Introduction

This is an attempt to simulate the acoustic implications of certain types of hearing loss.

The idea is to use a medically obtained audiogram and apply its values to arbitrary audiofiles to give a 3rd person the possibility to get an impression how the owner of that audiogram experiences their world

Install Requirements

• python 3.7 or higher
  • https://www.python.org/downloads/
  • Or use the inbuilt package manager, i.e
    • sudo apt install python3
• scipy and matplotlib
  • Run below command in cmd prompt/powershell/terminal
    • python3 -m pip install scipy matplotlib
• ffmpeg
  • Download from https://ffbinaries.com/downloads and place into same folder as AA.py or on the $PATH.
  • Or via inbuilt package manager, i.e
    • sudo apt install ffmpeg

Usage:

python AA.py -i "path/to/media/file"
             -o "path/to/output/folder"
             -a "path/to/audiogram/file"
             [--anim] 

• Paths can be relative or absolute
• Output folder will be created if it doesn't exist
• --anim is an optional flag, produces an updating animation of the spectrums, however it takes a while to process

Audiogram:

The audio processing is based on an audiogram file.

This file contains 2 arrays, one for the left audio channel, one for the right audio channel.

Each channel array contains a number of frequency/dB pairs which will be applied to the audio signal by the Python script.

The format of the file is:

[
    # Left channel parameters
    [
        [Hz, dB],
        [Hz, dB],
        [Hz, dB],
        [Hz, dB],
        ...,
        ...
    ],

    # Right channel parameters
    [
        [Hz, dB],
        [Hz, dB],
        [Hz, dB],
        [Hz, dB],
        ...,
        ...
    ]
]

Since most audiograms will not contain points beyond ~8 KHz, you'll have to extend the curve beyond that point manually, for example by repeating the last point's dB for 12, 16, 18 and 22 Khz.

You can also increase the attenuation (dB values) gradually if that fits your hearing curve better.

Acknowledgements

Thank you E@gle for your continuous support and aid throughout <3