, you can effortlessly obtain pristine instrumental versions for remixes or karaoke performances.

* High-Fidelity Output: We understand the importance of preserving audio quality. With DeVoX-Isolate, you can rest assured that your audio will retain its original brilliance, even after removing the vocals. Our tool ensures minimal loss of audio quality during the process.

* User-Friendly Interface: DeVoX-Isolate boasts a user-friendly interface that enables both novices and professionals to work with ease. Simply drag-and-drop your audio file on our application and the screen shows you status updates.

* Multiple File Format Support: Whether it's MP3, WAV, FLAC, or any other popular audio format, DeVoX-Isolate accommodates a wide range of file types. This flexibility ensures that you can work with your preferred audio files without any hassle.

As this is a full-beta software, we do not provide a GUI yet, but you can already use our core functionality.

TODO :
* Batch Processing: Time is of the essence, and DeVoX-Isolate respects that. Our tool supports batch processing, allowing you to remove vocals from multiple audio files simultaneously, significantly reducing editing time.

* Real-time Preview: Get a sneak peek of the final result before committing to the changes. DeVoX-Isolate provides a real-time preview, empowering you to fine-tune your settings and achieve the perfect audio output.

DeVoX-Isolate is a must-have tool for musicians, content creators, podcasters, and audio enthusiasts alike. Whether you're a professional sound engineer or an aspiring artist, this software caters to all your vocal removal needs.

Join the revolution of audio editing with DeVoX-Isolate and unlock a world of possibilities. Experience the power and convenience of this remarkable tool today!

# Usage

When you're on Windows you can download and use our release directly.

You can run the batch file `devox.bat` with just one parameter the input file, e.g. an mp3.

Or simply drag and drop it on the `devox.bat` file

# Building

To build `devox` for yourself and your OS. First download this repository.

## 0. Install Python

We need Python 3.10 at least installed (it was tested with this version to say the least).
Expand Down Expand Up @@ -29,9 +70,10 @@ You can now remove the contents of the `demucs\venv` folder as it is no longer n
First download the `ffmpeg` and `ffprobe` executables, e.g.:

Windows: https://www.gyan.dev/ffmpeg/builds/ffmpeg-release-essentials.zip

MacOS: https://evermeet.cx/ffmpeg/ffmpeg-109428-g10a56363a7.zip and https://evermeet.cx/ffmpeg/ffprobe-109428-g10a56363a7.zip

Then copy them to `demuncs\dist`
Then copy them to the `demucs\dist` folder

## 4. Download models

Expand All @@ -51,7 +93,8 @@ and place them in the `models\htdemucs_ft` folder

Note: List of all filenames (in case you need a model other than htdemucs_ft) https://raw.githubusercontent.com/facebookresearch/demucs/main/demucs/remote/files.txt

demucs-cxfreeze.exe -n htdemucs_ft --repo ..\..\models\htdemucs_ft
## 5. Build Release

Run `build-release.bat` to copy all the required files to the `dist` folder

demucs-cxfreeze.exe -n htdemucs_ft --repos ..\..\models\htdemucs_ft -o ..\..\output --two-stems vocal INPUTFILE
`devox.bat` is now available
10 changes: 10 additions & 0 deletions build-release.bat
Original file line number Diff line number Diff line change
@@ -0,0 +1,10 @@
pushd %~dp0
mkdir dist
copy devox.bat dist
mkdir dist\demucs
mkdir dist\demucs\dist
xcopy /s/e demucs\dist dist\demucs\dist
mkdir dist\models\htdemucs
xcopy /s/e models\htdemucs_ft\*.th dist\models\htdemucs_ft\
xcopy /s/e models\htdemucs_ft\*.yaml dist\models\htdemucs_ft\
popd
2 changes: 1 addition & 1 deletion devox.bat
Original file line number Diff line number Diff line change
@@ -1,3 +1,3 @@
pushd %~dp0\demucs\dist
demucs-cxfreeze.exe -n htdemucs_ft --repo ..\..\models\htdemucs_ft -o ..\..\output --two-stems vocals ""%1""
demucs-cxfreeze.exe -n htdemucs_ft --repo ..\..\models\htdemucs_ft -o ..\..\output --two-stems vocals %1
popd

0 comments on commit ecb3928

Please sign in to comment.