Log in Register FAQ Memberlist Search pcHDTV Forum Index
pcHDTV Forum

pcHDTV Forum Index -> General pcHDTV topics -> pchdtvr 1.0-rc11 Goto page Previous  1, 2, 3
Post new topic  This topic is locked: you cannot edit posts or make replies. View previous topic :: View next topic 
PostPosted: Wed Mar 15, 2006 10:06 am Reply with quote
tupari
 
Joined: 29 Dec 2004
Posts: 38




Guess what? atscut also core dumped.

(gdb) bt
#0 0x08048df7 in huffman_decode (d=0xbfb0d420 "Javier; D4L; Beyonchok�",
s=0xbfb0d620 "���\222c\035����\202\202��ܥ/e�L@\235\003o\237\017�001ɠ", dlen=489, slen=35, comp=2) at atscut.c:1113
#1 0x08050c99 in test_ett_mss (
r=0xaa1b62d "���\222c\035����\202\202��ܥ/e�L@\235\003o\237\017�001ɠ\027�\004�, '�' <repeats 123 times>, n=29, etmid=197194) at atscut.c:4455
#2 0x08051403 in test_ett (p=0x805f700 "GRK\027", n=29) at atscut.c:4611
#3 0x080548a5 in test_mgt_tt_packet (p=0x805f700 "GRK\027", pid=4683)
at atscut.c:5789
#4 0x080551b0 in test_packet (p=0x805f700 "GRK\027") at atscut.c:6121
#5 0x0805582d in read_loop () at atscut.c:6417
#6 0x08058588 in main (argc=5, argv=0xbfb0da24) at atscut.c:7772


BTW you don't have to quote every post you reply to.
View user's profile Send private message Visit poster's website
double yikes!
PostPosted: Wed Mar 15, 2006 6:25 pm Reply with quote
inkling
 
Joined: 05 Feb 2004
Posts: 342




What is WB doing to their users up there in NY?

Looks like atscut got farther with it. Can see actual readable text in the destination string. Almost like they swapped the Title for the Description tables...

Go ahead and stick a return at the top of atscut.c huffman_decode too. May have to add a command line option to enable it, if the stations are going to be non-compliant like that.

Same/similar code in both programs so it shouldn't crash simply because huffman decode returns a blank string.

It should then let you finish the -kmgt extract. It looks big, like 8M or more maybe, but bzip2 turns it into about 20-30k, something that is easily mailed.

Pretty odd that the top market in the country has broken PSIP. Guess early-adopter phase is still in effect for WB.

KHOU CBS transmitter here has odd 29 second glitch.

Hopefully you can get that -kmgt extract file to me so I can see what exactly is going on over there at WPIX.

KHWB here had a hard time getting PSIP working for all of 2004, but since mid-2005 has been one of the best program guides in town. 5 days with no huffman compression.

I'll go test some more locals for their huffman compliance.

Maybe we should make an ATSC Compliance Report form to send to all the stations in our areas to let them know that we know when they are running outside of specification.


Thanks for the feedback,

-ink




tupari wrote:
Guess what? atscut also core dumped.

(gdb) bt
#0 0x08048df7 in huffman_decode (d=0xbfb0d420 "Javier; D4L; Beyonchok�",
s=0xbfb0d620 "���\222c\035����\202.
View user's profile Send private message
PostPosted: Wed Mar 15, 2006 7:52 pm Reply with quote
tupari
 
Joined: 29 Dec 2004
Posts: 38




Alright my emails to you always bounce, but you can get it at :

http://www.tupari.net/WPIX.cap-00.ts.bz2
View user's profile Send private message Visit poster's website
my email borked again?
PostPosted: Wed Mar 15, 2006 8:22 pm Reply with quote
inkling
 
Joined: 05 Feb 2004
Posts: 342




Thanks for the heads-up on the email issue. Did not know.

I put range checking in huffman decode to fail gracefully if the lookup table index goes crazy.

Will get that file and start dissecting it now.

Can't wait to see what they're doing wrong.

Will let you know soon.

Other than huffman crashing things, is it working OK on the non-huffman stations?


-ink
View user's profile Send private message
Re: my email borked again?
PostPosted: Wed Mar 15, 2006 8:31 pm Reply with quote
tupari
 
Joined: 29 Dec 2004
Posts: 38




inkling wrote:

Other than huffman crashing things, is it working OK on the non-huffman stations?


