Strange pack arrangement in VOB file (menushrink does not like it!)

[jeanl]

Guys,
Check this out:

http://img77.echo.cx/img77/3751/foo9bp.png (http://www.imageshack.us)

This is a still menu with audio created by TMPGenc DVD Author. This LBA is in the middle of a cell. Notice the first video pack with a GOP header, then 9 video packs that contain only 0's (except for the pack header and video stream start code)!!!
Then the 10th video pack contains the I-frame!

MenuShrink complains because of the absence of a GOP header, then goes on to produce a useless VOB (unless you select the first I-frame, in which case everything works OK).

I'm simply wondering... Is this strange layout legit (I'm guessing it is). It's highly unusual, (first time I've seen that)... Have you see that kind of stuff before?

jeanl
EDIT: The problem is that MenuShrink replaces the first I-frame with an I-frame pack that does not have a GOP header. It only has a picture header, and a picture coding extension (no sequence extension, no sequence display extension). I'm guessing that the resulting MPEG stream isn't valid, because one of these extensions must be present in the first video frame (but which ones)...

This is annoying because ideally, menushrink should check for that... :(

[mpucoder]

Send me a copy, if you can (you know where) I'd like to see what the verifier thinks. And is this a bug in VobEdit, or is the pts missing? Specs say it should be in the first video pes header of the vobu. But I don't see it there, or the I frame pack.

[jeanl]

Originally posted by mpucoder
Send me a copy, if you can (you know where) I'd like to see what the verifier thinks. And is this a bug in VobEdit, or is the pts missing? Specs say it should be in the first video pes header of the vobu. But I don't see it there, or the I frame pack.
I checked, the pes header in the first video pack of the VOBU does not have a PTS. I'm sending you a snippet of it...
I muxed the menu with muxman (originally) but I'm pretty sure TMPGenc DVD Author demuxed and remuxed because the audio went from ac3 to LPCM! :(

jeanl

[mpucoder]

That seems to be a problem with IfoEdit, the pts is in the I-frame pack.
The verifier did not complain about the zeroes, but it did complain a lot about other things. For one the color space is PAL in an NTSC resolution/framerate. And lots of buffer overflows (all those zeroes get buffered initially, until the I frame is decoded). Here's some of the output:

>>> [DVD] WARNING 3376 (ref. DVD-3 5.4.1.2.1 (6)) :
Sequence_display_extension : colour_primaries value is 5,
which is different from the default values 4 or 6
for video sequence (0) at byte 91 bit 0;
PES ($E0) byte 113 (byte 113 of packet 0);
byte 127 of pack 1 (PS stream byte 2175).

>>> [DVD] WARNING 3376 (ref. DVD-3 5.4.1.2.1 (6)) :
Sequence_display_extension : transfer_characteristics value is 5,
which is different from the default values 4 or 6
for video sequence (0) at byte 92 bit 0;
PES ($E0) byte 114 (byte 114 of packet 0);
byte 128 of pack 1 (PS stream byte 2176).

>>> [DVD] WARNING 3376 (ref. DVD-3 5.4.1.2.1 (6)) :
Sequence_display_extension : matrix_coefficients value is 4,
which is different from the default values 5 or 6
for video sequence (0) at byte 93 bit 0;
PES ($E0) byte 115 (byte 115 of packet 0);
byte 129 of pack 1 (PS stream byte 2177).

>>> [DVD] ERROR 3460 (ref. DVD-3 5.4.1.2-1) :
picture_header: vbv_delay value (0x5A85) must be 0xFFFF for DVD
for video sequence (0) at byte 128 bit 5;
PES ($E0) byte 150 (byte 150 of packet 0);
byte 164 of pack 1 (PS stream byte 2212).

>>> [MPEG] ERROR 1449 (ref. MPEG Systems 2.4.5.1 | 2.5.2.3) :
PES STD buffer overflow for stream 0xE0, at decoding time 1054.867 [msec] of AU 25 (1095 bytes, buffer size 237568)
in PES stream 0xE0 at byte 838424 bit 0 (byte 2034 of packet 412);
byte 2048 of pack 495 (PS stream byte 1015808).

>>> [MPEG] ERROR 1449 (ref. MPEG Systems 2.4.5.1 | 2.5.2.3) :
PES STD buffer overflow for stream 0xE0, at decoding time 1058.122 [msec] of AU 25 (3120 bytes, buffer size 237568)
in PES stream 0xE0 at byte 840458 bit 0 (byte 2034 of packet 413);
byte 2048 of pack 497 (PS stream byte 1019904).

>>> [MPEG] ERROR 1449 (ref. MPEG Systems 2.4.5.1 | 2.5.2.3) :
PES STD buffer overflow for stream 0xE0, at decoding time 1059.744 [msec] of AU 25 (5145 bytes, buffer size 237568)
in PES stream 0xE0 at byte 842492 bit 0 (byte 2034 of packet 414);
byte 2048 of pack 498 (PS stream byte 1021952).

>>> [MPEG] ERROR 1449 (ref. MPEG Systems 2.4.5.1 | 2.5.2.3) :
PES STD buffer overflow for stream 0xE0, at decoding time 1061.367 [msec] of AU 25 (7170 bytes, buffer size 237568)
in PES stream 0xE0 at byte 844526 bit 0 (byte 2034 of packet 415);
byte 2048 of pack 499 (PS stream byte 1024000).

>>> [MPEG] ERROR 1449 (ref. MPEG Systems 2.4.5.1 | 2.5.2.3) :
PES STD buffer overflow for stream 0xE0, at decoding time 1063.000 [msec] of AU 25 (9195 bytes, buffer size 237568)
in PES stream 0xE0 at byte 846560 bit 0 (byte 2034 of packet 416);
byte 2048 of pack 500 (PS stream byte 1026048).

[mpucoder]

Just wondering, does TMPGenc DVD Author have any kind of project file, preferrably text? I might be able to write an import routine for Muxman - design on TMPGenc DVD Author and mux in Muxman.

[jeanl]

Yes it does, but it's binary. I can send you an example, if you want...
jeanl

[mpucoder]

I'll pass on that. Binary is too hard to figure out what means what, especially without the program (which I'm not terribly interested in, either) and subject to change.