pyAudioTest

Prerequisites

• Python 3.x
• FFmpeg: This tool is used for audio preprocessing. Please ensure it is installed and accessible in your system's PATH.
  • On macOS: brew install ffmpeg
  • On Linux (Debian/Ubuntu): sudo apt-get install ffmpeg
  • On Windows: Download from ffmpeg.org and add to PATH.

Installation

1. Clone the repository:

   git clone <repository_url>
   cd <repository_name>

2. Install Python dependencies:

   pip install -r requirements.txt

Usage

To classify an audio file, run the audio_classifier.py script:

python src/audio_classifier.py path/to/your/audiofile.wav

Arguments:

• input_audio_file: (Required) Path to the input audio file (e.g., WAV, MP3, MP4, MOV).
• --model_path: (Optional) Path to a specific pyAudioAnalysis SVM model file (e.g., path/to/your/svm_rbf_sm). The script attempts to auto-detect the default speech/music model (svm_rbf_sm) included with your pyAudioAnalysis installation. If this detection fails, or if you want to use a custom model, you must provide this path.
• --output_dir: (Optional) Directory where the output CSV and .dcsx files will be saved. Defaults to the same directory as the input audio file.

Example:

Assuming you are in the project root directory:

python src/audio_classifier.py sample_audio/test_sample.wav --output_dir output_files

(Ensure you create an output_files directory first if you use this example, or point to an existing one. The sample_audio/test_sample.wav is provided in the repository.)

Outputs

The script generates two files in the specified output directory (or the input file's directory by default):

1. <input_filename>_logic_pro.csv:

   • A CSV file formatted for import as markers into Logic Pro X.
   • Columns: Marker,Position,Name
   • Position is the start timecode of the segment (HH:MM:SS.mmm).
   • Name is the classification label (e.g., "speech", "music").

2. <input_filename>.dcsx:

   • A JSON file containing structured information about the classified segments.
   • Includes the source audio filename and a list of cues, each with:
     • CueID
     • StartTimecode
     • Label (classification)

Running Tests

The project includes unit tests to verify basic functionality. Ensure you have installed dependencies first.

To run the tests, navigate to the project root directory and run:

python -m unittest discover -s tests

Troubleshooting

Model File Not Found

The script relies on pyAudioAnalysis and its pre-trained models (e.g., svm_rbf_sm for speech/music classification).

• Automatic Detection: The script first tries to automatically locate the default models from your pyAudioAnalysis installation.
• Manual Override: If the automatic detection fails (you'll see an error message like "Model file not found..."), or if you wish to use a different model, you must use the --model_path argument to provide the full path to the model file.

Finding your pyAudioAnalysis models:

The models are usually located within the pyAudioAnalysis/data directory of your installed pyAudioAnalysis package. The exact location can vary depending on your Python environment and installation method.

1. You can try to find the pyAudioAnalysis installation path (specifically the directory containing the data folder) by running this Python snippet:

   import os
   try:
       import pyAudioAnalysis.audioSegmentation
       # The 'data' directory is typically alongside audioSegmentation.py
       paa_module_dir = os.path.dirname(pyAudioAnalysis.audioSegmentation.__file__)
       print(f"pyAudioAnalysis module directory (contains 'data' folder): {paa_module_dir}")
       print(f"Potential model path: {os.path.join(paa_module_dir, 'data', 'svm_rbf_sm')}")
   except ImportError:
       print("pyAudioAnalysis.audioSegmentation not found. Is pyAudioAnalysis installed correctly?")

   Then, look for a data subdirectory within the printed path. The speech/music SVM model is typically named svm_rbf_sm.
2. If you cloned the pyAudioAnalysis repository for development, the models are usually in pyAudioAnalysis/pyAudioAnalysis/data/.

Once you locate the svm_rbf_sm file (or your custom model file), provide its full path to the --model_path argument.