Which TMAPTI is correct ? (VobBlanker, DvdReMakePro, FixVTS, PgcEdit)

[jinjin_jp]

I had the problem about output of DVD2one Join-Mode(seamless) like below.
http://img241.imageshack.us/img241/4421/97488689xe1.th.jpg (http://img241.imageshack.us/my.php?image=97488689xe1.jpg)

It seems to be because C_ADT and VOBU_ADMAP is incorrect.
It could be fixed by
(1)VobBlanker
(2)DvdReMakePro
(3)FixVTS
C_ADT and VOBU_ADMAP is same about 3 softwares,
but TMAPTI is different about each 3.

So I tried to rebuild TMAPTI by PgcEdit.
"VobBlanker+PgcEdit"'s and "DvdReMakePro+PgcEdit"'s is same.
But "FixVTS"'s had warning like below and different.
http://img167.imageshack.us/img167/8128/51618613jv3.th.jpg (http://img167.imageshack.us/my.php?image=51618613jv3.jpg)

There are 5 kinds of TMAPTI.
Which TMAPTI is correct ? (VobBlanker, DvdReMakePro, FixVTS, PgcEdit)

originalVB rebuild DRM rebuild FV rebuild
Entry 1 185 369 369 185 369 185 369
Entry 2 555 740 740 555 740 555 740
Entry 3 925 1111 1111 925 1111 925 1111
Entry 4 1449 1449 1449 1296 1449 1449 1449
Entry 5 1732 1918 1732 1732 1732 1732 1732
Entry 6 2104 2289 2104 2104 2104 2104 2104
Entry 7 2474 2660 2474 2474 2474 2474 2474
Entry 8 2845 3031 2845 2845 2845 2845 2845
Entry 9 3216 3402 3216 3216 3216 3216 3216
Entry 10 3587 3772 3587 3587 3587 3587 3587
Entry 11 3959 4144 3959 3959 3959 3959 3959
Entry 12 4329 4515 4329 4329 4329 4329 4329
Entry 13 4700 4885 4700 4700 4700 4700 4700
Entry 14 5071 5256 5071 5071 5071 5071 5071
Entry 15 5441 5627 5441 5441 5441 5441 5441
Entry 16 5812 5998 5812 5812 5812 5812 5812
Entry 17 6183 6369 6183 6183 6183 6183 6183
Entry 18 6555 6741 6555 6555 6555 6555 6555
Entry 19 6925 7111 6925 6925 6925 6925 6925
Entry 20 7295 7479 7295 7295 7295 7295 7295
Entry 21 7664 7850 7664 7664 7664 7664 7664
Entry 22 8034 8219 8034 8034 8034 8034 8034
Entry 23 8404 8590 8404 8404 8404 8404 8404
Entry 24 8775 8961 8775 8775 8775 8775 8775
Entry 25 9145 9331 9145 9145 9145 9145 9145
Entry 26 9516 9701 9516 9516 9516 9516 9516
Entry 27 9886 10072 9886 9886 9886 9886 9886
Entry 28 10258 10443 10258 10258 10258 10258 10258
Entry 29 10628 10813 10628 10628 10628 10628 10628
Entry 30 10998 11184 10998 10998 10998 10998 10998
Entry 31 11370 11554 11370 11370 11370 11370 11370
Entry 32 11740 11927 11740 11740 11740 11740 11740
Entry 33 12111 12290 12111 12111 12111 12111 12111
Entry 34 12476 12661 12476 12476 12476 12476 12476
Entry 35 12847 13032 12847 12847 12847 12847 12847
Entry 36 13217 13403 13217 13217 13217 13217 13217
Entry 37 13589 13774 13589 13589 13589 13589 13589
Entry 38 13959 14144 13959 13959 13959 13959 13959
Entry 39 14329 14515 14329 14329 14329 14329 14329
Entry 40 14701 14901 14701 14701 14701 14701 14701
Entry 41 15100 15300 15100 15100 15100 15100 15100
Entry 42 15499 15699 15499 15499 15499 15499 15499
Entry 43 15935 15935 15935 16147 15935 15935 15935
Entry 44 16345 16345 16345 16546 16345 16345 16345
Entry 45 16745 16745 16745 16943 16745 16745 16745
Entry 46 17142 17142 17142 17340 17142 17142 17142
Entry 47 17539 17539 17539 17738 17539 17539 17539
Entry 48 17936 17936 17936 18333 17936 17936 17936
Entry 49 18333 18333 18333 18731 18333 18333 18333
Entry 50 18731 18731 18731 19130 18731 18731 18731
Entry 51 19130 19130 19130 19530 19130 19130 19130
Entry 52 19530 19530 19530 20127 19530 19530 19530
Entry 53 19929 19929 19929 20525 19929 19929 19929
Entry 54 20326 20326 20326 20922 20326 20326 20326
Entry 55 20723 20723 20723 21319 20723 20723 20723
Entry 56 21121 21121 21121 21717 21121 21121 21121
Entry 57 21518 21518 21518 22314 21518 21518 21518
Entry 58 21917 21917 21917 22712 21917 21917 21917
Entry 59 22314 22314 22314 23111 22314 22314 22314
Entry 60 22712 22712 22712 23510 22712 22712 22712
Entry 61 23111 23111 23111 23907 23111 23111 23111
Entry 62 23510 23510 23510 24504 23510 23510 23510
Entry 63 23907 23907 23907 24900 23907 23907 23907
Entry 64 24306 24306 24306 25300 24306 24306 24306
Entry 65 24700 24700 24700 25698 24700 24700 24700
Entry 66 25099 25099 25099 26096 25099 25099 25099
Entry 67 25499 25499 25499 26692 25499 25499 25499
Entry 68 25896 25896 25896 27091 25896 25896 25896
Entry 69 26294 26294 26294 27491 26294 26294 26294
Entry 70 26692 26692 26692 27889 26692 26692 26692
Entry 71 27091 27091 27091 28290 27091 27091 27091
Entry 72 27491 27491 27491 28654 27491 27491 27491
Entry 73 27889 27889 27889 28853 27889 27889 27889
Entry 74 28290 28290 28290 29052 28290 28290 28290
Entry 75 28453 28454 28453 29253 28453 28453 28453
Entry 76 28985 28853 28985 29453 28985 28985 28654
Entry 77 29383 29253 29383 29784 29383 29383 29052
Entry 78 29784 29652 29784 29983 29784 29784 29453
Entry 79 30183 30183 30183 30183 30183 29852

