Spaces:
Sleeping
Sleeping
File size: 4,101 Bytes
4ca706f c1e08a0 4ca706f c1e08a0 4ca706f c1e08a0 4ca706f c1e08a0 4ca706f c1e08a0 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 |
---
title: Improvisation Lab
emoji: 🎵
python_version: 3.11
colorFrom: blue
colorTo: purple
sdk: gradio
sdk_version: 5.7.1
app_file: main.py
pinned: false
license: mit
---
# Improvisation Lab
A Python package for generating musical improvisation melodies based on music theory principles. The package specializes in creating natural-sounding melodic phrases that follow chord progressions while respecting musical rules, with real-time pitch detection for practice feedback.
Improvisation Lab Demo
https://github.com/user-attachments/assets/a4207f7e-166c-4f50-9c19-5bf5269fd04e
## Features
- Generate melodic phrases based on scales and chord progressions
- Support for multiple scale types:
- Major
- Natural minor
- Harmonic minor
- Diminished
- Support for various chord types:
- Major 7th (maj7)
- Minor 7th (min7)
- Dominant 7th (dom7)
- Half-diminished (min7b5)
- Diminished 7th (dim7)
- Intelligent note selection based on:
- Chord tones vs non-chord tones
- Scale degrees
- Previous note context
- Real-time pitch detection with FCPE (Fast Context-aware Pitch Estimation)
- Web-based and direct microphone input support
## Prerequisites
- Python 3.11 or higher
- A working microphone
- [Poetry](https://python-poetry.org/) for dependency management
## Installation
```bash
make install
```
## Quick Start
1. Create your configuration file:
```bash
cp config.yml.example config.yml
```
2. (Optional) Edit `config.yml` to customize settings like audio parameters and song selection
3. Run the script to start the melody generation and playback (default is web interface):
```bash
make run
```
- To run the console interface, use:
```bash
poetry run python main.py --app_type console
```
4. Follow the displayed melody phrases and sing along with real-time feedback
### Configuration
The application can be customized through `config.yml` with the following options:
#### Audio Settings
- `sample_rate`: Audio sampling rate (default: 44100 Hz)
- `buffer_duration`: Duration of audio processing buffer (default: 0.2 seconds)
- `note_duration`: How long to display each note during practice (default: 3 seconds)
- `pitch_detector`: Configuration for the pitch detection algorithm
- `hop_length`: Hop length for the pitch detection algorithm (default: 512)
- `threshold`: Threshold for the pitch detection algorithm (default: 0.006)
- `f0_min`: Minimum frequency for the pitch detection algorithm (default: 80 Hz)
- `f0_max`: Maximum frequency for the pitch detection algorithm (default: 880 Hz)
- `device`: Device to use for the pitch detection algorithm (default: "cpu")
#### Song Selection
- `selected_song`: Name of the song to practice
- `chord_progressions`: Dictionary of songs and their progressions
- Format: `[scale_root, scale_type, chord_root, chord_type, duration]`
- Example:
```yaml
fly_me_to_the_moon:
- ["A", "natural_minor", "A", "min7", 4]
- ["A", "natural_minor", "D", "min7", 4]
- ["C", "major", "G", "dom7", 4]
```
## How It Works
### Melody Generation
The melody generation follows these principles:
1. Notes are selected based on their relationship to the current chord and scale
2. Chord tones have more freedom in movement
3. Non-chord tones are restricted to moving to adjacent scale notes
4. Phrases are connected naturally by considering the previous note
5. All generated notes stay within the specified scale
### Real-time Feedback
Pitch Detection Demo:
https://github.com/user-attachments/assets/fd9e6e3f-85f1-42be-a6c8-b757da478854
The application provides real-time feedback by:
1. Capturing audio from your microphone
2. Detecting the pitch using FCPE (Fast Context-aware Pitch Estimation)
3. Converting the frequency to the nearest musical note
4. Displaying both the target note and your sung note in real-time
## Development
### Running Lint
```bash
make lint
```
### Running Format
```bash
make format
```
### Running Tests
```bash
make test
```
## License
MIT License
## Contributing
Contributions are welcome! Please feel free to submit a Pull Request.
|