Update: IsoPuzzle 1.6 - Recovery program for scratched damaged DVD Data-CD

[marsoupilami]

You need a working ASPI layer on Your system!
This program has been tested with the newest version of WNASPI32.DLL of NERO AG (Version 2.0.1.74).
Please visit the homepage of NERO AG to get the DLL and read the terms of use!
http://www.nero.com/nero6/eng/WNASPI32.DLL.html

NEW VERSION AVAILABLE HERE
http://www.geocities.com/marsoupilamis/

Please put the WNASPI32.dll together to Your IsoPuzzle.exe folder - do not copy it to Your system folders. Please note, that the behaviour of this program may depend on this dll.

@ admins:
========
If You like to host this program here please feel free to do it!

Short Description:
=============
The goal of this program is to recover as much as possible data of damaged, bad burned or scratched CDs or DVDs.
This version is experimental and supports only data CDs and DVDs with sectorsize of 2048 byte - Audio CD, VCD and SVCD are not supported.

How it works:
==========
All readable data is stored in an "ISO" image file which can be read by various software such as "Daemon Tools", "WinRar" (needs iso-plugin) and most burner software like "Nero" etc.
Please note, that image files greater 4GB need to reside on a NTFS disk!

Every succsessfully read sector is stored in the image file (*.ISO) and will be marked "good" in the corresponding "flag" file (*.FLG).
If a read error is encountered a dummy sector will be written which will remain marked "bad". In this case a random number of sectors are skipped for faster progress.
Reaching the end of the disk a retry loop will be entered until all sectors have been read or a selectable "Give Up Time" has been reached.

After selecting the "Source" drive and the "Destination" folder the copy process can be started by pressing "Start".
The filenames are taken automatically from the DVDs volume label if available.

The copy process can be interrupted at any time by pressing the "Pause" button and can be continued by pressing "Start" again.
This way You can try to complete Your image file even on different machines with different CD drives - often one drive can read some sectors which are unreadable for an other one.

For using different drives You can share the image file *.ISO (together with the flag file *.FLG) on a network drive or put it on a portable disk (USB).

Import function:
============
To avoid dealing with very large files You can use the "Import" function:
Assuming that You started on machine "A" and no more data comes from Your DVD.
.) Copy ONLY (!) the *.FLG file (e.g. via USB stick) to machine "B".
.) On "B" select "Destination" folder, where the *.FLG is located.
.) "Start" will create an "import file" *.IMP in the same folder. All readable sectors are now stored here.
.) After completion on "B" transfer the *.IMP to a location where it is accessible from machine "A".
.) On "A" it can be merged into the existing *.ISO image with "Import File"

Force Disksize - WARNING - Use with care!
=========================================
.) Some DVD drives report slightly different size of the same DVD/CD.
For ripping to an existing *.Iso/*Flg pair (or *.Imp/*.Flg) You can set
the disk size to "Flagfile reported".
(You can try "Estimated" to get out, which drive is wrong ...)
.) In some cases "empty disk" is reported although the disk is readable.
Setting the size to "Estimated" would be the choice here.
This function may be slow and unreliable.

Be careful and be sure You know what You are doing when using this function!!!
By changing the (drive reported) disk size You might accidentially destroy
an existing image file - make sure the given filenames are correct!

Cylce Tray
==========
It your responsibility NOT to change the disk while the tray is open!

History:
========
1.4 Added options: Cycle tray on multiple errors, Select iso-filename,
Force disksize (solves "different sectornumber" and "empty disk")
1.3 Progressbar & other cosmetics
1.2 Crash on missing host-adapter-0 fixed
1.1 A few bugs corrected
1.0 First release

[kumi]

Interesting! I've been recently hunting for something like this.

If I remember correctly, DVD-Video discs use 2048-byte sectors, so your tool should work on these, right? I have a valuable home movie burned on a degraded DVD+R (read errors occur near the starting sectors.) I'm hoping that if the outer sectors are readable, your tool will help me recover some of the footage.

Thanks marsoupilami

[marsoupilami]

Yes! VideoDVD use 2048-byte sectors - that's right!

Until now I'm not sure what's the best way if You are not able to complete without read errors (missing sectors).
In case of a Video-DVD I guess You can try "mock stripping" with IfoEdit.

[Mug Funky]

i don't suppose this handles CSS'd discs then? (2054 byte sectors IIRC). there wouldn't be much point ripping these without decryption anyway :)

real good work - the number of bad discs i've had to try on all different drives is quite large. this program should help a lot :)

[kumi]

I couldn't get IsoPuzzle to read out any sectors of my disc, I let it run for about 30 minutes and I got 100% read failure rate. Both DVDDecrypter (in ASPI or SPTI mode) and IsoBuster can read all but about 120 sectors.

[marsoupilami]

@Mug Funky
If You will get off the image You can try mounting it with DaemonTools as a virtual disk and there running DVDDecrypter... but I didn't test this.
@kumi
Maybe You have a RPC protected disk?
I've heard in such case the drive could be unlocked by opening the disk with a software player or DVDDecrypter ;)

[kumi]

Maybe You have a RPC protected disk?Not RPC. I replaced my wnaspi32.dll (I think it came with windows XP?), and IsoPuzzle started working.

Non-working wnaspi32.dll (100% fail-rate)
Description: ASPI for Win32 DLL
Company: Adaptec
Version: 4.71.0002.0000

Working wnaspi32.dll:
Description: Win32 Nero Aspi Library
Company: Ahead Software AG
Version: 2.00.0001.0074

Edit: it's strange, the Adaptec wnaspi32.dll is the same one DVDDecrypter is using, and it works in that program.

[LakersFan]

So basically it's like this: http://forum.doom9.org/showthread.php?t=106778 except in English?

I love that program, it's saved me a few times. :)

[marsoupilami]

@kumi
It is NOT necessary to change anything in Your system:
Simply keep the working dll together with the exe in the same folder!
I didn't test many versions of different dlls... If You like mail me the non-working one - I will try to fix this in a next version!

@LakersFan
I had the idea to write such a program a long while ago.
When a friend told me about that german program I presented it here.
This one - I hope - is more user friendly because of it's GUI, it's 'GiveUp' and it's 'shutdown' function and it takes the filenames from the disk automatically (which could be a useful feature late at night) :)
I also tried to find a faster an more efficient retry strategy with pausing or spinning up the drive after some read errors.

For the future I have planned to support VCD and maybe audio - but we'll see....

