I am looking to do some integration with watson for text to speech. Not sure if anyone here is familiar with it but I was wondering if it possible to send a binary stream of the audio within a call to a 3rd party resource. I dont mind build a middle man app that would allow me to implement the following 

It seems watson has to be notified of the audio pt so if that would be acquired somehow that would be helpful. Also I don't wan't to deal with removing rtp headers or decryption DTLS if possible.

 Please advise