Guitutor - Real-Time Performance Feedback

Welcome to Guitutor - Guitar Learning System! This project is designed to help beginners learn guitar by providing real-time feedback on their performance. The system allows users to select a song, which is processed and broken down into its musical components using a Convolutional Neural Network (CNN) model. Users can then play along with the deconstructed song, and their performance is analyzed using an AI classification model (Random Forest Classifier - RFC) and music-based algorithms to identify the chords played. The system compares these chords to the original song and returns feedback on the user's performance.

Features

• Song Selection: Users can select a song to practice.
• Real-Time Feedback: Provides real-time feedback on user performance.
• Chord Analysis: Identifies chords played by the user and compares them to the original song.
• Performance Analysis: Analyzes user performance using an AI classification model.
• User-Friendly Interface: Easy-to-use interface for beginners.

Usage

1. Upload an Audio File: Users upload an audio file of the song they want to practice.
2. Analyze and Save Data: The system processes the file, breaking it down into its musical components and saving the data on the server.
3. Play Along: Users play along with the deconstructed song.
4. Record and Upload Performance: The system records the user's performance and uploads it for analysis.
5. Receive Feedback: The system analyzes the performance, identifies the chords played, compares them to the original song, and provides feedback.

Installation

1. Clone Repository:

   git clone https://github.com/Dvora-K/Guitutor.git
   

2. Navigate toGuitutor
3. Run in Docker platform

     docker-compose build
     docker-compose up

Technologies Used

• Server-Side:
  • Python (Flask)
  • Libraries and Tools:
    • librosa, torch_audio, wavio, Pydub, Scipy
    • CNN and RFC Model
• Client-Side: React, TypeScript, SCSS

Some Screenshots