[LakersFan]

Yes, I agree this is MUCH more user friendly for me, specifically the GUI. Thanks for taking the time to write it. I'm looking forward to trying it out. (Then again, hopefully I won't have to any time soon.)

Take care,
LF

[marsoupilami]

A few bugfixes ...
NEW VERSION - ALSO AVAILABLE HERE
http://www.geocities.com/marsoupilamis/

Please visit the homepage of NERO AG to get the WNASPI.DLL and read the terms of use
http://www.nero.com/nero6/eng/WNASPI32.DLL.html
:)

[mgpc]

Thank you!!! :-)

It's nice, I like it, I need it.

Thanks again.

Mark

I ran it last night over an 'unreadable' DVD and it's pulled off all but 200 sectors so far (170 passes) and it's still getting data off, great!

A couple of minor things, it would be good (for me) if the scroller in the log window would stay at the bottom, ie always show the last messages.

My drive seems to cope better, and read more data, if I cycle the tray occasionally, i.e it just picked up another ~20 sectors after I cycled the tray, after >20 passes with getting no data. Could you consider an option to cycle the tray every so many passes, (or maybe after 'n' passes with 0 sectors read)?

Thanks

Mark

[Trahald]

this program is better than that other one (which i tried a few months ago.) the auto loop and import are great! (my wifi isnt stable enough for the network option)

[LakersFan]

Really really minor bug, but the 'X" button in the upper right corner doesn't work for me...

Other than that, :thanks:

[cnn66]

Interesting! I've been recently hunting for something like this.
LOL

[marsoupilami]

@LakersFan
Yes - I disabled the small "x" button but did'nt grey it out...

@mgpc
The main reason for a new version was an crash on initialisation with "host adapter 0 not present"...

The "cycle tray" option I had considered when I started to write this program - but I encountered a lot of problems especially when a disk makes troubles while mounting.
When I played around with some "bad burned" disks I experienced that only cycling the tray doesn't have a great effect - don't ask me why.
A really good practice was to cool down the disk putting it into the freezer for some minutes and to reload it manually. Maybe there are microscopic air bubbles (?) inside the medium which are decreasing in size temporarily at low temperature.
However - I'll think on it when i continue optimizing the retry strategy...

[marsoupilami]

I've just put the final 1.3 on my page.

* The small "x" button is working now.
* The logwindow is scrolling down
* A progressbar is added - not very precise but it's obvious now what's going on ...

@mgpc
With the "cycle tray" option I didn't succseed - A further problem is how to "re-unlock" the drive if You like to backup a RCP protected disk... (The first unlocking would be easy with every player)
I leghtened the pause times instead - for me it seems to be more effective.

[Mug Funky]

i just tried it on a poorly burnt Z-CLV disc (ie the 3rd speed zone and beyond would not read at all on my 111D), and it worked nicely. excellent.

so could this be extended to CSS'd discs? i'm still not clear if it can rip the full 2054 sectors, and if so whether it's possible to deCSS the image file made (i'm guessing probably not unless the DVD is still in the drive). not a big deal because most of the bad discs i get are burns, but every now and then you get a badly pressed disc, or one that's been buggered in transit or by careless users.

maybe i'll just try one and see what happens :) i wouldn't expect you to implement deCSS yourself, as that could put a legal compromise on a program that's too useful for that kind of bother.

[cd090580]

Hi,
juste one question. How many pass it does on bad sectors ?????

20000 bad sectors = 20000 passes ???? I hope not

[marsoupilami]

... i wouldn't expect you to implement deCSS yourself, as ...

I didn't spent much time on thinking of RCP, deCSS, decryption and so on - I have no idea how that works and I guess there could be a simple way:

An RCP disk can be "opened" with every player software or by selecting the drive with DVDD.
The resulting image will be still encrypted - but with Daemon Tools You can mount it and DVDD might decrypt the virtual drive?
Many thanks to LighningUK for that great program:thanks:

Maybe it's also possible to load AnyDVD with "ignore read errors" turned off - but I never tried all this...

[mgpc]

Hi,
juste one question. How many pass it does on bad sectors ?????

20000 bad sectors = 20000 passes ???? I hope not

Click the '+' next to 'normal' in the 'Log Window' section, and change the Normal to Verbose, and the log window will show how many reads and failures it makes on each pass.

But yeah, 20000 bad sectors will take a while. Mind you I had a disk than ran to ~100 passes with 2000 bad sectors and not many reads, then it seemed to finish quickly from that point.

Mark

[bloodrush]

Wow!! that's all I can say. Greatest program ever. This program has recovered a DVD that I've kept in my closet, praying I could someday recover the data off of it. Thank you so much, worked like a charm. It took 10 hours but it fully recovered data that no other program has gotten close to.
Also, paired with AnyDVD..you can get data off of CSS encrypted discs. I had a retail dvd that always had glitches at the end of the movie due to scratches. I was never able to get DVDdecrypter to fully read it. Loaded up AnyDVD and ISOPuzzle, it was able to pull off those last few bytes to give me a complete ISO finally. Burned off a new copy and....yeah...complete movie. Thanks for the great program.

[Mug Funky]

just a small feature request - would it be possible to specify a filename for the ISO file that gets ripped?

[marsoupilami]

@bloodrush
Nice to get some positive feedback - Anjong Haseo!

@Mug Funky
Currently I don't have much free time... For one of the next versions I've planned to support VCD which would require some changes (e.g. a differerent fileformat - mds, cue/bin? I have no idea right now...)
But I'll think on it - as I'll do with a "cycle tray" option.

[outsiders71]

marsoupilami,

I just wanted to applaud your work! This program is an essential tool! Thank you!

[sirzur]

Thanks for this great tool. I have a question about Sectornumber. I have a writeable DVD-R which I am trying to rip. I have used one drive to rip and then tried it in another. However I am getting "Sectornumber sould be XXX. Drive reports XXX"

I presume that the Sectornumber is to make sure that information on one disk is not accidently overwritten onto another. Even if the sectonumbers are different, would it still be possible to use the import function or use another drive to rip the problem disk? If it is possible, can the program be modified such that it would continue even if there ae different sectornumbers. Of course, if that is the case, then there should be warnings about the effect of overwriting one disk on another etc etc.

