Hello,
I would like to know if it is possible to recover the raw audio data buffer in order to apply post-processing to it before it is sent via the peerConnection.
Without alterations to support processing the audio device/manager, not currently.
But it is indeed possible and may be supported in the future.