RTMP parsing with multiple Audio Video Session in the pcap

I have to write a RTMP parser which will handle the packets captured form a RTMP stream on wireshark and i will extract the data from the pcap.

I have gone through the specs ad i am able to understand the handshake process and also able to locate the media in TCP packets but i am confused in case of Multiple Audio/Video session which are interleaved within a single pcap, how we can handle that in the parsing so as make our parser able to parse multiple stream simultaneously. Any uniqueness will be very helpful for the simultaneous parsing of the different RTMP streams.

EDIT (after @Martin Redmond's answer): yeah that I am able to figure out but it seems like some FLV data is being streamed over the RTMp but that FLV header is missing and there seems to be different handshake and FLV data is streaming for same IP with different ports. So, i am not able to fi开发者_如何学Gond if its the real FLV file or only header as if i extract only the header and the other data, i am not able to make a FLV file from it.

Any way to validate or extract the media from that RTMP stream???

The header information for each chunk of data lets you figure out which stream the chunk belongs to. It's not straight forward though. The header information gets compressed and the relevant info may have only been sent at the begining of the stream so you need have a context for each chunk.

The important part is the streamid. Video and audio from the same source will have the same streamid but will have different channel numbers and datatypes.

In the spec. the streamid is referred to as the message stream id (section 6.1.2.1) and is only sent with a type 0 header.