[RESOLVED] The red thin line

[Guest 2]

There must be something wrong about "The red thin line".

If I try to index it, I get:

Code: Select all

DGAVCIndexFileNV14 DGIndexNV 2042 X32
D:\eseguibili\media\dgdecnv\

E:\raw\2_50 sottile linea rossa, La\BDMV\STREAM\00800.m2ts 40924200960

DEVICE 0
DECODE_MODES 1,1,1
STREAM 1
PKTSIZ 192
VPID 4113
CLIP 0 0 0 8
RANGE 0 0 40924200959 0
AUDIO 1100,1101,1102,1103,1104,1105

SPS 993
PPS 1040
RPS 1134
IDR 1142
0:FRM 7 0 0 1
1:FRM 6 1 0 1
2:FRM 5 2 0 1
3:FRM 6 3 0 1
4:FRM 6 4 0 1
5:FRM 5 5 0 1
6:FRM 6 6 0 1
7:FRM 6 7 0 1
8:FRM 5 8 0 1
9:FRM 6 9 0 1
10:FRM 6 10 0 1
11:FRM 5 11 0 1
12:FRM 6 12 0 1
13:FRM 6 13 0 1
14:FRM 5 14 0 1
15:FRM 6 15 0 1
16:FRM 6 16 0 1
17:FRM 5 17 0 1
18:FRM 6 18 0 1
19:FRM 5 20 0 1
SIZ 1920 x 1088 
FPS 24000 / 1001
CODED 20
PLAYBACK 20
0.00% FILM
ORDER 0

http://www.mediafire.com/?jyc5spqbjphljg9

[Guest 2]

FYI - It works if you demux the video stream with TSMuxer first.

Thanks but it's a "bug" report, not a help request.

[admin]

That's bizarre. I'll check it out ASAP.

[admin]

OK, I have it fixed. I'll slipstream it tonight or tomorrow morning.

[Guest 2]

OK, I have it fixed. I'll slipstream it tonight or tomorrow morning.

Which kind of error was it?

[admin]

I have slipstreamed the fix. Please re-download 2042 and update DGIndexNV.

The problem: I had some code that tried to detect and correct the case of video start codes split across packets. But when the main buffer got refilled after running out of data, the pointers I was using became invalidated. That caused general havoc and the get NALU function returned an error during indexing, which caused an abort.