Getting started
From a fresh Python environment to your first audio tensor.
Install Phonotensor
Use Python 3.11 or newer. A virtual environment keeps your audio project’s dependencies isolated. These docs target the published 0.2.2 release.
python -m venv .venv
# macOS / Linux
source .venv/bin/activate
# Windows PowerShell: .venv\Scripts\Activate.ps1
python -m pip install phonotensor==0.2.2NumPy, SciPy, soundfile, and sounddevice are installed as dependencies. The package imports sounddevice at startup, so PortAudio must be available even for file-only workflows. On Debian/Ubuntu, install the runtime with sudo apt install libportaudio2; some systems also need libsndfile1.
Create, shape, and save a signal
Generate a quiet sine wave, add short fades, and write a WAV file. This example runs without recording or playing audio.
from phonotensor import FileIO, ops
# Create a two-second tone. No microphone needed.
audio = ops.sine(440, duration=2.0, amplitude=0.25)
audio = ops.fade_in(audio, duration=0.02)
audio = ops.fade_out(audio, duration=0.1)
FileIO.write(audio, "tone.wav")
print(audio.shape) # (88200, 1)
print(audio.sample_rate) # 44100Choose your workflow
Read the tensor guide to understand shapes and sample rates. The audio I/O guide covers files, recording, and playback. Signal processing covers filters and features; the API reference lists parameters, defaults, and return types.
from importlib.metadata import version
print(version("phonotensor")) # Installed distribution versionUse distribution metadata to check your version. The 0.2.2 wheel currently contains an internal __version__ value of 0.2.1.