Regards.

[r0lZ]

Obviously, there is a problem inside a cell. I can't say which program is right, but PgcEdit is probably wrong.
Try to demux everything, and remux with muxman. The problem should disappear, but verify carefully if there are no A/V synchro problems.

[jinjin_jp]

Thanks, r0lZ.
I tried to demux and remux. (Set audio delay of Muxman according to what is shown by Pgcdemux.)
Output of (1),(2) and (3) are all same about file size, VOBU_ADMAP and TMAPTI.
But file size and VOBU_ADMAP are different from before processed. I can't judge which is correct about (1),(2) and (3). Demux/remux is correct ?

Regards.

[r0lZ]

I suppose that muxman has reorganized the V/CIDs correctly.
IMO, the file size is not a good criteria. The number of nav packs may be different for example. (I gave noticed that if you remux a movie shrunk with DVDShrink, the new VOB files are significantly shorter.)

[jinjin_jp]

Thanks again, r0lZ.

I think the reason of the problem seems to be DVD2one broke VOB as you pointed.
(File size is decreasing when joining by DVD2one.)
I tried anothor method, i.e; replacing by VOB which is before joined by DVD2one.
(1)replace by VobBlanker.
(2)replace by VobBlanker, and rebuild TMAPTI by PgcEdit.
(3)replace by DvdReMakePro.
(4)replace by DvdReMakePro, and rebuild TMAPTI by PgcEdit.

*C_ADT and VOBU_ADMAP is all same.
*TMAPTI is same about (2), (3) and (4). Only (1) is different.
I think DvdReMakePro seems to adjust when replacing, because it shows in log "different stream...".

I think (2) and (3) is reliable method for fixing DVD2one's output.

Regards.

[r0lZ]

Note also that having some little differences in TMAPTI is probably harmless. The value in the table are rounded. The rounding method might be different, and produce slightly different numbers. You don't have to worry, as long as you can seek to the correct duration.

[jinjin_jp]

I looked at IFO once more.
TMAPTI of (1)replace by VobBlanker seems to have a big problem.
Sectors which are described in TMAPTI don't exist in VOBU_ADMAP.
So others are all same.

Regards.

[r0lZ]

Good to know...
Thanks!

[jinjin_jp]

Please forgive me to add,

About TMAPTI of (1')replace by VobBlanker after "SCAN", sectors which are described in TMAPTI exist in VOBU_ADMAP.
It seems to be better to "SCAN" previously, even if all cells are replaced.

Regards.

