Dumping DVDs with bad sectors

[NeonMan]

Hi!

Some time ago Darksoul thought about using ddrescue to copy dvd with damaged sectors, the aproach seems effective but quite slow.
Yesterday I've tried something simmilar but usind dvdisaster, a CD/DVD backup, restoration and protection (against damages, not copy) program, for both windows and linux, which uses a special reading mode for 'defective media' called adaptive reading.

The program works fine and pretty fast with CDs with the damaged ring and it shoul work with almost any bad sector crap. I would rent a newline DVD to check if it does read xprotect.

Some screenshots, inner red blocks are the damaged ring proctection, outer red block are actual scratches on the disk's surface.
http://www.imagechile.net/imgt/img5_1219754941y.png (http://www.imagechile.net/vista.php?id=img5_1219754941y.png) http://www.imagechile.net/imgt/img5_1219754958v.png (http://www.imagechile.net/vista.php?id=img5_1219754958v.png)
# Readable sectors
# Unreadable/Skipped sectors
# Now reading
# Not read
---------------------------
dvdisaster
http://dvdisaster.net/

<Updated>
Patch available for windows in order to allow dvdisaster v 0.70 reading encrypted and/or pressed DVDs (See post #11 (http://forum.doom9.org/showpost.php?p=1177276&postcount=11))

[RunningSkittle]

"This software does not support DVD-ROM type media"


Not usable types (image can not be extracted):

DVD-ROM (pressed DVDs), CD-Audio and CD-Video.

[NeonMan]

It cannot generate Error Correction Codes (ECC files) for pressed DVDs but it can read them. I've tried with Queen Live at Wembley Stadium and Aliens, Only CSS both of them, but read flawlessly and the generated iso worked perfectly.

I've not tried Cd-Audio and XA-mode CDs
I'm using version 0.71~devel23-6, latest on Ubuntu, I will try the stable version

<edit>
Windows build can't read DVD-Rom Disks (Don't ask me why) but on linux can (Asking myself why)

[RunningSkittle]

using the windows binaries it cannot read them at all.

[NeonMan]

I've spotted the problem after running the backup of the queen's dvd for the second time.
After booting linux I started dvdisaster and on the first sectors an error shown on the status bar Reading scrambled sector without authentication, this error didn't show the first time I've done the test so I guessed that the autoplay feature of totem, which plays the disk if it is inserted AFTER booting, authenticated the drive AND allowed dvdisaster full acces to the disk. To check it out I ran Totem and the paused the playing, after that dvdisaster could make an iso image of the disk. Note that even after closing Totem, dvdisaster can still read the whole disk.

http://www.imagechile.net/imgt/img5_1219763794b.jpg (http://www.imagechile.net/vista.php?id=img5_1219763794b.jpg)
Keeping totem paused, the darker green blocks were previously read without authentication, the ongoing failed with the previously mentioned error the first time.

http://www.imagechile.net/imgt/img8_1219763821g.png (http://www.imagechile.net/vista.php?id=img8_1219763821g.png)
Finished, totem is closed after the 3rd Gib

The interface on the screenshots is different because here I'm using the linear (dd-like) reading mode, but adaptive mode also works with the totem-trick

dvdisaster on linux stars reading the disk but, if totem hadn't run before, it starts to fail (skippink) after a bunch of sectors have been read.

<edit>
Windows version of dvdisaster keeps on rejecting DVD-ROM disks, even after authenticating the drive (Windows media player, VLC) and also rejects the CSS-disabled drive from AnyDVD. I supose it is a hardcoded limitation on the windows version so this solution is at present Linux-Only

[setarip_old]

@Neon Man
Hi!tried with Queen Live at Wembley Stadium and Aliens, Only CSS both of themso this solution is at present Linux-OnlyPerhaps I'm not understanding something in your posts but surely, circumventing CSS, which was accomplished years ago, is not something in need of a new "solution", is it?

[NeonMan]

@Neon Man
Hi!Perhaps I'm not understanding something in your posts but surely, circumventing CSS, which was accomplished years ago, is not something in need of a new "solution", is it?

CSS is not the target of this program, but rather the bad-sectors stuff, I will have a newline DVD (Surely with al the anticopy crap) to try if the bad sectors and the bad filesystem structures can be skipped better with the adaptive mode of dvdisaster. Once with an iso image of the DVD on the hard disk CSS is still present but it doesn't matter because, as you said, there are LOTS of tools which decrypt CSS under linux and/or windows. (ie. Vobcopy)

the authentication 'problem' was an unexpected error, simply by running and closing Totem/vlc/xine/anything-that-reads-css-DVD, dvdisaster stops complaining at the DVD-Rom disks so it could make the image of the disk.

