This Python-based tool extracts and visualizes musical features from MIDI and WAV files. It provides insights into melody, rhythm, chords, and spectral analysis using advanced signal processing and MIDI analysis techniques.
- Melody Extraction: Detects pitch contour, note onset times, and durations.
- Chord Detection: Extracts chords and labels them with their names and positions.
- MIDI Rhythm Extraction: Identifies beat positions and normalizes them into a 4-beat measure.
- Audio Rhythm Extraction: Uses
librosa.beat.beat_track()to detect tempo and beat positions.
- Piano Roll with Chord Labels:
- Notes are displayed with blue bars.
- Chords are labeled above their onset points.
- Time Grid for Beat Visualization:
- Displays detected beats as red markers on a structured time grid.
- Spectrogram & MFCCs for Audio:
- Visualizes spectrogram (STFT) and MFCCs for audio feature analysis.
- Zoom Feature: Adjust visualization scale dynamically using a slider.
- View Selector Dropdown:
- Toggle between Piano Roll, Time Grid, and Spectrogram.
- MIDI Playback:
- Extracted melodies can be played using
pygame.midi. - Implements note on/off but needs precise timing refinement.
- Extracted melodies can be played using
- Read and parse MIDI files (extract notes, velocity, tempo, time signature)
- Extract rhythmic/melodic features
- Extract spectral features from audio for texture analysis
- Generate simple classifications (e.g., rhythm type, key detection)
- Train/test models for rhythm & melody style classification
- Implement real-time tagging of input patterns
- Develop UI for users to view analysis results
- Implement rhythm variation algorithm (Markov chains, VAEs)
- Implement melody transformation (style-aware modifications)
- Implement texture-based sound analysis
- Build GUI to allow user interaction
- Implement MIDI export for DAW integration
- Optimize UI for usability
- Refine MIDI Playback Timing
- Ensure accurate note durations.
- Improve note-off timing for a natural playback feel.
- Implement a tempo-sync feature to align playback speed with detected tempo.
- Improve Beat Quantization & Grid Alignment
- Enhance rhythm accuracy for extracted beats.
- Add bar subdivisions for better beat visualization.
- Implement adjustable quantization levels (e.g., 16th-note, triplets, etc.).
- Optimize Spectrogram & Beat Grid Overlays
- Allow simultaneous spectrogram & MIDI playback.
- Overlay detected beats on the spectrogram for better rhythm analysis.
- Add interactive controls to toggle between different analysis layers.
- Export MIDI & Audio Analysis Results
- Add functionality to save extracted MIDI data.
- Export beat timing data as a text file or MIDI.
- Implement batch processing for multiple files at once.
- Enable Real-Time Processing (Stretch Goal)
- Implement real-time rhythm extraction from microphone input.
- Visualize live MIDI performance in the piano roll.
- Develop a real-time MIDI loop generator to generate variations on the fly.
- Machine Learning-Based Enhancements
- Train a deep learning model to suggest chord progressions based on melody.
- Implement style-transfer for melody & rhythm generation.
- Develop AI-assisted MIDI improvisation based on user input.
- Advanced DAW Integration
- Enable drag-and-drop MIDI export for seamless DAW workflow.
- Support Ableton Link for real-time tempo synchronization.
- Provide MIDI effect processing options (quantization, swing, etc.).
- Python 3.x
- Required Libraries:
pip install mido music21 librosa pygame numpy matplotlib PyQt5
python main.py- Click 'Choose File' to select a MIDI or WAV file.
- Use the View Selector to choose between Piano Roll, Time Grid, or Spectrogram.
- Adjust the Zoom Slider to modify the visualization scale.
- Click 'Extract Features' to analyze the file.
- Use 'Play MIDI' to listen to extracted melodies.
Contributions are welcome! If you want to improve this tool, feel free to fork this repo and submit a pull request.
MIT License