Requesting feedback on DVD Shrink

[dvdshrink]

Hello,

I'm afraid it has been many years since I visited this forum.

Would anybody be able to give me a quick rundown on known DVD Shrink problems?

I'm aware that there is some new kind of copy protection which DVD Shrink cannot handle, but I do not know any specific details. Also I understand there may be a problem with ISO file output, many years ago I was asked to change something about this, but I cannot remember what.

(I should add that I have no plans to release a new version, I'm requesting this information because I am working now on a related project).

Any feedback would be really appreciated.

[dialysis1]

First off I would like to say it's great to see you back.
I have no real issues with Shrink although I would like to see a program that can combine titles into one title instead of two. It would be great for the so called flipper discs.
As of ISO's, I think creating an ISO with the proper handling of the layer break would be nice.

[Chetwood]

Man, it's good to have you back!

I don't know of any ISO related problems but having Shrink do 32 KB IFO/BUP padding would be nice and save an additional working step (putting out to files and building padded ISO with Imgburn, that is). The rare occasions I do rip to DVD-9, ImgBurn sets the layer-break for me upon burning so it's maybe not that important to add layer-break support.

Naturally Shrink doesn't handle new protections but that's not what it's designed for so no problem there either.

Of course it would be nice of you to release a new version 3.2.0.15i which would call up ImgBurn instead of DVD Decrypter when burning and maybe even upgrade the DeCSS-free legal versions to abort upon detection of Macrovision protected DVDs so there would be no need to use other hacks.

[blutach]

Nice to see you back dvdshrink and hope you are well!

The new protections are somewhat complex and basically necessitate an independent ripper for these titles. However, most titles (99%+) remain simply CSS protected, and so can be handled with DVD Shrink.

The ISO file output is, as Chetwood said, one of padding (especially in a reauthored DVD). This is not critical. 3rd party tools such as PgcEdit, VobBlanker and ImgBurn take care of this. More reading here.

As for layer breaks, it is not really relevant as DVD Shrink is not a tool primarily designed for DVD 9 to DVD 9 backup. However, if you were concerned, the issue would be of making a LB on an ECC boundary (sector multiple of 16). This can involve padding. Again, 3rd party tools such as PgcEdit and ImgBurn take care of this perfectly.

There is one tiny issue in respect of reauthored DVDs with more than one Vob ID. The number of Vob IDs in the C_ADT table is always set to 1. This is not major either and does not affect playback.

IIRC, if the framerate was incorrectly set (or not set at all) in VMGM_MAT --> FP-PGC, this was not fixed when backing up.

I'm sure there are other small issues, which we will post when we think of them.

Now, members should take note of what dvdshrink said - there are no plans to do a new version. Hence, feature requests are not relevant in this or any other thread. Please do not inundute this thread with what "would be nice to see" etc etc. Simply give dvdshrink the feedback he wants.

Regards

[cynthia_old]

Always nice to see old members return!

Here are some issues collected during the latest years.

When I use author mode to trim the start and end of the movie or chapter, the ac3 timestamp header is not reset to 00:00:00:00 this cause error whe i try to put differents episodes or chapters together in ONE pgc. I have to do it manualy, using HeadAc3he (demux audio-video) and then remux. (is a long process but work 100%)

After shrink a dvd i get all chapter +1
dvdshrink append an empty chapter ! This is in re-author mode.

Just one small issue: after a title is previewed, the audio device isn't released. Most people wouldn't notice, but my card is a semi-pro card that allows you to change its internal sample rate. The problem is that when an app opens the audio device, the card stays fixed at the requested sample rate until it closes it again - so throughout the long backup phase, I can't work at any other rate.

When replacing a titleset with a still image, the audio streams disappear when "Still Image" is selected from the drop down menu, and while the audio is removed, the new VTS_XX_0.IFO and VIDEO_TS.IFO still list the audio streams as present.

When Menu files *_0.VOB of size 0KB exist in the DVD source, they dont get copied over by Shrink 3.2 to the output file. This problem causes a break in the DVD structure and fatal error when burning with Nero.

