Alternative to reading m4a files

I am currently having another go at Julia and decided to do a small project working on audio data. Currently running into the situation of not being able to read m4a files. These are coming from my OSX setup and I am doing all my julia stuff on my ml linux box. I understand libsndfile (@ssfrr) will not read these so I am wondering what is the best format to have these saved in python? I know I read these files python and export them into different formats. Is wav my best approach? I’d rather use a compressed format but still being able to load it as audio-esque files, as I would want to perform MFCC, etc. I guess I am aiming to as much audio manipulation as possible in Julia.

For starters, I was currently thinking of saving them all as wav 22khz stereo files and loading these up in Julia, anyone else have other suggestions? :slight_smile:

I’ve used https://github.com/dancasimiro/WAV.jl successfully in the past, but I’d be curious to hear if other people have had success with compressed audio.

For most of my audio analysis work I use FLAC files because they’re compressed but lossless. They’re supported by LibSndFile.jl.

There’s a list if supported file types here:

If you want loss compression Ogg Vorbis is a good option.

3 Likes