I hope this post explains the actual purpose of the thread.

[setarip_old]

I will have a newline DVD (Surely with al the anticopy crap)If memory serves correctly, a few of the most recent "New Line" releases did NOT contain the "XProtect" copy protection - so when you select such a DVD for your work, I'd suggest you make certain that the directory of the DVD contains duplicate .VOBs and absurdly sized and mismatched .IFOs and .BUPs, as these are the sure signs of such protection (You might want to "Search" the "Decrypting" and "Newbies" sub-forums here for specific "New Line" titles containing "XProtect" copy protection).

I've tried something simmilar but usind dvdisaster, a CD/DVD backup, restoration and protection (against damages, not copy) program, for both windows and linux, which uses a special reading mode for 'defective media' called adaptive reading.Just as a general observation, it's hard to believe that an already-existing generalized disc repair program ("dvdisaster") would be capable of dealing with the various forms of DVD copy protection, in light of the fact that there is required constant ongoing effort of authors/publishers to update and modify programs written specifically to overcome DVD copy protection...

[NeonMan]

Reasonable, I expect I will get from dumping xprotect is all the data on the DVD but the filesystem completly screwed.
The bad sectors can be read and padded but I don't think I'll be lucky enough to get a clean filesystem after the dump.

[setarip_old]

By all means, let us know how your project evolves ;>}

[NeonMan]

I've started investigating the protections and making some advances. I can now dump the entire mpeg stream from an image with corrupted filesystem (from dvdisaster) and now I'm trying to order the tings a little bit.

By the way, the official dvdisaster is coded to reject dvd-rom and encrypted dvds in order to stay away from laysuits, however, the gpl allows the modification of the program so most distribution packed versions have it disabled.

For the windows version I've made two patches for both command line and GUI version which eliminates the DVD-Rom restriction.

Command-line
http://www.megaupload.com/?d=241SLIBF

GUI
http://www.megaupload.com/?d=89T9LWWU

I don't really know if the gpl allows the modification in binary form of the program, that's what the patches basicaly do, if someone knows I would like to hear from him it.

[Oopho2ei]

I don't really know if the gpl allows the modification in binary form of the program, that's what the patches basicaly do, if someone knows I would like to hear from him it.
afaik you have to make the source code (in this case probably the patch you applied to the source) available to the public. You may not distribute the modified program (which is under gpl) as binary only. That's what i remember from Richard Stallman's speech...

[setarip_old]

Sounds like you're making some headway ;>}

Please clarify - Is your ultimate desire with this project to generate a complete functioning duplicate DVD (menus, chapters, movie, alternate audio languages, subtitles, etc.), or an MPEG2 of "movie only"?

[NeonMan]

@Oopho2ei
I'll post the patches to the source, tested on linux, but since I don't have the development tools for windows, the posted binary patches were actualy made over the original exes wit olly-dbg. much like the way warez cracks are made. (Quite simple, a couple of nops & jmp)

scsi_layer.c
Looking for the diff


fingerpront.md5
Looking for the diff


@setarip_old
The initial thought is a GPLed program able to clean the filesystem of the dvds, allowing to play the copy flawlessly on software players. CSS decryption might be supported once the primary target is met.
At presen I can extract CSS scrambled VOB streams (not files yet) from DVDs, so a VOB may contain (and it does) parts from the previous and next VOB.

[Oopho2ei]

@Oopho2ei
I'll post the patches to the source, tested on linux, but since I don't have the development tools for windows, the posted binary patches were actualy made over the original exes wit olly-dbg. much like the way warez cracks are made. (Quite simple, a couple of nops & jmp)
What people do here is legally questionable anyway. It's just because you asked. Hmmm, have you some experience debugging windows applications? If yes you could help us getting new processing keys or reverse engineering BD+. We need smart people :)

[NeonMan]

The things we make here are legally questionable, that's true, but I prefer to avoid doing anything ilegal if it is perfectly avoidable, using gpled code as a base for a gpled project.

