Parsing Microphone Audio Data in PDM Mode for Edge Impulse Models


We are working on an audio project and need assistance with parsing audio data collected using the IMP34DT05TR (STMicroelectronics) microphone in PDM mode with Edge Impulse models.

Context/Use case:

We are currently collecting audio data from the IMP34DT05TR microphone in PDM mode and successfully capturing the raw audio data. However, we are facing challenges in parsing and preparing this data for use with Edge Impulse models. Specifically, we need guidance on the following:

  1. How to format and preprocess the raw PDM audio data for compatibility with Edge Impulse’s model requirements.
  2. Any specific libraries or tools recommended for this process.
  3. Best practices for ensuring high-quality data input for accurate model performance.

We would appreciate detailed steps or references to relevant documentation to help us integrate our audio data with Edge Impulse models effectively.

Thank you!


You can follow the thread here on the format required (PCM) Data format for audio recogniton



Thank you @Eoin

Will take a look and update!