PS I have also been following advise on freezing the diska nd then reripping
Cheers

[marsoupilami]

sirzur,

You are right! The sectornumber is checked to prevent ripping different disks to the same image and also to determine the size of the flagfile.
Overwriting the drive reported sectornumber will be possible in the next version for a different reason: some erroneous disks report size 0 (I guess because the session is unclosed).
If you don't like to wait a few weeks for the next version there is a - not really recommended - workaround (please don't beat me, if it's not working):
Make the 1st ripping attempt on that drive, which reports the larger size. You can edit the "total number of sectors" which is the first double word in the flagfile. This will produce only a "checksum error" which can be ignored...
Edit:
However I am getting "Sectornumber sould be XXX. Drive reports XXX"
What difference is reported? Only a few sectors?

[Piyoko]

I just have to thank you for this awesome application! I had a bonus disc from a retail DVD that came loose during shipping and was badly scratched. Tried DVD Decrypter on a bunch of different drives and it was just doing retry upon retry upon retry without getting anywhere. So I also put AnyDVD in the background to get rid of the CSS and let this app run, and after only 10 minutes and a couple of passes over the problematic section I had a perfect iso on my hard drive. Awesome! :)

[sirzur]

I can't remember the difference between the sector numbers but it is quite small.

I think I will wait for a new version rather than editing the flag file

[marsoupilami]

Well!
The new version 1.4 will have:
.) A Force Volumesize dialog for fooling the disksize
(for solving the "empty disk" and the "different sectornumber" problems) - be careful!
.) A Cycle Tray checkbox
.) The Filenames are now only suggested

[LakersFan]

Sounds great. Looking forward to using this. I love the cycle tray option.

Thank you.

[ilpippo80]

Hello Masoupilami,
first of all thanks, great program, it's really useful!
One small feature request:
I wanted to tinker with the number of skipped sectors to see if I can get any speed improvement with it... Do you think it's possible for you to add a switch (maybe just in the ini file so probably it would be easier for you) to choose a fixed number and not a random one for it?
Thanks!

[LProof]

Thank you so much for this wonderful program. I've already recovered 3 bad DVDs I had no hope of recovering before!

Possibly something obvious that I'm missing - but is there a way to tell exactly which sectors are missing - a program I can use to read the *.flg file? Wordpad gives me gibberish.

[marsoupilami]

@ilpippo80
There would be a lot of switches & adjustments - but I tried to keep the use as simple as possible.
Besides: The random number of skipped sectors is adapted to the missing part - the more sectors are missing the larger the steps.

@LProof
If You have a look to the flagfile with a hexeditor: following the header every bit represents one sector

[snn47]

Is there a way to use DVD-RAM with the programm?

While I had had no problem so far in reading DVD-RAM, unlike CD- & DVD-r(w), I am wondering is there a way at all to recover data from scratched/unaccessable DVD-RAM, if it should happen?

[mgpc]

And is there any way to get it to read a disc (DVD-R), which burned and verified ok, including a check with Nero Cd/Dvd-Speed Scandisk...

and now, about a year later is showing as an empty disc! It's *not*, I don't LABEL discs until I've verified I can read the content, ever.

And I've got about a dozen of them, and they are all Datawrite Yellow's with absolutely useless CMC MAG AE1 dye. Imgburn says they are empty, so does Isobuster and IsoPuzzle, but they AREN'T. Visibly they aren't!

I will never buy anything but Taiyo-Yudens again, even if they are more expensive.

marsoupilami, is there any way to force IP to just read off all the data, even if the disc appears to be empty? Maybe then I can get IsoBuster to scan the ISO for me and extract at least some of the data.

Mark

Oops, just looked at the Changelog for 1.4, it looks like it already does this, let's go and check.....

Umm that's recovered 2 for 2 so far, with the third proceeding happily, though on one disc it took ages (10min+) to go off Drive Not Ready.

[marsoupilami]

@snn47
Sorry - I have no DVD-RAM ... neither at home nor in our company

@mgpc
Sorry but maybe I'll dissapoint You too...

A few years ago I also had a couple of SWISSTEC disks which verified ok and a few month later they seemed to be 'empty'.
Although the surface shows exactly which region of the disks had been written to - I could not recover even one sector!
I definitely did not store these disks in light or hot areas - they seem to have been rotten in the dark....

I hope You will have more luck and you got it in the beginning of the dissolving process!

[mgpc]

marsoupilami, I'm currently at 4 discs read fine, 1 that had a couple of read errors so I've paused it at the moment, and 1 more on the way, plus one I haven't been able to get anything off yet because IP never comes back saying Drive Not Ready, even though ImgBurn reads the disc info quite quickly, but I guess it's reading elsewhere on the disc.

Thanks again.

Mark

[mgpc]

For those discs that don't seem to come back from DriveNotReady, any chance we could have an option to NOT try to get any details off the disc, like the drive reported sectors or the label, and just start reading, to the full size of the media. I realise this will get lots of errors towards the end of the disc, but might still get enough data to be useful.
Force Size = Max for this media maybe? With possibly an option to reduce the size when we figure out where the end of the written data is? And I've just come across another disc that IP won't even start on because it can't read the label. Does it ever timeout?

Now at 5 read ok, one with some errors and 3 saying DriveNotReady.

IP is brilliant!

<pedant>BTW, on the doco page for IP1.4 you have a heading of 'Cylce Tray' instead of 'Cycle Tray'.</pedant>

[mgpc]

I have now recovered 8 out of 9 discs that Imgburn says are empty, the last one IsoPuzzle sat trying to read the disc details for about 10 hours before I gave up. It sometimes took half a dozen disc mounts before IP found the details, but then it recovered them perfectly, even the disc that had errors reading first time.

Mark

[setarip_old]

@mgpc

Hi!I don't LABEL discs until I've verified I can read the contentAre you referring to adhesive labels? If so, be advised that they may contract with time, causing your discs to warp and make them (temporarily) unreadable.

Removing such labels usually makes them once again readable...

[snn47]

Some feedback from my test, in short I was unlucky in recovering a badly scratched DVD (dull surface ) and got a variety of error messages.

I tried several drives drive brand No interface,
LG 4167 IDE,
Plextor 740 FW ??,
TEAC DV-W50 FW Prolific,
Pio KV14 USB2 ??, and
an older PIO 116 ROM FW Prolific.
System ASUS A7V266E, KT266A/xx33 SB, AMD XP1800, PCI TI FW controller.
XP-Pro, Nero 6.6 & InCD (turned on/off made no difference when errors were received)

Isopuzzle didn't start with the ROM, only with DVD-Burners.

I made 2 ISO-files one with the LG (in under an hour), the Plextor took about 1 1/2 days for the second ISO-file.

