Audio Utils : a Python library for classic audio utils

generate_sine_waves(filename, length, sr=44100, freqs=[(0.2, 440)])

This function generates a wav file of sample rate sr. It is composed of addition of the freq present in the list of tuples freqs, each tuple being a couple (Amplitude, Frequency) [amp in [0,1] - freq in Hz]

Arguments :
----------
	  - filename : path and file name to write the resulting wav
	  - length : length in seconds
	  - sr : sample rate (default 44.1 kHz)
	  - freqs : a list of couple (Amplitude, Frequency) w/ amp in [0,1]

WARNING !!!

this is a deprecated way of creating a sine wav it is using a lot of memory WAVTABLE would be better

find_n_harmonics(data, sr, n, samples_per_slice=1024, overlap=100, n_local_max=5, local_steps=1, local_min_height=1):

Function that extracts the n main frequencies of a wav file

This function extracts the N main harmonics from a WAV file it proceeds by slicing the audio for short term freq domain conversion extracts the local max and then proceed to the global max search

  Arguments:
  ----------
         - data: a numpy array that contains the audio data (MONO)
         - sr : sample rate
         - n : number of harmonics to retrieve

         - samples per slices : for computing FFT (default=1024)
         - overlap : windowing (default=100)
         - n_local_max : number of harmonics to store for each slice (default=5)

         - local_steps, local_min_height : parameters for the find_peak function. 
           If no harmonic is found, tweak those param

filter_n_harmonics(data, sr, harmo, Q=50):

Function that writes a wav file containing the audio data filtered by bandstop filters centered on the freq stored in harmo

   Arguments:
   ----------
          - data: a numpy array that contains the audio data (MONO)
          - sr : sample rate
          - harmo : an array containing the center frequencies of the filters
          - Q: quality factor of the filters

Examples

filename = './440.wav'
sr, data = wav.read(filename)

############ process stereo to mono (ToDO) try: channels = data.shape[1] except IndexError: channels = 1

if channels > 1: print("WARNING Stereo File Detected.... we keep only one channel") data = data[:, 0] print(f"SR={sr} - Nb Channels={channels}")

########### finding the n main frequencies nb_harmo = 2 main_harmo = find_n_harmonics(data, sr, nb_harmo, local_min_height=1, local_steps=1)

########### filtering res = filter_n_harmonics(data, sr, harmo=main_harmo)

######## writing filtered signal wav.write('./res.wav', sr, res)