Audio_capture not from a device, but straight from wav file (source is a .wav file not microphone)

[antondim]

Hello, as title suggests I'm trying to use a pre-recorded .wav file as audio source and publish it's data as a signed int16[] or byte[] audio stream, to /audio (or /audio/audio) topic.

Documentation seems to ignore this case, even though it's stated in description of "audio_capture" package.

[knorth55]

currently, we do not support the feature, I think.
but we are welcoming your contribution to enhance audio_common repo.

[antondim]

could this be a node that's raised inside a python script?

[knorth55]

what do you mean? I don't get what you wan to say.
audio_capture node is written in c++.
https://github.com/ros-drivers/audio_common/blob/master/audio_capture/src/audio_capture.cpp

[antondim]

It would be easier for me to manipulate audio (devices, stream, data) through python, so I was thinking to call an extra python node via the audio capture launchfile . This node would publish the stream data from python's pyaudio module to a separate topic etc etc . But ok, will look into it in my free time