Both didn't recover the data, so I tried the other drives, using the flag-file or ammending the ISO.

Isopuzzle started
Continue 4167.iso
Loop 1 read 2006176 Errors 0
and stoped with the error message
Exit 0x5c: Disk full

When I deleted the iso-files and tried to create a new one I had
create .iso
Exit 0x0: 0k
Exit 0x0: Disk is Empty

trying an old flagfile
Disksize changed to "Existing Flagfile" (2253343 Sectors)
Continue .iso
Exit 0x2: FOpen *.iso

Does this make any sence to you?
I tried every combination I could think of just short of a new XP installation.

Does ISOpuzzle use the registry?

[marsoupilami]

Hi there!
I didn't understand everything - but I'll try...

Isopuzzle started
Continue 4167.iso
Loop 1 read 2006176 Errors 0
and stoped with the error message
Exit 0x5c: Disk full

Maybe You tried to write to a non-NTFS partition?
(The warning is often supressed because of the unreliable media identifier returned by the system)
Writing more than 4GB to a non-NTFS disk will produce a write error and the 'disk full' message!


When I deleted the iso-files and tried to create a new one I had
create .iso
Exit 0x0: 0k
Exit 0x0: Disk is Empty

Although a DVD is readable some drives return 'zero size' (= empty disk) - that was one reason for 'force size' dialog


trying an old flagfile
Disksize changed to "Existing Flagfile" (2253343 Sectors)
Continue .iso
Exit 0x2: FOpen *.iso

On my XP (both 'home' and 'pro') I have a file locking problem very often - especially (but not only) if I terminate a program via the taskmanager the accessed file remaines locked (deny write) until the system is rebooted - this might cause the 'file open error'.... i guess this could be a reason
Another possibility could be 4GB-problem on an non-NTFS (as above) ??


Does ISOpuzzle use the registry?

No! The 'destination path' etc. is stored in the *.ini

[snn47]

:thanks: Thank you for the fast reply.

Since I do own such "nice" very badly scratched DVD I can't wait to try again next weekend.:devil:

You are correct that I terminated isopuzzle via the taskmanager and did not reboot.
However while my HDD has mix of NTFS and FAT32, I am pretty shure that I had the problem with the NTFS partition.

When I changed DVD-drives I use a second folder on the same partition. However when I copyied the flag-file from one folder to the other and started Isopuzzle new it still asked for the isofile.

[rookandpawn]

Lets see, Ive tried DvdDecrypter in iso mode, ive tried Cd/DVD data Recovery, ive tried CD Check, ....

None are as fast and do what i want more than IsoPuzzle.

Way to go man!

[poprock]

I have tried Stellar Phoenix CD Data Recovery (http://www.stellarinfo.com) Software It Provides recovery of data and file, lost due to unreadable, scratched, corrupt or defective CD media. Phoenix CD Data Recovery Software supports ISO9660 and Joliet file system.

[Politiken]

Hi - The concept of marking good/bad sectors using the flag file together with the import function is a great idea !!!

I tested a scrathed dvd - it fullfilled 98 % and then I exited the application (using the exit button). I later restarted the job but IsoPuzzle fails with these messages:

regards Pol

[marsoupilami]

Hi Politiken!

Sorry about the very late answer ... I have been on holidays.

Obviously your *.flg has been corrupted (don't know the reason...)
The drive reports 3847037 sectors - multiplied by 2k => 7.8 GB (this looks plausible for a DVD-DL)

Besides ... do you have more than 8 drives connected? (physical + virtual)

regards

[saulob]

A few bugfixes ...
NEW VERSION - ALSO AVAILABLE HERE
http://www.geocities.com/marsoupilamis/

Please visit the homepage of NERO AG to get the WNASPI.DLL and read the terms of use
http://www.nero.com/nero6/eng/WNASPI32.DLL.html
:)

marsoupilami, thanks for the program, but... the WNASPI32.DLL download link from Nero.com is gone....

Any mirror ?

thanks :)

[marsoupilami]

The page with the link has been removed but the file is still there - it is fixed now!

[saulob]

The page with the link has been removed but the file is still there - it is fixed now!

marsoupilami, thanks for the mirror on your site.

the program is working fine with it :thanks:

[imatasion]

Great program and much needed! Unfortunately, I may have stumbled upon a serious bug (1). Also, here's a feature request (2).

(1)
My old WinXP reinstall disc (CD - 448MB), aside from the usual light surface scuffs, had gone really bad (don't ask me how). I used IsoPuzzle v.1.4 on it, starting on one computer, and continuing through 3 others until it was completely read (the last computer's drive was the best; it zoomed through those remaining tough-to-read sectors).

I burned the resulting .ISO to a CD-RW and attempted a test reinstall. At one point, I received a fatal-install message about an error from some file line-number. I looked into that file and discovered it was full of hex-zeroes. The same file from the original disc (easy to read, so it was not in a bad sector area) had all the proper data. Somehow, IsoPuzzle read the original disc sector, which contained this file, but didn't write the retrieved data into the restored image area.

Retrying the restore exclusively on that last computer produced the same results on the very same file. Finding and using your latest version 1.6 and trying the restore on that same last computer, again, produced the same no-data error in that same file.

Could you look into IsoPuzzle and see if you can find such a bug? As yet, I have not manually searched through my .ISO (on my test, rewritable CD) to find other missing-data files.

(2)
Is there any way for IsoPuzzle to temporarily disable Window's autorun of the source disc while "Cycle Tray" is active (and, of course, return it to it's normal state when Cycle Tray is unchecked or IsoPuzzle exits)? As you might guess, everytime the tray is cycled, Windows will run an autorun disc (which happened to me during this restore attempt).

Thanks for any help!

[imatasion]

After doing some experimenting while waiting for the author's reply, it looks like IsoPuzzle definitely has a problem. A little Internet searching brought up IsoBuster as good software for disc recovery. Running it on the system that most easily read my damaged disc, I was able to recover it with only a few "retries" for the difficult areas.

Reading and comparing information on the 2 recoveries showed IsoPuzzle clearly lacking in size of total blocks recovered.

<strike>IsoBuster v.2.3</strike> IsoPuzzle v.1.6 recovered .ISO file
file size - 474857472
blocks - 231864
block size - 2048

<strike>IsoPuzzle v.1.6</strike> IsoBuster v.2.3 recovered .ISO file
file size - 474859520
blocks - 231865
block size - 2048

And, at least in one instance (described in my previous post), IsoPuzzle didn't write into the recovery file what it had read from the disc. The final proof was using the recoveries -- IsoBuster's recovered disc worked, IsoPuzzle's didn't.

