GNN Audio Generation (Step 15)

Purpose

Generates audio representations (sonification) of GNN models, mapping model structures, state transitions, and matrix values to sound parameters using SAPF, Pedalboard, and other audio backends.

Key Commands

# Run audio generation
python src/15_audio.py --target-dir input/gnn_files --output-dir output --verbose

# As part of pipeline
python src/main.py --only-steps 15 --verbose

API

from audio import (
    AudioGenerator, process_audio, generate_audio_from_gnn,
    create_sonification, analyze_audio_characteristics,
    SAPFGNNProcessor, SyntheticAudioGenerator,
    check_audio_backends, generate_oscillator_audio,
    apply_envelope, mix_audio_channels,
    convert_gnn_to_sapf, generate_audio_from_sapf
)

# Process audio step (used by pipeline)
process_audio(target_dir, output_dir, verbose=True)

# Generate audio from GNN model
result = generate_audio_from_gnn(gnn_content, output_dir="output/")

# Create sonification
audio_data = create_sonification(model_data)

# Use AudioGenerator class
gen = AudioGenerator()

# Convert GNN to SAPF format and generate audio
sapf_code = convert_gnn_to_sapf(gnn_content)
generate_audio_from_sapf(sapf_code, output_path="model.wav")

# Check available backends
backends = check_audio_backends()
# Returns: {'librosa': {'available': True}, 'soundfile': {...}, 'pedalboard': {...}}

API

from audio import ( AudioGenerator, process_audio, generate_audio_from_gnn, create_sonification, analyze_audio_characteristics, SAPFGNNProcessor, SyntheticAudioGenerator, check_audio_backends, generate_oscillator_audio, apply_envelope, mix_audio_channels, convert_gnn_to_sapf, generate_audio_from_sapf ) # Process audio step (used by pipeline) process_audio(target_dir, output_dir, verbose=True) # Generate audio from GNN model result = generate_audio_from_gnn(gnn_content, output_dir="output/") # Create sonification audio_data = create_sonification(model_data) # Use AudioGenerator class gen = AudioGenerator() # Convert GNN to SAPF format and generate audio sapf_code = convert_gnn_to_sapf(gnn_content) generate_audio_from_sapf(sapf_code, output_path="model.wav") # Check available backends backends = check_audio_backends() # Returns: {'librosa': {'available': True}, 'soundfile': {...}, 'pedalboard': {...}}

Gnn Audio Generation