demuxed m2v shows wrong real-time length, good frame count

[psantucc]

Greetings All-
I suspect the solution to my problem is simple, but I have been unable as yet to

find an answer in the many fine guides and forums available. After substantial

research on my own, I am at last able to present a single technical question to

this group, in the hope that a reader knows the answer. I bet it'll be slk001 or

defense B-}=.

I am trying to back up Beauty and the Beast Special Edition , R1 NTSC. My goal is

a DVD5 with the special edition, all three soundtracks, both subtitles, and

working chapter points, but no menus or extras. Since the elementary streams

themselves are small enough to fit a DVD5 with room to spare, I want to avoid

trascoding and the commensurate quality loss.

Using SmartRipper, I have been able to extract valid files for all of the streams

in which I am interested and the chapter points. I can remux and reauthor them in

IFOEdit and get a file set that plays properly in PowerDVD XP, including

observation of chapter points.

The Problem is the reported length of the m2v file and any vobs made from it. The

movie iteself is 91 minutes long, but WMP 6.4 and PowerDVD XP both report it as

72 minutes.

A preview operation in DVD2AVI on the m2v reports 95% FILM with both progressive

and interlaced frames.

It doesn't take long to count the frames (131220, if memory serves) and do the

math; if the file is treated as 23.576 fps, it will show 91 minutes - if 29.97

fps, 72 minutes. Something in one of my steps is preventing players from

understanding how to count the length, and I don't know how to find it.

PowerDVD XP is tolerant - though it reports a wrong time (as though the file were

to be counted in 29.97 frames) it can continue to play and chapter operations

succeed. My standalone player is not tolerant. Though a DVD made from this

fileset will play, any calculation dependant operation (such as chapter search,

scan, or even OSD of time elapsed) will cause the player to freeze.

Any ideas?

[fingernailX]

Get pulldown.exe and pulldown gui (Under Tools on the left frame) Under options set drop_frame_flag for NTSC set to false

I sometimes have to do this in MainConcept Encoder when reencoding an m2v to fit 4.3 gb. It would be great if all mpeg programs would calculate the time based on number of frames X framerate but they don't.

see http://mpucoder.kewlhair.com/DVD/mpeghdrs.html under GOP for full description of drop frame flag

[psantucc]

Thanks for the tip, fingernailX. I got closer than I have ever been, but no cigar yet; I hope you or others might have a next step. It is very frustrating to be missing only one piece of information and have no idea where to find it...

Using pulldown.exe set for drop_frame true or drop_frame false, windows media player is convinced that the resultant m2v is film, and shows accurate time code and permits scrubbing. PowerDVD has no difference in its treatment from before the pulldown operation; no total time count is shown, and the drag bar stops at about the point where the end would be if the stream really were true 29.97 NTSC.

Authoring a DVD in IFOEdit using the pulled down m2v yields the same result as not pulled down; neither PowerDVD nor my standalone treat the runtime properly, and the standalone chokes on that as a result.

Examining the IFO file of the re-authored DVD reveals that IFOEdit thought the file was the wrong run length on the PGC:
PGC (program chain):
[00000002] Number of Programs 22 [16]
[00000003] Number of Cells 22 [16]
[00000004] Playback time (BCD) 17978325 [011253d5]
Playback time (hh:mms.frame) 01:12:53.15 / 30 fps

But editing the IFO to match the original playback time does not fix the playback problems. I tentatively conclude that the proper information is missing from the m2v, and IFOEdit authors based on that.

So the question now is: what do PowerDVD and IFOEdit want that isn't in my m2v? It isn't the drop frame flag, and I've tried restoring the rff flags with pulldown.exe at the same time (even though DVD2AVI's report of 95% film should indicate that they were there already). What the hell do I have to do to identify the m2v stream properly?

As a corrollary, doesn't the fact that the information was somehow lost indicate a flaw in SmartRipper's demux?

[dhluke]

Hmm... IFOEdit can't handle Pulldown flags, therefor it can't do NTSC...

it sounds like DVD2ONE would do a nice job here... Heck, even ripping the DVD in IFO mode (DVD Decrypter) and fixing the IFOs will do. it'll probably take 10-20 mins to rip and 10 seconds to fix the IFOs... and 15-60 mins to burn...

ohh and powerdvd -- that thing sucks for reading time... even windvd sucks at it... I never trust any of those programs for reading time.

dhluke