That's all I have for now. Until the author checks through these findings, I caution IsoPuzzle users not to rely on it exclusively for bad disc recovery.

[marsoupilami]

@imatasion
Thanks for your detailed description....
...but please keep in mind, that im working on this program in my free time!
I try to answer any question but there is no need sending pm, email and a two postings simultaneously - I am definitely not anyones hotline!

If you are satisfied with Isobuster feel free to use it and don't forget to pay for it!

If you are still interested in my answers to your previous questions let me know.
kind regards

[imatasion]

@marsoupilami

Hi. I reported what appears to be a serious recovery software error (glad to help!). A pm and later an e-mail were just to get your attention to the forum (I did notice how your postings tended to be months apart). I'm sure the posters here, who have relied on your software to recover their important data, wouldn't think that excessive. :) (Also, I should point out that my postings weren't simultaneous nor duplicated. Rather, a week apart, and as I discovered and reported additional findings. Nothing improper there, either.)

Don't get this wrong, but I hope there IS a bug and that you can fix it. Time permitting, of course. ;)

[blutach]

@imatasion

Please remember that ISOPuzzle is freeware and marsoupilami owes you (and everyone else), absolutely nothing. Whether people rely on ISOPuzzle to "recover their important data" or not, nowhere has marsoupilami indicated that he is there to provide support free of charge. In fact, explicitly, the readme states:

! YOU ARE USING THIS PROGRAM ON YOUR OWN RISK !
===============================================
By using this program you accept that no warranty of any kind can be given.
Neither functionality nor data integrity can be guaranteed.
The author cannot guarantee that this program is free of viruses, trojans,
spyware or any other harmful software.
The author cannot held responsible for malfunctions, data loss of any kind,
corruption of any other software installed on your computer, permanent
hardware damage or profit loss.

While I agree that you have not peppered the forum with multiple posts, it really is up to marsoupilami to respond or not. He has now indicated that he will reply when he gets round to it.

Regards

[marsoupilami]

Hi there - Don't worry! ;)
This is my 1st day back in the office (after x-mas) where I have fast internet access and more than one pc for testing something.
I agree, that it is really frustrating downloading a couple of programs only for discovering that they don't work properly.

I have tested all error conditions which I could imagine very carefully and I didn't touch the part of code which writes to the iso file since the earliest versions - thatfore I doubt hopefully that the described behaviour could be an Isopuzzle error.
The only bug I know currently is:
If you have more than 8 drives connected (both physical and virtual CD/DVD drives) the behaviour is unpredictable.
I never tested what happens but I guess it will crash - this will be fixed in 1.7.

As I understand all the erroneous 'zero-data' had been produced by the same (the last) CD drive.
Generally Isopuzzle relies on informations returned by the ASPI driver: drive status, read data, read errors etc.
Since wnaspi32.dll came along with Neros 'Burning Rom' (as long as W98 was supported) - I treat it as an 'error free' part of a well tested commercial product!
The most suspect part in my opinion could be the 'fast reading' drive itself (especially its fimware).
I remember an old combo drive which could 'read' everything errorfree - even a test CD with a black marked area!
I'll try to get it back for testing - but I expect random data instead of all zeroes.

What I think what could had happened is:
Is it possible that you accidentially didn't transfer both the *.iso an the *.flg file to your 3rd computer?
If you tranfered only the *.flg this would create an *.imp file which has to be 'imported' to your iso image. This would explain the 'zero data'.
Please search your disk for such an *.imp and let me know!

I cannot believe, that such an error - which I had never heard about - occurs twice...

Regarding your feature request I'm very sorry to disappoint you:
First of all I think all 'auto-whatever' are (really boring) Windows 'features' - Please understand, that I don't like to grab into registries and so on.
Last but not least I'm trying to keep IsoPuzze as much as possible independent to Windows versions - I cannot (and don't like to) buy and install a lot of different ones for testing purpose.
Thatfore I recommend to disable either the 'autoplay' function or the 'cycle tray' checkbox.

kind regards

[imatasion]

@blutach

I know all that, thanks anyway. ;)

@marsoupilami

To eliminate some debugging complexity, I'll ignore my initial results with multi and single drive recovery, on the now obsolete version 1.4. (BTW, I did correctly follow instructions on using multiple drives; you made it easy & convenient to do -- good design.)

I used the links from the first page of your thread to get the required software: IsoPuzzle-16.zip & wnaspi32.dll. These were expanded (as needed) and placed in an IsoPlay1.6 directory.

So, for repeatability testing, I again set up a "new install" of IsoPuzzle 1.6. It was run without changing the default settings, except to un-check "Allow Cooldown" and to set "Destination". The recovery ran quick with only a few retries.

Resulting IsoPuzzle 1.6 recovered ISO was 452 MB (474,857,472 bytes).

Again, the IsoBuster 3.32 recovered ISO was 452 MB (474,859,520 bytes).

Now, if I try to view the "empty file" (that bad-recover file) in 7-Zip, the file comes up empty. In IsoBuster, the file is "x" marked and it shows no data.

Here is something interesting!

The IsoPuzzle recovered ISO has the empty file is at sector 231864. In IsoBuster's viewer, that sector comes up as a "device error". Why? If I navigate backward, sector data shows normally. So the actual recovered sectors are 0 thru 231863 (or 231864 sectors). A full sector is 800h or 2048 bytes. 231864 x 2048 = 474,857,472 bytes. This is the recovered amount from IsoPuzzle's ISO. The missing file data comes from a sector outside the ISO!

However, the IsoBuster recovered ISO also has that file at sector 231864, but with it's data intact. Navigating forward, there are no more sectors -- it's the end of the ISO. So the actual recovered sectors are 0 thru 231864 (or 231865 sectors). 231865 x 2048 = 474,859,520 bytes. This is the recovered amount from IsoBuster's ISO. The file data exists because the ISO is big enough to contain it!

Conclusion? IsoPuzzle is not writing the final recovered sector into it's ISO. This is where you should start your code checking. These kind of "boundary errors" often show up in programming and it looks like it bit you, too.

Hope this helps. ;)

BTW, when you come out with v.1.7, I recommend you include a critical warning not to use any previous version, because of this bug.

[marsoupilami]

@imatasion - thank you very much for giving such a detailed description and spending that much time!

First I was a little confused because of your previous post (where the IsoPuzzle recovered file should be the larger one...).

