The goal is to attempt to synthesize tones from a defined sequence of swars, while retaining as simple a notation as possible.
I had to install python3-pyaudio to get PyAudio to work.
On Fedora and similar:
sudo dnf build-dep python3-pyaudio
On Debian systems, replace dnf with apt-get
Each swar defines a note. A collection of swars sound melodic when they are destributed across beats, with one or more swars to a beat, also known as a matra. In this project, each beat is separated by a ','.
The seven basic swars (Sa,Re,Ga,Ma,Pa,Dha,Ni) are represented a (s,r,g,m,p,d,n). A swar on the immediate lower octave is represented by prefixing a '<', and that on the immediate higher octave with a '>'. Example, <n,s,g,m,d,n,>s start from the lower Ni and ends on the upper Sa.
The above seven swars are all shuddh. The non-shuddh swars are represented by suffixing any single character. I typically use '^'. So komal Re is 'r^'.
Call the swar_synthesizer.convert() function with the path to the input file, and optionally the base frequency, which will define your scale. (Default=440Hz)
I have a main.py function that allows me to call the script from the command line as follows:
python main.py -i "PATH/TO/INPUT"