Chapter Points in DVD2DVD Don't Match Original

[mr_edavid]

I've recently started using the DVD2SVCD tool to make DVD backups. The quality looks great. Lately, I've been alternating between Instant Copy and DVD2DVD to make my backups. This is where I found a slight problem.
I noticed the chapter points in DVD2DVD are slightly off the original (maybe 10-15 seconds). Can you recommend an adjustment to solve this?

Thanks.

[r6d2]

Hi, mr_edavid, and welcome to the forum.

You're right, they don't match perfectly. But 10-15 seconds is quite a bit too much. Try checking your VCX Builder log file. It will show lines like this:

--DEBUG: requested entry point (id=entry-chapt-02) at 433.160000, closest possible entry point at 433.299533

The explanation is that I-frames are needed for an entry point, and they are not always where you need them. But they should be about within 1 second or so, unless you are using really, and I mean really long GOPs (not recommended).

You may have spotted a problem but before crying wolf, wolf!, do some search on the topic and also see if your 10-15 second differences are really that big.

If the situation subsists, then read the stickies and post everything we may need to help you.

[mr_edavid]

r6d2,

I was not exaggerating when I posted a discrepancy of 10-15 seconds on the chapter points. In fact it's even greater than that. I recently completed doing a backup of BLADE II. I compared chapter 4 in each DVD (original and mine) and noticed a discrepancy of 2 min., 15 S.

I did some digging up and noticed that the suggested GOP settings based on the guides here (http://www.doom9.org/mpg/dvd2svcd.htm) and the DVD2SVCD tool are different. The guide states GOP Sequence M = 3 M/N= 5, however on the DVD2SVCD encoder tab, under the advanced setting dropdown is set for GOP Sequence M = 3 M/N= 4. Could this be the source of my frustrations? Please advise. Thanks

[r6d2]

Originally posted by mr_edavid
The guide states GOP Sequence M = 3 M/N= 5, however on the DVD2SVCD encoder tab, under the advanced setting dropdown is set for GOP Sequence M = 3 M/N= 4. Could this be the source of my frustrations?
Not likely. GOPs are 15 and 12 frames in those cases, respectively, which means about one 10th of a second difference.

Again, please read the stickies, post your log, and also include the VCX build log.

[mr_edavid]

R6D2,

OK. I've re-read all of the stickies. and did not find an appropriate solution to my issue on hand. A VCX build log was not generated because I am using Scenarist to generate my image, not VCDXBuild.

Can you please recommend a solution? Is this solution "solution" to my issue so obvious to you that it's not worth your time to share. I never made any changes to the DVD2DVD tool once it was installed. I took all of the defaults.

[DDogg]

Easy there, mr_edavid :) This ain't your paid tech support :) (I wish it were!)

I think r6d2 is saying that this problem has never been reported before. It will take one of us hours of work to retrace and verify your findings. Before we do that, please do another encode and verify it also has these timing discrepancies.

If you have already double verified this issue then let me know and I'll try to find time to check it out. I did make sure D2S saw your report. He replied that he spent a lot of time on that area and thought he got it as accurate as possible. He did say he would take a look at it if additional reports came in.

[mr_edavid]

My apologies for coming off as brash and impatient. I'm just hoping one of us has come across something like this and would be kind enough to share their solutions. From what I gather through hours of searching through previous posts, FAQs, and stickies, nobody has made a mention of a similar scenario. Maybe I am just being too much of a perfectionist (in trying to duplicate a DVD 100%).

Other than visual verification of chapter point mismatches, how can I verify these chapter points to show the discrepencies? Please advise.

Can somebody verify a chapter on one your backups if this discrepancy is only through a fault of mine and not DVD2DVDs? Thanks.

Before the advent of Instant Copy, I followed TRILIGHTS method in making DVD backups. One of the steps involved prior to encoding was to take the chapter points and multiply the number of frames by a factor of 80% (I think). This was done on *.ecl file. I can't recall who came with the solution.

Thanks, again.

[DDogg]

Maybe I am just being too much of a perfectionist Only in the sense that you may have found something that others didn't care about enough to see. Dunno, but I will try to look at this soon.

[r6d2]

Originally posted by mr_edavid
OK. I've re-read all of the stickies. and did not find an appropriate solution to my issue on hand.
The idea was that you learned how to do a good post (http://forum.doom9.org/showthread.php?s=&threadid=19566). Shameful enough, you have just proved you are able to read but unable to understand and to follow simple instructions.

If you had understood, you'd have posted you log file in the first place, which had shown me you do not use VCX Build. I perhaps might have infered it from the "DVD2DVD" name, but I don't have a cristal ball and if I had one, I'd be fortune telling making myself a millionaire and not moderating a forum for free, having to deal with tantrums of people like you, which fortunately are the minority.
Can you please recommend a solution? Is this solution "solution" to my issue so obvious to you that it's not worth your time to share.
I really don't know why, when in doubt, you prefer to attibute me mean intentions. This may have something to do with the "difficulty to understand" thing.

If I knew what the solution was of course I'd have given it to you. You will not find that kind of behaviour while I'm moderating this forum, even if the answer was obvious.

You violated already several rules on this thread: R1, R4, R12 (at its best) and R16. And you will not see this kind of behaviour in this forum either. So, considering you sort of apologized for your impatience I'll give you only 2 strikes and not 4, which would suspend you immediately.

One of the steps involved prior to encoding was to take the chapter points and multiply the number of frames by a factor of 80% (I think). This was done on *.ecl file. I can't recall who came with the solution.
This is done when converting from 30-fps to 24-fps. If the differences on chapter locations are sistematically 25% displaced to the right, this may turn out to be the cause of your problem.