I grabbed into the code and - No! I didn't make the beginners mistake of forgetting the last loop.
Obviously the Aspi driver reported disk size is a little sloppy - this is both disk and drive dependent!
The value in the progress field prints out the reported sectornumber 'as is' - also the greyed out field in 'force disksize' shows this value.
Besides - my 'force size' -> 'estimated' function has really an error: it detemines the last readable sectornumber which is size-1 - so I trusted both.:devil:

I could now simply add one to this number - I found out that some drives have the capability to read ahead beyond the reported size without an error but some do not!
The good news:
It seems that most images have at least one 'zero sector' at the end of the image - maybe that's the reason why!
E.g. Nero 6 (which I use) appends a lot of - I burned an image and 'isupuzzled' it and burned it again: it was growing and growing!

I'll try a workaround in the final 1.7 like reading ahead beyond the end of the disk with a few retries... what else should I do? (Oh how I hate things like this!)
For the next version I will look into the SPTI interface hoping that it is more reliable - but I expect this would require a little more time.

Btw I could fix the problem which was reported by Politken in beginning of September:
It can occur on a ripping attempt into a locked image file.

Thank's again and I'll keep your disable autoplay suggestion in mind - maybe someone can tell me a moretheless simple solution ;)
Nice weekend!

[setarip_old]

I'll keep your disable autoplay suggestion in mind - maybe someone can tell me a moretheless simple solutionJust hold down the "Shift" key while the tray recycles...

[JohnnyPintas]

Just hold down the "Shift" key while the tray recycles...

Thanks for a most usefull tip, it does indeed work :D

[[]]

[setarip_old]

Thanks for a most usefull tip, it does indeed workYou're quite welcome ;>}

[imatasion]

@marsoupilami

Hi, and sorry about the "cut & paste" typo showing ISOBuster having produced the 1-sector-short ISO file. I've edited that text to correct the mistake.

Also, sorry to see this may be a third-party problem (harder to fix). Well, here's some more information that may help:

The main test was on a Windows2000 system with a Lite-On DVDRW drive.

ISOPuzzle, with Nero's wnaspi32.dll (v.2.0.1.74), started the recovery stating 231864 sectors to fill (1 sector short of the actual amount). I searched the hard drive and found another wnaspi32.dll (v.1.0.0.14) from Acronis hard drive backup software, which I put in place of your recommended Nero version. Still, ISOPuzzle reported there were 231864 sectors to fill.

Then, I installed ISOPuzzle on a WindowsXP system with a no-name CDRW drive.

Again ISOPuzzle, with Nero's wnaspi32.dll (v.2.0.1.74), started the recovery with 231864 sectors to fill (still 1 sector short). I searched the hard drive and found another wnaspi32.dll (v.2.0.1.59) from Nero disc burning software, which I put in place of your newer Nero version. Still, ISOPuzzle noted there were only 231864 sectors to fill.

BTW, I looked into the ISOBuster folder but found no wnaspi32.dll. They must have implemented something internally.

To sum up, with 2 operating systems, 2 disc drives, 3 drivers, ISOPuzzle is still seeing 1 sector short. Just more sauce for the goose (as Spock would say in some Star Trek movie). :)


@setarip_old

The old "hold down the shift key" trick only works if one is watching the recovery. As some recoveries go on for many hours or even over night (hence the "cool down" check-box), that might not be practical. I'm afraid one must manually toggle Windows's "CD autoplay" option in those instances.

[marsoupilami]

@imatasion

Thanks for your findings - seems that's not a Neros bug...
If I misinterpreted the reported size ... shame on me!

What I found out during the weekend:
1.) Both Nero Aspi and Frog Aspi reports '0' for an empty disk while size for disks with content is one sector more.
2.) This 'extra' or 'hidden' sector is sometimes not present.
3.) This 'extra' sector is sometimes not readable by specific drives.

What I'm doing now - I hope this would be clean solution:
1.) This sector is included from now on.
2.) On starting up copy process this sector will be read first.
3.) If it is not readable at the first attempt a warning box pops up and the user can decide to limit the retries for this specific sector (if it's the only missing one).
This prevents senseless torturing the drive while keeping alive all other functions.

I don't know what IsoBuster does but I think it will use IOCTL interface.
As I stated in a previous post I'll play around a little with IOCTL in the future (as I know it is part of Windows but not available in W9x).

Currently I'm trying to find out if this sector often contains important data - until now it doesn't look like.
Even if this would be true it wouldn't be very helpful for you... I know.

However - I treat this missing sector as serious and will suspend my work on some new features.

EDIT: '1.7 Beta' is up now!

Sorry about the inconvenience and thanks for your time!:thanks:

[imatasion]

@marsoupilami

Thanks for all your work! I ran a (quick) test on my scratched disc and discovered the new beta has corrected it's missing sector problem. Hopefully, that also will handle the various disc-reading scenarios you described.

-

This last test recovery on my scratched-up disc went "quick". Why? I searched around for a disc-repair solution to make, if not for any other reason, the scratched disc easier to read for recovery. There a several commercial products (and home remedies) that claim to do the job. Over at Burning Issues (.net), CD Repair Kits - versus - various goops from kitchen & bathroom (http://www.burningissues.net/how_to/scratchrepair/scratchrepair.htm) has comparative tests and shows what actually works. Their conclusion? Brasso Metal Polish, available at your local supermarket or hardware store, beats them all in both results and price (under $3). I tried it on my disc. Three rubbings really reduced the scratches and made the overall surface like-new shiny. Better still, this time, ISOPuzzle read the entire disc without a single retry!

For this type of unreadable disc problem and Brasso, you may not even need ISOPuzzle! (Sorry.)

:)

[snn47]

@marsoupilami

So far I could not save any defective Disc but I am still trying out each new version.

I normally try four drives two slimline Pioneer DVR-K06, and Matshita UJ-824s, and a TEAC WD-50, LG 4167.

While three worked fine one gave me this message
"Loop 1 Read 255277 Error 0
Exit 0x0 0k"

I compared the ISO-files with windiff.exe and all are identical.

What does this mean?

[marsoupilami]

@snn47

Especially during debugging & testing the "log window" was very useful for me... obviously you turned it to "verbose" once:
Loop 1 Read 255277 Error 0
Exit 0x0 0k means: 255277 sectors read in 1 loop (no retries) with no errors.
Exit 0 is the return code "OK" of the copy task

Your files should be identical.
:)

[snn47]

Thank you,
I was confused because with the other drives I did not get this message just the exit and that the flagfile was deleted at the end.

