This library is already unmaintained. Please check the following new library that is in development.

https://github.com/hassaku/audio-plot-lib

---

Converts a line graph to sound and returns an object that can be played in Jupyter notebook or Google Colab.

Values are represented by pitches, and the timeline is represented by left and right pans.

It was created to make data science fun for the visually impaired.

Dependency

ffmpeg library is needed to provide a descriptive guide to the graph.

Google Colab

No additional installation is required.

Linux

$ sudo apt-get install libavformat-dev libavfilter-dev libavdevice-dev ffmpeg

Others

It's under investigation. If you know of any, please make a pull request. Thanks.

Install

$ pip install audio-plot  # !pip install audio-plot for Colab or Notebook

Usage

See and run demo notebook also.

y = np.sin(np.arange(0, np.pi*2, 0.1))
yy = np.array([y, -y]).T
plot(yy)  # audio control will be appeared on notebook.

[The audio is as follows for this example]

tts > minimum value is -1.0
(Low sinusoidal sound)
tts > maximum value is 1.0
(High sinusoidal sound)
tts > line 1
(Sound in response to changes in the line 1 graph)
tts > line 2
(Sound in response to changes in the line 2 graph)

Example Results Page

Supplement

If using Jupyter Notebook with a screen reader is inconvenient for you, you may want to consider using the following NVDA add-on.

https://github.com/mltony/nvda-browser-nav

Use in script

from pydub.playback import play

# generate graph sound
audio = audio_plot.plot(two_inverted_sin)

# play
play(audio)

# save to audio file
audio.export("graph.wav", format="wav")