The workaround is to manually copy the 0KB .VOB files from the source over to the Shrink output.

A little bug I discovered in 3.17 and it seems to happen in 3.2 too. When shrinking a DVD folder, the last file open from the source (usually the last VOB from the folder) is not being closed. The file is released only after closing DVD Shrink.

Number of VOBIDs in VOB incorrect after re-author
I have some titles with the last cell being a blank and a different VOB-ID. So, each one has 2 originally, VOB-ID 1 & 2. To prove to you I had 2 VOB-IDs, I opened up my original files in IfoEdit and looked at the VTS_C_ADT table and it should show 2. But after reauthoring, each title now shows 1 in their VTS_C_ADT table even if it shows VOB-ID 2 at the bottom of this table.

If anyone wants to strip out a VOBID in IfoEdit now, it would show only the first VOB ID only instead of both. Editing the VTS_C_ADT table in the DVD Shrink IFOs as the same as the original would have you seeing both again.

as I just recently noticed that Shrink does remember the heigth of the preview window after a restart. Unfortunately the program window size & position and the width of the preview window is not retained so you have to rejuggle the layout after each start on your screen. Looks like shrink does not remember it was closed with the window maximized and behaves like you explained.

I'll try to look in my old papers to find some more stuff.

[Chetwood]

Originally Posted by blutach

There is one tiny issue in respect of reauthored DVDs with more than one Vob ID. The number of Vob IDs in the C_ADT table is always set to 1.

I guess this is what caused the error message in previous PgcEdit versions:

Warning: the number of VOB IDs stored stored in the VTS_C_ADT table of VTS 3 (1) don't match the real number of VOB IDs in the table (2).
(This is a known bug of DVD Shrink in reauthor mode)


Maybe r0lZ can chime in as there seem to be a few more fixes done in his current DVD Shrink-plugin for PgcEdit. For the sake of completeness I'm posting an old bug list I compiled from forum entries when we were doing the german version of Shrink:

• When I use author mode to trim the start and end of the movie or chapter, the ac3 timestamp header is not reset to 00:00:00:00 this cause error whe i try to put differents episodes or chapters together in ONE pgc. I have to do it manualy, using HeadAc3he demux audio-video) and then remux.
  
  
• Similar to this, when you reauthor episodic discs (or strip episodes from them with Shrink for encoding with AutoGK) that have all eps of a DVD in the first PGC, DVD Shrink does not reset the time code. Thus if you open ep 2 in AutoGK both sound and subtitles start at 43 or so minutes which in in turn results in no subtitles being displayed.
  
  
• I found a small glitch in dvd shrink nothing important really when the file browser window on the left is sized so small that there is need for a vertical scrollbar, and so short that there is no room for the fixed heading row, the program enters an infinite loop between what looks like resizing the fixed row and adding a scrollbar even restarting the program the problem doesnot go away, although it should with a reinstall nitpicking really but thought yall should know.

[setarip_old]

@dvdShrink

Hi!

I'm aware that there is some new kind of copy protection which DVD Shrink cannot handle

Actually, there are several, including:

ARccOS

RipGuard

XProtect (or CrossProtect)

[blutach]

There's one more I remembered overnight and it relates to multiple titles as dialysis1 has hinted at.

If you take a not one sequential title (with many PGCs) and re-author it, it will come out as several titles. This breaks the nav, if for example, it is a game with BOVs, that typically have commands that say Link PGCN x (in the re-authored DVD, each PGC is a title with just one PGC).

Originally Posted by cynthia

After shrink a dvd i get all chapter +1
dvdshrink append an empty chapter ! This is in re-author mode.

I have not seen this.

When I use author mode to trim the start and end of the movie or chapter, the ac3 timestamp header is not reset to 00:00:00:00 this cause error whe i try to put differents episodes or chapters together in ONE pgc. I have to do it manualy, using HeadAc3he (demux audio-video) and then remux. (is a long process but work 100%)

This is an important restriction in many transcoders that only a full demux/remux fixes. It sometime has bad playback consequences. Good find!