I don't have a BluRay (nor an HDDVD) drive or hardware player to play with it so I think I won't be of much use until I have one. As soon as I have one I would like se the insights of the unbreakable for 10 years (http://forum.doom9.org/showthread.php?t=127827) protection

The promised patches :P, made against dvdisaster-0.70.6
scsi-layer.c (It basicaly comments 3 lines of code)

347c347
< //dh->subType = UNSUPPORTED;
---
> dh->subType = UNSUPPORTED;
1185,1186c1185,1186
< { //CloseDevice(dh);
< //Stop(_("This software does not support encrypted media.\n"));
---
> { CloseDevice(dh);
> Stop(_("This software does not support encrypted media.\n"));


fingerprint.md5 (Adjusting the md5 of the previous file)

48c48
< d838c1b55a100a5f59b190c715a89ebd *scsi-layer.c
---
> e68152e2ffdd4b12f8f5be7800fc9e9f *scsi-layer.c


pd. I couldn't edit my last post, so the patches are here instead

[setarip_old]

...to clean the filesystem of the dvds, allowing to play the copy flawlessly on software players. CSS decryption might be supported once the primary target is met.Unless I'm misunderstanding you, I'd suggest that it would be imperative to decrypt the CSS because, from what I've been told by authors of ripping programs, while accessible by a DVD-ROM and/or burner for READING, the area where the CSS key codes are stored on a DVD is not accessible for WRITING. Therefore, any disc you burn (without decrypting the CSS) would fail due to the lack of keycodes necessary for descrambling.

If I'm misunderstanding you, please clarify...

[NeonMan]

Tha primary target is fixing the filesystem, with the fixed copy of the stll-scrambled you can either decrypt it (with another program at present), to be able to use it on Hardware players or play it with a software player like vlc. Software players like vlc wont complain at scrambled copies of a dvd since it could guess the key even without the DVD.
Maybe you are right that if I burn a dvd still scrambled (I allways descramble the DVD before copying), a hardware player would reject the disk but, I thought CSS was a play protection rather than a copy protection, vulnerable to bit-by-bit copy. Correct me if I'm wrong.

<edit>
Excuse me if I didn't explain myself correctly, english is not my nature language by far.
CSS decription will be added, but it's not a priority because it is a long-dead protection with lots of tool an libraries to descramble it.

<edit2>
http://en.wikipedia.org/wiki/Content_Scramble_System
CSS can only be added to replicated DVDs. A duplicated DVD, i.e. DVD burned from recordable media such as DVD-R or DVD+R cannot have CSS. The lead-in area of these media does not have the proper track to store CSS information.

[setarip_old]

I thought CSS was a play protection rather than a copy protection, vulnerable to bit-by-bit copy. Correct me if I'm wrong.It's just a matter of semantics, I guess. CSS is copy protection in that the lack of a descrambling keycode is intended to prevent making a viable copy for playback...

[Ajax_Undone]

Wow its like a tricked out isopuzzle

[mc2man]

While using dvdisaster on x-protect is doable, the time frame is still beyond useful. Using 'hairspray' as an example took around 3.5 hrs.

There was a group of 30,000+, followed by 70,000+, and at the end of the disk 500,000+ unreadable sectors. The first 2 groups were skipped in blocks of 16 sectors @ 2 sec.'s per block, the final group went quickly, mostly as one bulk error. (wasn't paying close attention.
Also an issue in these methods is many drives after several read errors drop to a low rip speed for readable sectors (2x or so

In this case the orig. disk structure appears very normal, not so in the .iso. Basically all the .ifo's and .vob's have switched extensions.

What's somewhat interesting though is as long as the dump is kept in .iso format it plays perfectly with open source players, with full navigation, ect.

A typical ARccOS, while easily fixed, would probably be in the 1-1.5 hr. range min.
The various ripguards generally have very few bad sectors, maybe 5-10 min. extra at most (if using drive that can regain rip speed
Myself I'd just use a read try modified vobcopy on ripguards (also using blocks of 16 sectors

http://img79.imageshack.us/img79/236/screenshotdvdisaster071ja2.th.png (http://img79.imageshack.us/my.php?image=screenshotdvdisaster071ja2.png)


http://img224.imageshack.us/img224/5885/screenshotmediumisotu9.th.png (http://img224.imageshack.us/my.php?image=screenshotmediumisotu9.png)

[NeonMan]

Thanks for the feedback mc2man, the bad sectors are read out of the box.

What's somewhat interesting though is as long as the dump is kept in .iso format it plays perfectly with open source players, with full navigation, ect.
I've been studing both the Iso9660 and UDF table of contents and the issue you explain might be related with it. I supose licensed dvd players read the UDF structure only and the iso9660 is kept for compatibility issues.
I won't have much time until the end of september so the current research is slooooooow but any feedback is interesting.

<edit>
Did you set up dvdisaster to fill the unreadable sectors? (take a look in the image tab on the setup menu)

[jinjin_jp]

I tried to use dvdisaster (with patch).
But it can't read nor scan with below message.
http://img502.imageshack.us/img502/7798/15039875dn7.th.png (http://img502.imageshack.us/my.php?image=15039875dn7.png)
It is same about DVD-ROM and DVD-R.
My method to use has anything wrong?

Regards.

[NeonMan]

Tell me what hace you done (step by step).
I guess you are using any of the recovery options so you may use the read option, not any of the others so it wont need the ECC file (which you can leave blank).