Add:Sorry, (3) is diffent from what rebuild, too.
(1) (1') (3)
replace rebuild
VB VB DRM PgcEdit
(NoScan)(Scan)
Entry 1 140 369 185 369
Entry 2 507 740 555 740
Entry 3 877 1111 925 1111
Entry 4 1401 1450 1297 1450
Entry 5 1685 1919 1733 1733
Entry 6 2055 2290 2105 2105
Entry 7 2425 2661 2475 2475
Entry 8 2797 3032 2846 2846
Entry 9 3168 3403 3217 3217
Entry 10 3539 3773 3588 3588
Entry 11 3909 4145 3960 3960
Entry 12 4280 4516 4330 4330
Entry 13 4651 4886 4701 4701
Entry 14 5021 5257 5072 5072
Entry 15 5393 5628 5442 5442
Entry 16 5764 5999 5813 5813
Entry 17 6135 6370 6184 6184
Entry 18 6506 6742 6556 6556
Entry 19 6877 7113 6927 6927
Entry 20 7247 7483 7298 7298
Entry 21 7616 7855 7669 7669
Entry 22 7986 8225 8040 8040
Entry 23 8356 8596 8410 8410
Entry 24 8727 8968 8782 8782
Entry 25 9096 9338 9152 9152
Entry 26 9468 9709 9523 9523
Entry 27 9838 10080 9894 9894
Entry 28 10209 10451 10266 10266
Entry 29 10579 10821 10636 10636
Entry 30 10951 11192 11006 11006
Entry 31 11321 11562 11378 11378
Entry 32 11692 11935 11748 11748
Entry 33 12063 12299 12120 12120
Entry 34 12428 12670 12485 12485
Entry 35 12798 13041 12856 12856
Entry 36 13170 13412 13226 13226
Entry 37 13540 13783 13598 13598
Entry 38 13912 14154 13968 13968
Entry 39 14280 14525 14339 14339
Entry 40 14652 14911 14711 14711
Entry 41 15051 15310 15110 15110
Entry 42 15450 15709 15509 15509
Entry 43 15887 15945 15945 15945
Entry 44 16297 16357 16357 16357
Entry 45 16695 16758 16758 16758
Entry 46 17093 17157 17157 17157
Entry 47 17491 17556 17556 17556
Entry 48 17888 17955 17955 17955
Entry 49 18285 18354 18354 18354
Entry 50 18683 18754 18754 18754
Entry 51 19082 19153 19153 19153
Entry 52 19480 19553 19553 19553
Entry 53 19880 19953 19953 19953
Entry 54 20278 20352 20352 20352
Entry 55 20675 20751 20751 20751
Entry 56 21072 21151 21151 21151
Entry 57 21470 21550 21550 21550
Entry 58 21868 21950 21950 21950
Entry 59 22266 22349 22349 22349
Entry 60 22664 22748 22748 22748
Entry 61 23063 23148 23148 23148
Entry 62 23461 23548 23548 23548
Entry 63 23858 23947 23947 23947
Entry 64 24257 24347 24146 24347
Entry 65 24652 24742 24546 24742
Entry 66 25051 25141 24942 25141
Entry 67 25449 25542 25342 25542
Entry 68 25847 25941 25742 25941
Entry 69 26245 26340 26141 26340
Entry 70 26642 26739 26541 26739
Entry 71 27041 27139 26939 27139
Entry 72 27443 27539 27339 27539
Entry 73 27840 27937 27738 27937
Entry 74 28241 28338 28138 28338
Entry 75 28605 28833 28501 28501
Entry 76 29001 29232 29033 29033
Entry 77 29399 29633 29432 29432
Entry 78 29798 30032 29833 29833
Entry 79 30232 30232

[jinjin_jp]

I noticed the difference between each software about TMAPTI is caused from each feature.
(1)VobBlanker: go to next entry at *.27 second.
(2)DvdReMakePro: go to next entry at *.15 second.
(3)PgcEdit: go to next entry at *.21 second.
(4)FixVTS: go to next entry at *.15 second. (*.27 when also exist *.27.)
http://img112.imageshack.us/img112/2488/42605591yi0.th.jpg (http://img112.imageshack.us/my.php?image=42605591yi0.jpg)

Regards.

[blutach]

I think the entries work on the basis of a maximum of 2048 per time map, and so the differential in seconds is set that way. mpucoder posted the correct formula about a year ago and I believe this has been adopted in VobBlanker and PgcEdit.

Regards

[r0lZ]

jinjin_jp, did you your tests on NTSC or PAL DVDs?

[jinjin_jp]

@r0lZ
It's on NTSC.

@blutach
I believe this has been adopted in VobBlanker and PgcEdit
So, it is strange that VobBlanker and PgcEdit have different TMAPTI.
I'll confirm once more later.

Regards.

[jinjin_jp]

I tried to examine TMAPTI feature, but it was too dificult for me.
Below will not involve all cases, I'll saw partly.

