Fast Music Remover

Fast Music Remover aims to provide a lightweight and fast tool to remove music, sound effects, and noise from any internet media. Even at its early stages, it shows promise for near-realtime usage, with processing times around just 8% of the original media length meaning it can process a 10-minute video in less than a minute! While only offline analysis is currently supported, the project has a clear aim to support live feeds as well. Please consider contributing if this interests you!

https://github.com/user-attachments/assets/e46c161b-0178-4213-b468-245e9f829d5e

The original interview video is by Fisher College of Business, licensed under a Creative Commons Attribution license (reuse allowed).

Background

This project builds on an initial attempt using a demucs-based Python implementation, which had a filtering time cost greater than the source media length. This is due to demucs's advanced capabilities, such as identifying individual instruments, which are features beyond the scope of this project. That initial version can still be found here.

Use Cases

Fast Music Remover can serve a wide range of needs, such as:

Interview Editing: Extract clear voice tracks for professional editing of interviews, especially field reports, removing ambient noise without distorting speech.
Education & Lectures: Remove background noise and improve the overall audio quality of recorded lectures, making them clearer and easier for students to understand.
Streaming & Social Media: Clean up your media before posting it on your preferred platform, avoiding copyright issues by removing background music or sound effects.
Home Recording Enthusiasts: Enhance the quality of home-recorded content by removing unnecessary noise or isolating specific audio tracks for better post-production.
Regular Media Consumers: Remove distractions from your favorite podcast or video, allowing you to focus on the content without the sound effects.

Contributing

Contributions are always welcome! If you're interested in helping improve Fast Music Remover, there are several ways to get involved:

Report Bugs or Suggest Features: Open an issue to report a bug or suggest new features you'd like to see.
Improve the Code: Check the open issues and contribute with a pull request.
Testing: Help by manually testing the software or writing automated tests to improve code coverage and robustness.
Enhance Documentation: Improve existing documentation, add examples, or clarify unclear sections. This makes it easier for others to use and contribute.
Spread the Word: Share Fast Music Remover with others and participate in discussions to shape the project direction.

Please see the contributing guidelines for more details on how to get started.

Benchmarks

The processed audio quality shows no audible degradation so far. A formal benchmarking process for performance and quality is planned. If you're interested in contributing to the benchmarking process, please open an issue or reach out via the discussions!

Prerequisites

To get started with Fast Music Remover, ensure that you have the following software installed on your system. These dependencies are necessary for running the backend server, compiling the C++ processor, and handling media files.

Python 3.9+: Required for running the backend server.
FFmpeg: For extracting, probing, and processing audio files.
CMake: Needed to compile the C++ MediaProcessor.
nlohmann-json: A JSON library required for parsing configuration files in the MediaProcessor.
Docker and Docker Compose (optional but recommended for a quick setup):

Installation Commands

FFmpeg:

On Ubuntu/Debian:

sudo apt update
sudo apt install ffmpeg

On macOS:
```
brew install ffmpeg
```
After installing FFmpeg, ensure the correct path is set in the config.json file. By default, it is set to /usr/bin/ffmpeg. If you are using macOS and installed FFmpeg via Homebrew, update the path in config.json to:
```
"ffmpeg_path": "/opt/homebrew/bin/ffmpeg"
```

CMake:

On Ubuntu/Debian:
```
sudo apt update
sudo apt install cmake
```
On macOS:
```
brew install cmake
```

nlohmann-json:

On Ubuntu/Debian:

sudo apt update
sudo apt install nlohmann-json3-dev

On macOS:
```
brew install nlohmann-json
```

Docker and Docker Compose:

On Ubuntu:

sudo apt install docker.io docker-compose

On macOS:

brew install docker
brew install docker-compose

Note: Ensure all the above dependencies are installed before proceeding with the setup.

Getting Started

To get started with Fast Music Remover, you have two main options: running it directly via Docker (quick and easy) or installing all the necessary dependencies manually for greater customization and control.

Option 1: Quick Start with Docker

Ensure Docker is installed by running:

docker --version

If Docker is installed, you can quickly try Fast Music Remover by running the following command:

docker-compose up --build

Note: You may need sudo to run this command, depending on your system.

Once it's up, open http://localhost:8080 in your browser, and you can test the service right away by submitting a URL.

This Docker setup will spin up a containerized version of the app with everything set up: the Flask backend, the C++ MediaProcessor, and the frontend. Once processing is done, you'll be able to watch the filtered video in the media player.

Option 2: Manual Installation

For those who want more control or are looking to contribute to development, follow these steps to set up Fast Music Remover manually.

Step 1: Ensure Dependencies Are Installed

Ensure all dependencies mentioned in the Prerequisites section are installed before proceeding.

Step 2: Install Python Dependencies

Install the Python dependencies with:

pip install -r requirements.txt

Step 3: Compile the Media Processor (C++)

Navigate to the MediaProcessor directory:

cd MediaProcessor

Make a build directory and navigate into it:

mkdir build
cd build

Run CMake and compile

cmake ..
make

Note: Ensure there are no errors during compilation. If you encounter an error, double-check that all prerequisites are installed.

Step 4: Start the Backend Server

After setting up the dependencies and compiling the C++ project, go back to the project root and start the backend server:

python3 app.py

Note: The server should indicate it is running on http://127.0.0.1:8080. Visit this address in your browser and submit a URL to begin.

License

Fast Music Remover is released under the MIT license. The license can be found in the LICENSE file.

Badges

Extracted from project README's

fast-music-remover