Regards

[dvdshrink]

Thanks everyone for a really positive response :-)

Lots of bugs! This feedback is just perfect. I have a little to say about some of the issues, and the rest, I've definitely taken notice. Not releasing file handles (or audio devices) is just plain bad programming, sorry about that.

If you take a not one sequential title (with many PGCs) and re-author it, it will come out as several titles

Yes, DVD Shrink is limited to only re-authoring a single PGC at a time. It cannot handle multiple PGC titles, or even multiple angles in single PGC titles. Definitely this could have been done better.

When I use author mode to trim the start and end of the movie or chapter, the ac3 timestamp header is not reset to 00:00:00:00

It was left like this on the assumption that DVD players don't really care - since they all support jumping into the middle of a stream anyway, with Link#PTTN command. The problem is that you guys all want to do more advanced stuff with the output, and not just play the stream!

blutach is correct when he says this requires a full demux/remux, but I was thinking about this, and you know, I'm not even convinced it is possible. DVD Shrink can start a re-authored title at any arbitrary VOBU index, but audio frames are not typically aligned to VOBU boundaries. So you need to discard any partial audio frames, and start the first complete frame with a 00:00:00 timestamp, but what if you have multiple streams with different bitrates/formats? E.g. AC3 and LPCM. It may be impossible to find a "first frame" of each which is exactly aligned in time, so to maintain sync with each other, surely at least one stream would have to be started with a non-zero timestamp? Also I was wondering if the video timestamp for the first I-picture in the vobu could end up being negative after adjustment, or something bad like this...

There is one tiny issue in respect of reauthored DVDs with more than one Vob ID. The number of Vob IDs in the C_ADT table is always set to 1.

I was looking at exactly that piece of code the other day, and thinking, this can't be right!

After shrink a dvd i get all chapter +1
dvdshrink append an empty chapter ! This is in re-author mode.

An earlier version of shrink may have tried this as a feature. I don't remember exactly, but I recall experimenting with this. It means you can chapter skip to the end of the title.

The ISO file output is, as Chetwood said, one of padding

Just to make sure I fully understand this: the IFO and BUP files shouldn't actually be changed in size or padded as such, but should be arranged in the ISO so that there is at least a 16 sector gap between them? (And the sector offsets in each IFO adjusted accordingly).

ARccOS
RipGuard
XProtect (or CrossProtect)

Great, thanks for the feedback. I'm not sure which is which but I gained some understanding now of how these work.

[blutach]

Originally Posted by dvdshrink

Just to make sure I fully understand this: the IFO and BUP files shouldn't actually be changed in size or padded as such, but should be arranged in the ISO so that there is at least a 16 sector gap between them? (And the sector offsets in each IFO adjusted accordingly).

Yes - simply adjust the pointers in the VMG_PTT and the VTSI_MAT tables. For an example of how this works, take any reauthored DVD, and save it in PgcEdit with the option to save with 32k gaps on. Examine the differences between the pointers in the various titlesets compared to the DVD Shrink output.

In respect of the timestamps on "cut DVDs", it is not just the AC3 headers but also the cell elapsed time of the DSI stream matching that of the PCI stream (see 0x0045 and 0x0423 of the navpacks). In particular, if a flag is incorrect and the STC clock is not reset, very unpredicable playback can occur.

In respect of adding a tiny cell as an end chapter, this is not necessary anyway for 2 reasons:

1. Subject to 2 below, by making the Next PGCN = the current PGCN, you can skip to the post commands
2. Even if there is a tiny last cell, if PUOp=128 is set (Next PG) in either or both of the IFO or VOB for that cell, you can't skip. You need to eliminate this PUO.

I had 2 other issues pop to mind as well. The first has to do with the jump/link/call flags - this is very minor and most technicians agree it is entirely useless and ignored by almost all players. Please see this thread, with the definitive answer by (as you would expect) mpucoder in post 17.

