Frame Duration

[encrxpted]

Hi, is there any way to control the length of each frame that is sent to the audio rostopic? I am trying to integrate ROS with WebRTCVAD, which only accepts frames of 10, 20, or 30ms. Thanks!

[knorth55]

can you try throttle node to stabilize the topic frame rate?
http://wiki.ros.org/topic_tools/throttle

[encrxpted]

can you try throttle node to stabilize the topic frame rate? http://wiki.ros.org/topic_tools/throttle

I think that throttle will still send over the same messages, right? Won't each frame will still be the wrong duration since throttle just republishes the ros messages?

[knorth55]

yes, but if /audio topic ha s higher frequency, it is worth trying.
but the sound may not be continuous.
if you are talking about audio_capture node, the frequency of the node loop is controlled by gstreamer, and we currently don't support the frequency change.

[encrxpted]

Then is there any way for me to "resize" the chunks after I receive the audio data?

[knorth55]

currently no. we need to implement it.

[encrxpted]

currently no. we need to implement it.

Is there any way I could accurately implement something like this:
Say that the size of each chunk needs to be n bytes. Could I receive the audio data, and once there are n bytes received, send those n bytes to be processed by WebRTCVAD. Then, I'll do this to the next n bytes I receive and so on, basically just splicing the bytes objects that I receive.

[encrxpted]

Adding to the above:
We are able to bag the audio data and save it as a wav file. Plus, I am able to take audio data from a wav file, splice it into the chunk sizes I need, and successfully use the VAD library to process those chunks. Thus, I know that you can certainly "resize" the chunks via the wav file. My question now is, how can I directly just change the chunk size of the audio data when I receive enough bytes?

[knorth55]

I have no idea, sorry.
I think just concatenating several topics is enough, but I'm not sure.