pysdt 0.1.1

A thin Python wrapper around the Sound Design Toolkit

PySDT: Python bindings for the Sound Design Toolkit

Thin Python bindings for the Sound Design Toolkit (SDT) library.

The bindings consist of two parts:

• C++ wrapper classes (contained in src/wrappers/) encapsulate SDT's C API, especially the manual memory management
• nanobind bindings (contained in src/pysdt.cpp) make the wrappers available in Python

Proper documentation for the Python library will be added at some point. Until then, the SDT documentation should be helpful as well, since the bindings follow the C API and module structure very closely. Getter/setter pairs are usually captured as Python @properties.

Features

Most of the SDT functionality is available in the bindings, except for

• OSC support
• JSON export/import
• anything with a straightforward numpy equivalent, e.g. array means or FFT calculation

Installation

Packaged wheels are available for Linux, MacOS and Windows.

pip install pysdt

Alternatively, you can build the bindings yourself. The SDT is included as a submodule in this repository (with slight modifications to ensure MSVC compatibility). Two submodules of the SDT itself will need to be initialized as well.

git clone [e-mail]:dsuedholt/pysdt.git
cd pysdt && git submodule update --init SDT
cd SDT && git submodule update --init 3rdparty/json-builder 3rdparty/json-parser
cd .. && pip install .

Usage

By default, the global sampling rate is set to 0 Hz. It is important to set the global sampling rate explicitly at the start of your script. If you change the sampling rate later on, you may need to call update() on some objects; check the SDT documentation.

Here's a brief example of using the DCMotor class to generate one second of engine sound:

import pysdt
import numpy as np

sr = 44100

pysdt.common.set_sampling_rate(sr)

# buffer length of an internal comb filter is a constructor argument
motor = pysdt.dcmotor.DCMotor(1024)

# properties bind to getRpm / setRpm etc
motor.rpm = 3000
motor.load = 0.5

result = np.zeros(sr)
for i in range(sr):
    result[i] = motor.dsp()

# we have sound! can now process, play or save it, e.g:
import soundfile as sf
sf.write("engine.wav", result, sr)