audloop

audloop is a library in development. Its purpose is to enable rapid prototyping of DSP algorithms and/or audio synthesis experimentation in a Python programming environment.

audloop will run all audio DSP offline. Audio is meant to be read in as a wav file or synthesized from scratch. The audible result will be a newly written wav file.

This library aims to achieve similar functionality to that of sound specific languages/programs such as SuperCollider, FAUST, or MAX-MSP. Unlike those programs, audloop is written entirely in the general purpose language Python. This means that nothing is abstracted from the user. Because Python is general purpose, algorithms written in audloop can be ported to other general purpose languages such as C++. To further enable coherency with audloop and realtime DSP environment, all audloop modules are coded as if they are running in a realtime audio loop. This design choice was specifically made so that algorithms written with audloop could be ported into a realtime DSP environment via another general purpose language.

Quick Start

The code below is a suggested use case. Simply create a class that inherits from the AudioProcessorGeneric class. The three methods on_start_up(), process_block(), and post_process() need to be implemented. You need to, at the very least, put pass in each method to get the class to run. The names suggest when each one will be executed at runtime.

Applying A Flanger Effect To An Imported Audio File

from audloop import AudioProcessorGeneric
import audloop as pst


class AudioProcessor(AudioProcessorGeneric):
    def on_start_up(self):
        self.fio.read_wav("EXAMPLE_FILE.wav")  # import wav file
        self.set_process_time(self.fio.length_samples,
                              'samples')  # set the audio loop to be the same length as the imported audio. 
        self.fl = pst.Flanger()  # initialize the Flanger module

    def process_block(self):
        for i in range(self.process_time):
            self.output_stream[i] += self.fl.flanger(self.fio.imported_audio[i], 2, 0.6,
                                                     0.3)  # calling the flanger module in the audio loop 

    def post_process(self):
        self.fio.write_wav("EXAMPLE_OUTPUT.wav", self.output_stream)

Creating A Random Melody Using A Saw Wave

class PTAudioProcessor(AudioProcessorGeneric):
    def on_start_up(self):
        self.set_process_time(10)
        self.s1 = pst.Sawosc()                     # Initialize Sawosc module
        self.en = pst.EnvGen([(0.01, 1),(0.2, 0)]) # Generate a linear envelope from breakpoints
        self.cl = pst.Clock()                      # Intialize Clock
        self.no = pst.WhiteNoise()                 # Initialize Whitenoise

    def process_block(self):
        for i in range(self.process_time):
            f = self.no.whitenoise_range(200, 1000, 2)                                  # Value between 200 and 1000 every 2Hz 
            self.output_stream[i] += self.s1.sawosc(f) * self.en.env(self.cl.clock(2))  # clock will trigger envelope every 2Hz 

    def post_process(self):
        self.fio.write_wav('saw_melody.wav', self.output_stream)

Current Modules Available

AudioProcessorGeneric

Oscillators

Sinosc Cososc Sawosc Triosc Sqrosc

Noise

Whitenoise

Envelopes / Clocks

EnvGen Clock Score

Analysis

RMSDetector PeakDetector

Delays

Delay FDelay

Effects

Flanger

Utility

rad2deg deg2rad lin_pan const_pwr_pan

patrickTumulty / audloop