I have already done proper research, but still lack information on the thing I would like to achieve.
So I would like to program an application where the user can record
At this point, if i had to accept camera ( raw stream ) and immediately make it available to a set of clients, i would go the google hangouts route and use WebRTC. see ondello 'platform section' for the toolset/SDK. During your evaluation, you should have looked at comparative merit of WebRTC v RTSP.
IMO with its statefulness, RTSP will be a nightware behind firewalls and with NAT. AFAIK on 3G/4G the use of RTP in 3rd party apps is a bit risky.
That said, i put up on git an old android/rtp/rtsp/sdp project using libs from netty and 'efflux'. I think that this project was trying to retrieve and play just the audio track within the container ( vid track ignored and not pulled via network ) from Youtube videos all of which were encoded for RTSP at the time. I think there were some packet and frame header issues and i got fed up with RTSP and dropped it.
If you must pursue RTP/RTSP some of the packet and frame level stuff that other posters have mentioned is right there in the android classes and in the test cases that come with efflux