Audio acquisition for API and model training

Hello, I would like to ask about the functionality of Edge Impulse so that I could use it in my project. I want to use a microcontroller with a microphone to record and send audio training data, update the model and push it back onto the microcontroller (e.g. for new users to create custom commands). Is it possible using Edge Impulse API?
If yes, in Data acquisition format, audio files are to be encoded in binary payloads. I am trying to sample a short .wav audio file, append CBOR header, termination and signature. Will the model be trained with these payloads? Can they coexist with regular .wav audio files? Any recommendations on how to sample the audio?
Thank you for reading and have a nice day

Hi @kotsatka,

In theory, CBOR and wav files should be able to coexist in a project. That being said, it’s probably much easier if you stick to a single file type. If you use the ingestion API, you can upload files directly, so long as Edge Impulse supports that file type (.wav is OK). See here for how to use the ingestion API: Ingestion API