The goal of this application is to allow users to upload various sound files(.wav) and approximate the percentage of specific noise presences within the audio clips.
- On a Unix system while in the project's root directory:
Ensure FFTW is installed from https://www.fftw.org/download.html
On Unix once installed add it to the audio-processing directory and run these commands (ONLY NEED TO DO THIS ONCE) in its directory:
$ ./configure
$ make
$ make installNote: might have to run sudo make depending on your permissions.
Next compile the main c++ and python files from the project's root directory:
$ make
$ ./Audio-Insight
$ python python_scripts/main.py- Non-Unix instructions coming soon.
- Uses C++ for audio processing
- Uses python for audio analysis
- No plans for hosting as the program will be too data($) intensive so it will be a downloadable executable
- Plan is to run on large .wav files so will be data intensive
- Using C++ for data processing to expedite process
- Currently communicating between python and C++ files via binary files, may change in the future
- Visualizing the data in an efficient manner
- Currently using python for visualization (librosa), but this is slower than desired
- Bundle all commands into single executable (will differ based off OS I believe)
- Add band-pass filter (for outliers)
- Add distinct noise filtering
- Allow for spectrogram visualizations
Contributions are welcome. If you use this code in your own project please just provide credit.
MIT License