StereoPhaseNet: Phase Correction for Stereo Audio Using Deep Learning

Introduction

StereoPhaseNet is a deep learning project aimed at correcting phase issues in stereo audio recordings. The project focuses on creating a dataset of out-of-phase stereo audio stems and their in-phase counterparts, and then training a neural network to convert out-of-phase audio to in-phase audio. This ensures a balanced and full sound when played in both stereo and mono setups, enhancing the overall audio quality.

Project Description

Motivation

In audio production, phase issues between the left and right channels of stereo recordings can lead to undesirable effects such as frequency cancellations and a hollow sound. These problems become especially prominent when the audio is played back in mono. StereoPhaseNet aims to solve this problem by leveraging deep learning techniques to automatically correct phase discrepancies.

Objectives

Data Collection: Gather a diverse set of high-quality stereo audio files to create a comprehensive dataset.
Data Preparation: Process the audio files to generate pairs of out-of-phase and in-phase audio samples.
Model Development: Design and train a transformer-based neural network to learn the mapping from out-of-phase to in-phase audio.
Evaluation: Assess the model's performance using appropriate metrics and fine-tune it for optimal results.
Deployment: Develop an inference pipeline that can be used to correct phase issues in new audio recordings.

Features

Audio Processing: Efficient loading, normalization, and framing of stereo audio files.
Deep Learning: Transformer-based model tailored for audio phase correction.
Training and Evaluation: Comprehensive training loop with data loaders and evaluation metrics.
Inference Pipeline: Ready-to-use pipeline for applying the trained model to new audio data.

Getting Started

Follow the setup instructions to create the environment and install all necessary dependencies. Then, use the provided scripts to preprocess your audio data, train the model, and evaluate its performance.

Contributing

Contributions are welcome! Feel free to open issues or submit pull requests to enhance the functionality of this project.

License

This project is licensed under the MIT License. See the LICENSE file for more details.

By correcting phase issues in stereo audio recordings, StereoPhaseNet aims to improve the listening experience and ensure high-quality sound reproduction in any playback setup.

StereoPhaseNet Project Setup Instructions

Step 1: Create a New Anaconda Environment

Open your terminal (or Anaconda Prompt on Windows).
Create a new environment named stereosync with Python 3.8:
```
conda create --name stereosync python=3.8
```
Activate the new environment:
```
conda activate stereosync
```

Step 2: Install Required Packages

Librosa: For audio processing.
PyTorch: Deep learning framework.
Other dependencies: NumPy, etc.

Install the packages using the following commands:

```sh
conda install -c conda-forge librosa 
conda install ffmpeg pytorch torchvision torchaudio cudatoolkit=11.3 -c pytorch -c nvidia 
```

Step 3: Verify Installation

Verify the installation of each package:

python -c "import librosa; print('librosa installed')"
python -c "import torch; print('PyTorch installed')"

Step 4: Optional - Jupyter Notebook

If you plan to use Jupyter Notebook for experiments and development:

Install Jupyter:
```
conda install -c conda-forge notebook
```
Launch Jupyter Notebook:
```
jupyter notebook
```

Summary of Commands

Here’s a summary of all the commands to set up your environment and install the necessary packages:

```sh
# Create and activate the new environment
conda create --name stereosync python=3.8
conda activate stereosync

# Install required packages
conda install -c conda-forge librosa
conda install ffmpeg pytorch torchvision torchaudio cudatoolkit=11.3 -c pytorch -c nvidia

# Optional: Install Jupyter Notebook
conda install -c conda-forge notebook

# Verify installations
python -c "import librosa; print('librosa installed')"
python -c "import torch; print('PyTorch installed')"
```

By following these steps, you’ll have a fully set up environment ready for your StereoPhaseNet project.

karisigurd4/StereoPhaseNet