-ink


no core dumps since I turned off epg for WPIX. I also noticed that the program guide for WNYW is showing some things out of order:

0 .0 Wed 20:00 60 Bones [TV-14-D-L-V] Brennan winds >
1 .0 Wed 21:00 30 American Idol [TV-G] One of the 12 finalis>
2 .0 Wed 18:30 30 Geraldo at Large n/a >
3 .0 Wed 21:30 30 The Loop [TV-14-D-L] Sam, Sully and L>
4 .0 Wed 19:00 30 The Simpsons [TV-PG] Homer gets caught up>
5 .0 Wed 19:30 30 Seinfeld [TV-PG] George hopes to meet>


Is that their fault or pchdtvr's fault?
View user's profile Send private message Visit poster's website
WPIX
PostPosted: Wed Mar 15, 2006 8:36 pm Reply with quote
inkling
 
Joined: 05 Feb 2004
Posts: 342




atscut -a255 WPIX.ts | grep "bad huffman"

ETT-1D pg[0CE9] ETMID 0003024A [L23 C02 M00] bad huffman tree offset
ETT-1D pg[10E9] ETMID 0004024A [L23 C02 M00] bad huffman tree offset
ETT-1D pg[0CE9] ETMID 0003024A [L23 C02 M00] bad huffman tree offset
ETT-1D pg[10E9] ETMID 0004024A [L23 C02 M00] bad huffman tree offset
ETT-1D pg[0CE9] ETMID 0003024A [L23 C02 M00] bad huffman tree offset
ETT-1D pg[10E9] ETMID 0004024A [L23 C02 M00] bad huffman tree offset
ETT-1D pg[0CE9] ETMID 0003024A [L23 C02 M00] bad huffman tree offset
ETT-1D pg[10E9] ETMID 0004024A [L23 C02 M00] bad huffman tree offset
ETT-1D pg[0CE9] ETMID 0003024A [L23 C02 M00] bad huffman tree offset
ETT-1D pg[10E9] ETMID 0004024A [L23 C02 M00] bad huffman tree offset




A search for Javi, since we know that one died, gives:

ETT-1D VER 12 SLen 039 CNI 1 SecN 00 LSecN 00 etmID 0004024A
ETT-1D pg[10E9] ETMID 0004024A [L23 C02 M00] bad huffman tree offset
[eng] Huffman
[Javier; D4L; Beyonchok ]

OK, so the range check catches it and tosses it now instead of crashing. Looks like vast majority of your huffman data is good, except that one ETT.

I'll see if it's a typo in the specification (tables are copy/paste) in case I fat-fingered anything.

I'll try to get latest code posted to nop.org tonight.

Thanks for your help,

-ink
View user's profile Send private message
quick fix for huffman
PostPosted: Wed Mar 15, 2006 10:32 pm Reply with quote
inkling
 
Joined: 05 Feb 2004
Posts: 342




Hey your FOX screws up the EIT order too eh? pchdtvr EPG shows it in received order, so it's their fault.

Must be a national problem. It used to be OK, then one day they switched and it's been screwy ever since.

In Program Guide, [s] key to sort. It has a few sort options, just keep hitting [s] key to cycle through them.

Here's a quick fix for pchdtvr huffman decode, but it's not as stringent a range check as what will be posted to nop.org shortly.

You'll want to change asyslog() to fprintf() for atscut.





/* should be pointing to correct tables now */
for (i=0; i < (slen<<3); i++)
{
/*-------------------- new -----------------------------------------*/
if (p > 127) { /* sanity check, don't stray outside table */
/* change to asyslog if it hits too often */
asyslog( LOG_INFO, "%s BAD Tree Offset %02X", __FUNCTION__, p);
return;
}
/**********************************************************/
/* get tree offset for char p from order-1 tree byte offset table */
to = bo[ p ];




It should be obvious where it goes. p>127 is range check.

Your WB station has a whopping 10 days of guide. 2 bad descriptions out of 694 events is not too bad.

I used pchdtvr -r WPIX.ts to replay it and the guide was all there and looked ok with the exception of the junk on javi.

Cheers!

-ink
View user's profile Send private message
pchdtvr 1.0-rc11
  pcHDTV Forum Index -> General pcHDTV topics
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
All times are GMT - 7 Hours  
Page 3 of 3  
Goto page Previous  1, 2, 3
  
  
 Post new topic  This topic is locked: you cannot edit posts or make replies.  


Powered by phpBB © 2001-2003 phpBB Group
Theme created by Vjacheslav Trushkin