I compared
Source which is Re-Authored by DVDShrink.
Output which processed it by VobBlanker and DvdReMakePro.
Output which processed each 3 output by FixVTS and by PgcEdit.

(1)FixVTS does not modify TMAPTI anything. (It will modify only when VOB itself was fixed.)

(2)Source:left is more used as TMPTI entry point than right.
(0.16～1.00)＞(0.01～0.15)seconds
0.28>0.16, 0.29>0.17, 1.00>0.18

(3)PgcEdit: It is very similar to (2). The only difference is 0.18＞1.00
(0.16～1.00)＞(0.01～0.15)seconds
0.28>0.16, 0.29>0.17, 0.18＞1.00 ---1 of below Fig.

(4)VobBLanker:
(0.22～1.05)＞(0.06～0.23)
1.05＞0.23

(5)DvdRemakePro: I can't understand. It seems to have no rule.
Sometimes (1.01～1.04)＞(0.17～0.19) like (4). ---2 of below Fig.
Sometimes (1.01～1.03)<(0.18～0.20) like (2). ---3 of below Fig.

Please see with figure http://img165.imageshack.us/img165/3323/32894179rm1.th.jpg (http://img165.imageshack.us/my.php?image=32894179rm1.jpg)
("○"　means TMAPTI entry point of each output of software.)　

Regards.

[jinjin_jp]

I think the entries work on the basis of a maximum of 2048 per time map, and so the differential in seconds is set that way. mpucoder posted the correct formula about a year ago
The post which you say is this?
http://forum.doom9.org/showthread.php?p=823346&highlight=formula#post823346

Regards.

[r0lZ]

Theoretically, the entry in the time map should point to the nav pack (or VOBU) which is the more closer to the time represented by the entry. However, as anyway it's only an approximation, the player should find the right frame before or after the given nav pack, and start playing from that point. Also, don't forget that if the time unit is greater than 1 second (as it is for the main movie for example) and the user wants to jump to a time that is not exactly divisible by the time unit, the player will have to find the correct frame in another VOBU.

The computation is not really easy, as you have to take the delay at the beginning of the video stream into account, as well as the STC discontinuities when the VOB ID changes (and verify the errors like the one you have reported at the beginning of this thread.) Therefore, trying to guess which is right and which is wrong is not easy, and make little sense. As far as the entry points near the correct time, you can assume that it is right.

Personally, I may be wrong, but I consider that it is better to point before the right time than after, as the player has to read the I frame that is before the first frame to play anyway. Therefore, the player might be a little bit faster if it jumps directly to the right I frame.

I don't know how you have computed the timings in your tests, but the 14.00 is probably only a rounded value (especially in NTSC.) If that time is exact, then PgcEdit should probably have selected that point instead of the previous one. But if it is a little bit less, or if you don't have taken the starting delay into account, PgcEdit is right, given its computation method. Anyway, a good player will jump to the correct time without problem in both cases. Don't worry!

[r0lZ]

The post which you say is this?
http://forum.doom9.org/showthread.php?p=823346&highlight=formula#post823346
See this one: http://forum.doom9.org/showthread.php?t=111436

[jinjin_jp]

@r0lZ, Thanks for the explanation and the referenced URL.
Theoretically, the entry in the time map should point to the nav pack (or VOBU) which is the more closer to the time represented by the entry.
Also, don't forget that if the time unit is greater than 1 second
I don't know how you have computed the timings in your tests,
I thought TMPTI entry point is selected from the boundary of VOBUs.
I used IfoEdit and VobBlanker.
I read TMAPTI entry point by IfoEdit.
VobBlanker's preview of Prev/Cut can show both Time and sectors, and I confirmed every time of each VOBU boundary.
In above figure, all the time (like "12.24, 13.06, 13.18, 14.00, more) is from VobBlanker, and "○" (TMAPTI entry point) is from IfoEdit.

About above example (not first problematic example of DVD2one's output),
time unit is 1 second,
it is consist of 1 cell, so it will has no STC discontinuities,
and it doesn't have Audio/Video delay which I confirmed by PgcDemux.

It is very difficult to understand for me.
Anyway, I understood there is no problem even if little differece about saoftware exist. Thanks.

Regards.

[jsoto]

... as you have to take the delay at the beginning of the video stream into account...

VobBlanker is not taking into account that thing, only the number of frames.

The main reason is if you cut the begining movie with VobBlanker or DVDSrink, where the PTSs are not modified, resulting in an "not-true" delay in the video stream (in fact, in all the streams)

... as well as the STC discontinuities when the VOB ID changes..
Again, VobBlanker does not take this into account. Only the number of frames, assuming the video is played smothly, which usually is not the case...

jsoto