I would like to split up an audio stream (32bit floats) to chunks, encode them with a lossy compression and then decode it in the other side.
I have no experience with audio codecs, so I started to look for solution. What I've found is Opus and Vorbis. I was not able to find out the substantial difference between them.
The audio I would like to trasfer is mostly speech, but it may contain some other sounds too. The program may run in an environment where the computational power and network bandwidth is limited.
From this image it looks like that Opus can scale better and can compress audio more (by reducing quality) if for example the network bandwidth is low.