[jinjin_jp]

Thanks for the reply.

I insert media(DVD-ROM or DVD-R) to DVD drive, next click "Read" or "Scan" button.
I set read option like below following to http://dvdisaster.net/legacy/en/example82.html.
http://img169.imageshack.us/img169/7813/68395058vi7.th.png (http://img169.imageshack.us/my.php?image=68395058vi7.png)

Regards.

[NeonMan]

@mc2man
It would be possible for you to give me the first 2-3 MiB of the DVD?
Something like...dd if=/dev/scd0 bs=2048 count=1536 | bzip2 -9cf > image.iso.bz2
where /dev/scd0 is your cdrom driveIn this case it compress thne image on the fly.

@jinjin_jp
Try the same wit a non-pressed dvd, just to know if it is either patch's fault or a program's flaw.
I'll try the patched program under windows again.

<edit>
Try linear reading and when asked skip RS02 checking. Another detail, it doesn't matter but it will be better to let the drive spin up a few seconds, about 3-5 seconds
I think it will be better to recompile the program under windows ;)

<edit2>
I insert media(DVD-ROM or DVD-R) to DVD drive, next click "Read" or "Scan" button.
Strange, even with DVD-R media?

[mc2man]

Did notice that using a better drive times decreased by about 35%.
In the case of ripguard and ARccOS, considering how trivial they are to repair, could be some value there.

the issue you explain might be related with it

I didn't view it as an issue, rather a 'positive' due to fact the title played back perfectly with the s.p. intact. In some regards it's interesting in terms of DMCA because nothing has been circumvented - the iso is still encrypted, and the s.p. is still there. The only 'issue' in that regard would be use of unlicensed player for playback

I supose licensed dvd players read the UDF..... I think possibly the only issue with these players in regards to the iso is they won't 'brute force' the keys.

While I don't think it affects dvdisaster for the x-protect types you may want to (at least on debian systems) install either the latest libdvdnav (4.xxx) or install a patched one (libdvdnav4-ifo)
For mplayer (atm) patch dvdread before build

as for request I'll pm you later

[jinjin_jp]

@NeonMan

I re-installed, so it becomes not to have error.
:thanks:

But It can't read CSS ? All is red.
http://img526.imageshack.us/img526/4914/75549228ws7.th.png (http://img526.imageshack.us/my.php?image=75549228ws7.png)

Regards.

[NeonMan]

Did notice that using a better drive times decreased by about 35%.
Do you mean that better drives have slower reading speeds, or that a better drive skips errors faster?
On my laptop a normal data dvd can be read up to 12x but a movie DVD barely reaches 4x.
The name swapping is not a bad issue but interesting enough to study.
What do you mean by s.p.?

One more thing. My network admin has blocked megaupload and rapidshare to save bandwith so if you could upload files anywhere else it will save some extra work

<edit> @jinjin_jp
You have just replied while I was doing so.
The reading scrambled sector without autentication error can be fixed (at least on linux) by running the software dvd player before. I'm thinking in adding the css authentication and decryption into dvdisaster codebase to avoid this errors.

[mc2man]

Do you mean that better drives have slower reading speeds, or that a better drive skips errors faster?

Faster error skipping - down to blocks of 16 sectors @ 1.2 secs average or so
Resume to almost full normal rip speed after extended error skipping
(my case 12x+ @ layer break area

The reading scrambled sector without autentication error can be fixed (at least on linux) by running the software dvd player before
I didn't need to run player first but if so and trying an x-protect title you'll need the proper libdvdnav or your player will crash on opening

[jinjin_jp]

@NeonMan

It can scan and read by running the software dvd player before.
:thanks:, again.

Here is the result of RipGuard DVD.
It has 3 BadSectors in LBA: 77392 - 77749, 77920 - 81085, and 2979968 - 2983138.
The figure seems to be match with it.
http://img411.imageshack.us/img411/2866/3scanje6.th.png (http://img411.imageshack.us/my.php?image=3scanje6.png) http://img179.imageshack.us/img179/8330/3readeq8.th.png (http://img179.imageshack.us/my.php?image=3readeq8.png)
("Read" seems to be continue endless. It must be stopped manually.)

Regards.

[setarip_old]

@NeonMan

Anything new on this project?

[NeonMan]

The CSS auth is resolved but skipping the the bad sectors and pointing towards the correct ones is cumbersome, the only code recycled from dvdisaster is the reading (ioctl) layer.

I'm thinking in link against libdvdread to see if I can get somoe more info from the disk.

Also it would be of great help if someone knows of a gpled IFO reader or at least a command-line linux IFO reader.