I did not have any CD/DVD software installed except for DVD-RAM driver. When I have some time I will test isopuzzle with an naked-image xp-pro-sp2, where I have not even installed the DVD-RAM driver.

[laserfan]

...Brasso Metal Polish, available at your local supermarket or hardware store, beats them all in both results and price...The recent bumps caused this post to catch my eye. I too have used Brasso along with many other methods and I think it's (surprisingly) the best too.

[GrofLuigi]

Doesn't work on SATA for me. Shows Unknown Volume and reads all 0`s. I think it's due to the Nero Aspi. I tried also FrogAspi, that doesn't work even on IDE.

Anyone know a solution?

(I have 18000 sectors left to recover, on my big computer every pass yields 10-50 new sectors, and it's a torture for the drive even with cooldown - it has been running for 20 hours now).

GL

[snn47]

@GrofLuigi
I own two SATA drives and both work with isopuzzle, via the SATA controller or a SATA to USB 2.0 bridge, only with a bidirectional IDE to SATA bridge I have some problems which I have not been able to narrow down.



@marsoupilami

Over the past year I bought A LiteOn iHAS424 and a LG GSA H20, and received a few older ROM drives from a friend, but still no old Plextor drive 716 or older.

I ran additional tests over the holidays with all drives comparing Isopuzzle 1.7 to that of Isobuster professional version 2.6, and here is what I observed

- No drive was able to recover all my hard to read DVD.
While in average my Matshita provided the best overall results followed by my LiteOn and TEAC, there were some DVD on which the normally unsucessfull LG were more sucessful.
Some drives were never sucessfull.

- Some recovered ISO proved to be defective, when I play all contained mpg files, the playback stopped, while wthe ISO recovered from another drive worked.
While some drive provide “correct“ output in fact they do not . Prolonged reading provides more sectors read but not necessarily correct one.

- Isopuzzle also works with DVD-RAM.
Error correction responds on DVD-RAM not to well to fingerprints and dirt. I don’t know if this is also the case for normal DVD.

- You can’t recover what the drive is not able to read. In several tries I had my drives try it for a few days and still ~70000 sectors were missing, using a my slow Matshita slim-line drive I had an ISO in less then an hour, even with other programms .

- I still am not to shure how many sectors a DVD has if the drive does not provide a feedback. Furthermore while I have not read of any drive that is able to overburn, this might require a more flexible handling of ISO size as discussed below.

This brings me to the following questions

- Do you think reducing the speed of a drive might differ the results of the error correction?
I remember reading that some drives can even be made to read at 1x speed, but I have not tried if Nero Speed works.

- Is the error-correction dependent of the DVD specs detected by the drive or is this only of interest for burning DVD?

- Do you think it makes sense to compare images made with different drives of the same DVD.

So far I assume that different sectors that are more or less ok will have identical contense and could be identified by assigning a quality value (=number of identical reaults from different drives).

Hard to read sectors may differ from drive and error correction to drive and error correction.
If stored in a Matrix, only the identified hard to read sectors are the only ones that need to be verified on different drived, while those that remain identical on 2 or more drives would just improve the quality.

If the recovered information of e.g. n-drives is stored in an n-deep matrix, each read info x sectors wide and create the ISO out of the matrix rather then having to start a new iso. This way different ISOs could be created varying in number of sectors and which hard read sector result from which drive is used. Do you think such a Matrix-ISOpuzzle is workable?


Unlike ISOBuster, which operates more or less automatic only with a percentage as feedback, I’d say that Isopuzzle allows user to better judgement if waiting for a prolonged time seems promising or not. However neither ISObuster nor ISOpuzzle can guarnatee that the resulting ISO is ok.

[Ghitulescu]

I ran additional tests over the holidays with all drives comparing Isopuzzle 1.7 to that of Isobuster professional version 2.6, and here is what I observed

- No drive was able to recover all my hard to read DVD.
While in average my Matshita provided the best overall results followed by my LiteOn and TEAC, there were some DVD on which the normally unsucessfull LG were more sucessful.
Some drives were never sucessfull.
The best readers are the DVD-rewriters from LG. They beat by far any LiteOn, Matsushita or Teac (which are LiteOns). Older Pioneer (DVD-ROMs) are also quite good.
Nuance: there are several defects and disk types and thus there are many good drives, some are the best in reading pressed DVDs, other support the DVDRs better, also the LGs are more tolerant towards jitter, while the Pioneers can withstand scratches better.

While some drive provide “correct“ output in fact they do not . Prolonged reading provides more sectors read but not necessarily correct one.
Then either the FW is erroneous (it would yield defective data also on clean DVDs, it was a case with some drives not very long ago, in the same galaxy) or the data is beyond recovery (any reader would yield the same result).

- Do you think reducing the speed of a drive might differ the results of the error correction?
I remember reading that some drives can even be made to read at 1x speed, but I have not tried if Nero Speed works.
Yes.

- Is the error-correction dependent of the DVD specs detected by the drive or is this only of interest for burning DVD?
They're the same (for DVDs).

- Do you think it makes sense to compare images made with different drives of the same DVD.

Usually no, but it helps finding if one particular drive has problems in correctly extracting the data from a DVD.
If the recovered information of e.g. n-drives is stored in an n-deep matrix, each read info x sectors wide and create the ISO out of the matrix rather then having to start a new iso. This way different ISOs could be created varying in number of sectors and which hard read sector result from which drive is used. Do you think such a Matrix-ISOpuzzle is workable?

Yes.
Unlike ISOBuster, which operates more or less automatic only with a percentage as feedback, I’d say that Isopuzzle allows user to better judgement if waiting for a prolonged time seems promising or not. However neither ISObuster nor ISOpuzzle can guarnatee that the resulting ISO is ok.
No, it's the drive that guarantees that. ISOxuxxxx allows you to extract the best from your drive.

[GrofLuigi]

@GrofLuigi
I own two SATA drives and both work with isopuzzle, via the SATA controller or a SATA to USB 2.0 bridge, only with a bidirectional IDE to SATA bridge I have some problems which I have not been able to narrow down.
This is a laptop in SATA AHCI mode and nothing else (for example IDE emulation) selectable in the BIOS. :(

Otherwise, as for you other questions, in all discs I've recovered (around 10) there were no errors if IsoPuzlle said so - no bogus data.