The 2nd has to do with eliminating PUOs on a DVD with angles. It is natural to want control of your DVD, however, if the DVD has angles, on many players, if the angle change PUO is unset (in either the IFOs and most particularly, the VOBs) in such a DVD, a little angle icon comes on the screen, which is very annoying. On many players, this can not be surpressed. So, advising users not to eliminate POps where a DVD has angles is required, IMHO.

Hope this helps.

Regards

[dvdshrink]

I never thought of that. You certainly know your stuff. Maybe better to always leave the angle PUOP unchanged (uops &= 0x400000)? I can't imagine it's ever an inconvenience to have it set.

[blutach]

I agree re the PUOps. And coming from you, that's a very big compliment. Thank you.

Regards

[cynthia_old]

There seems also to be an issue with the automatic burning if Nero 8 is used. By reading this thread it seems that the link to call it up is broken by a name change of the .exe in Nero 8.

http://forums.afterdawn.com/thread_view.cfm/579890

[dvdshrink]

Well, the consensus at afterdawn seems to be that Nero should fix it, and I agree ;-)

I have another question - deep analysis. Does everybody use it? Or does everybody hate it?

The reason for deep analysis, of course, is that DVD Shrink can't know in advance exactly what is in the VOBS, until it's checked every sector. And it needs to know what's there, to optimize encoding decisions. End credits are a classic example - they are really hard to compress, and they come right at the end of the movie, so there's no margin for error remaining when you're trying to hit a specific size.

What I'm trying to work out is if people just accept that two-pass encoding is an unpleasant fact of life, or if you only enable it when really necessary, e.g. if Shrink failed to hit the target size?

[Calimari]

I sure love the deep analysis! DVD Shrink knocked down its competitors because of its great quality video output. Don't forget that people have fast machines these days (or at least they should ) So that extra few minutes won't matter. Besides, the user has the choice, use it or not. IMO it was one of the aspects that made Shrink so great.

[setarip_old]

What I'm trying to work out is if people just accept that two-pass encoding is an unpleasant fact of life, or if you only enable it when really necessary, e.g. if Shrink failed to hit the target size?

Based on postings I've read over the years, it seems to be (as we said in the '60s) a "mixed bag" - There are those that swear by it and use it every time they use DVD Shrink. On the other hand, there are those who use it only when the initial indication by DVD Shrink is that the total filesize cannot be reduced sufficiently. And, of course, there are those who use it selectively - e.g. not bothering to use it for CGI DVDs, but using it for DVDs that have high action and/or intense color.

Personally, I believe it to be a valuable available tool...

[Rogi]

Originally Posted by dvdshrink

I have another question - deep analysis. Does everybody use it? Or does everybody hate it?

Personally, I always use deep analysis.

In my experience, is clearly visible that dvd backup with deep analysis is better than without it.

[Vulcan94]

Originally Posted by dvdshrink

Well, the consensus at afterdawn seems to be that Nero should fix it, and I agree ;-)

I have another question - deep analysis. Does everybody use it? Or does everybody hate it?

<snip>

What I'm trying to work out is if people just accept that two-pass encoding is an unpleasant fact of life, or if you only enable it when really necessary, e.g. if Shrink failed to hit the target size?

I always use deep analysis. I want the best quality I can get, so I just take a break while DVDShrink does it thing. Great program by the way.


Vulcan94

[Chetwood]

Originally Posted by dvdshrink

I have another question - deep analysis. Does everybody use it? Or does everybody hate it?

It rocks and I use it all the time! Since it's an option which is also off by default on new installs I don't see why anyone should have a problem with it. Like already mentioned here, computers are a lot faster now so maybe the efficient use of multicore cpus (where possibe) is the way to go on your new project.

[dvdshrink]

Shrink is already multi-core - it'll create as many encoding threads as you have cores. I thought the bottleneck has always been drive speed, or maybe I just need a new drive? It takes over 20 minutes to make one pass.