Join Date: Oct 2002
Location: Bakersfield, CA
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
OK, now for my mega-dollar question:
I have been playing with this STB Firewire TS for awhile and have some questions and some assumptions which might need dispelling. The SA3250HD STB shows everything free and clear which may or may not be true in reality. The TS of all digital SD stations will record / playback and display live w/o dificulty. I have six (6) HD channels: NBC (43), ABC (48), DISC-HD (55), TNT-HD (56), HBO-HD1 (790) and HBO-HD2 (791) (I might not quite have my channel #'s quite right but no matter). HBO-HD2 (791) captures great. HBO-HD1 (790) comes in about 50% when viewing live (audio and video starts and stops). ABC-HD, NBC-HD, TNT-HD, and DISC-HD I can't get at all except for some occassional snippets of audio (live viewing). Recorded files from these are just "black". You would think I could get ABC-HD, NBC-HD, TNT-HD or DISC-HD before either of the HBO-HD channels.
Not knowing exactly what the problem is, I decided I needed to educate myself some. I read up on Transport Streams and how they are constructed. I now know enough to be dangerous. I understand the terms (PSI, PAT CAT, PMT, NIT, PID, PCR, etc.). I understand how the header and payload of each 188 byte packet is constructed.
I downloaded TSReader Lite to analyze some of the streams. I recorded 1 minute files from some of the SD channels for reference and all six of the HD channels. I opened multiple instances of TSReader and analyzed the different files and compared. The SD files analyze OK as does the one HD channel (HBO-HD2 (791)) that also views and records OK. In fact, the SD's and the one good HD files looked very much the same. Oddly, though, they all show about 260 PAT's and only 2 PMT's for a 1 minute file although the bit rate for the PAT's and PMT's are reported the same. The PAT's and PMT's are marked with "*" meaning continuity errors. According to my reading (the dangerous part), PAT's and PMT's should be coming through at least every 0.5 second. That would mean a minimum of 120 PAT's and 120 PMT's. Very few recognized PMT's might explain slow channel change response and channel change lockups since the decoder can't lock-on until at least one PAT AND one PMT come through.
The 1 minute recorded file for HBO-HD1 (790) (the one that comes through about 50%) will analyze but shows lots of continuity errors (41,789 of them to be exact) and hundreds of "stray" PID's. Since these are SPTS (Single Program Transport Steams), there will be a limited and low number of valid PID streams and values in this application. Besides the PAT (always PID=0x0000) and the PMT (PID=0x1394), I show only one audio PID stream and video PID steam for the "good" channels. The 50% HBO-HD1 (790) channel shows also shows two defined PID streams (1 - video and 1 - audio) from the PMT but has all these "stray" packets and continuity errors. The "stray" packets can not be null packets as these should not even be in these non-ATSC type streams and even if present, would all have the same value of 0x1111 (or perhaps 0x1fff). Besides, the valid and defined video and audio streams show continutity errors (missing or wrongly ordered packets) which I take are all the stray packets that are found.
As expected, ABC-HD, NBC-HD, TNT-HD and DISC-HD analyze even worse. While I can tell what should be the video stream PID and the audio stream PID by virtue of the relative packet quantities displayed, the number of discontinuities and "stray" packets is over the top. TSReader does not even find a PAT or PMT in the analysis so does not know and is unable to display what the valid stream PID values are. In fact, TSReader seems to think it is a DSS stream.
So, finally, my question is this: Am I seeing corruption or is this encryption? I do not understand enough about DTCP to know where the encryption layer is implemented at. I would think that encryption would be at the packet level or more precisely at the packet payload level and that the fundamental structure of the Transport Stream would remain intact. Or does DTCP encryption scramble everything including the TS structure?
For anyone wanting to know more about TS, I found "broadcastpapers.com" to be very helpful; technical enough but not overly. Also, try TSReader Lite for looking at problem streams and compare this to good streams. Very informative and free! (not news for some I'm sure)
This may not be a good place to pose this question and get comments. If there is a better thread to place this post, please advise.
Wrong information is worse than no information.