And I still can't grasp the concept of different sizes (what to do exactly in that case). :stupid: :)

GL

[snn47]

@Ghitulescu

The best readers are the DVD-rewriters from LG. They beat by far any LiteOn, Matsushita or Teac (which are LiteOns). Older Pioneer (DVD-ROMs) are also quite good.

I can speak only for the recovery of the bad DVD with the DVD Rom and RW I tested with. My TEAC W50 is, except for the firmware, identical with Pioneer drive (PIO 106). My Pio 116 DVD-ROM was even worse, and while I have another SCSI Pio DVD-ROM, I lack a XP driver for my Adaptec 2930 SCSI controller to test it.

the FW is erroneous (it would yield defective data also on clean DVDs, it was a case with some drives not very long ago, in the same galaxy) or the data is beyond recovery (any reader would yield the same result).

Unless I have a flawless copy I wonder how I should establish which image was correctly recovered. In case of AV Files, e.g. mpg2 the player would just stop working once the reach the corrupt data from bad sectors or showed distorded video. If I have data files of unkown contense or when corrupt archives are involved I wonder how you propose to establish which ones are correct.

@GrofLuigi
And I still can't grasp the concept of different sizes (what to do exactly in that case).

If you look at marsoupilami reply from the 14th January 2008 reported sector may differ or not reported correctly. So I was thinking, independent if you "free set" the recovery for to few or to many sectors, you always would have to recover the DVD again and again if the ISO was not of the correct size. I thought it would be easier if you don't get the correct size reported by the drive, to grab all sectors from a DVD and be able to vary the ISO size by beeing able to adjust the number of sectors.


PS.: While I have enough corrupt or hard to read DVD, luckily I have information backups, but none of those DVD is a clone to the bad ones.

[Ghitulescu]

My TEAC W50 is, except for the firmware, identical with Pioneer drive (PIO 106). My Pio 116 DVD-ROM was even worse, and while I have another SCSI Pio DVD-ROM, I lack a XP driver for my Adaptec 2930 SCSI controller to test it.

Most Teacs are LiteOns, with some exceptions, I'm no longer a Teac fan (since they left the SCSI field). Sorry for my mistake.

Teac DV-W516D is Pioneer 108
Teac DV-W516GA is LiteOn 1653
Teac DV-W516G is LiteOn 1633

TEAC DV-W58G is LiteOn LDW-811S
TEAC DV-W58G-A is LiteOn SOHW-812S

The same for W512 and so on....

Your W50 has 2 versions, according to the end letter D (+-RW, Pioneer 106) or E (-R/W only, Pioneer 105).

[Boulder]

I cannot get ISOPuzzle working on my notebook at work (OS is WinXP SP3). I wanted to try to read a badly scratched DVD there as my home computer doesn't seem to be able to recover it. I've copied wnaspi32.dll to ISOPuzzle's folder but I still get "No drives detected" upon startup, after which the program crashes.

[Ajax_Undone]

Is this Project still alive???

If not is the source code available cause there are a few things I would like to change....

[setarip_old]

@Ajax_Undone

Try contacting the author at:

mailto:Marsoupilamis@Yahoo.Gr
http://www.geocities.com/marsoupilamis/

[Ajax_Undone]

Thank You setarip_old the email might work but website went out with Geocities this last year or so...

But Yes I will send an email...

Indeed thank you very much =;}

Here is his new site btw if anyone cares.. http://www.marsoupilami.esmartweb.com/

[marsoupilami]

Hi there!

Is this Project still alive???


I only know a minor bug which doesn't need to be corrected. (It leads to slow reading, if you need more than 30 seconds for selecting a filename :) )

The second and more severe problem is that some drives are not recognized by the program. The reason is unknown to me because all my desktops and laptops which are accessible to me are working fine. I don't even know if it has something to with SATA, specific drive types, windows drivers or whatever.

Please do not expect me to buy something on my costs even if the donations would be sufficient - and indeed they are not.


If not is the source code available cause there are a few things I would like to change....

Would you release the source for free?

p.s. yes, the geocites site is down since one year...

[mvBarracuda]

Would you release the source for free?
I would actually do so :-) The main question is: do you consider it?

[spotter]

I think you should consider making your log format compatible with GNU ddrescue.

[Ajax_Undone]

I would release the source under GPL, but of course that is me.

Respect always

~Ajax

[Ghitulescu]

I know this is a very old thread, and that IsoPuzzle appears not to be developed anymore, but still...

My problem is that I have a DVD that is badly scratched. Shortening a long story, I managed to read most of the DVD, yet the image is longer than the actual data. I tried every solution I came up but either it started anew or spit an error (sector mismatch). I simply want to skip certain sectors from being read, those part of the image but actually pointing at unburned sections, sparing my drive.

I know the solution is to set the sectors to "free" but this means it starts anew and I needed 2 weeks (yes, 14 days) rotating some 50 drives to reach the actual 80% (of which only some 2000-3000 sectors right at the beginning are the problem, the rest are those "fake" ones). As I said, the situation is extremely strange, de disc was really really severely damaged.... I fairily know now, only after those days, how long the image should be...

Is there any solution to set the sector numbers AFTER the image was created? Sort of Truncate/Trim etc...
If not, will be any plans to continue the development?
Maybe the sources? With docs?

PS: I tried:
- setting the sectors after - error
- setting the sectore before - starts anew
- setting the sectors to anything else than what was in the flg/iso - error
- copy/paste/rename/reuse shorter images - error or starts anew

I know these are mechanisms against the idiot user, and few people actually met the situation in practice where the drives provide erroneous infos (because of the errors in the LeadIn/TOC).

PS2: well, no, these are home movies, can't be bought again.

[maetel99]

If you have a scratched up disc, you can clean it up pretty well using auto scratch remover products such as "Formula 1 Scratch Out" or "Nu Finish Scratch Doctor". Just polish it repeatedly with the scratch remover and you can clean up most of the damage. Do a final clean with alcohol then you should be able to read out more of the data without so many problems.

[StainlessS]

Smokers toothpaste [and a little water] is sometimes recommended.
Either round and round in little circles [as for bulling military boots], OR,
out from center, lines radiating from center, not following rotary direction of tracks.
[scratches radiating from center can be recovered by error correction, scratches in
direction of tracks can defeat error correction]

EDIT: I used to have a hand wind up thing which spins disk to clean it up with some cleaning solution, with slightly abrasive pad.
In the UK, CEX, a 2nd hand [ie used] I.T. shop that sells used DVD's, etc, also has a powered device that
does the same thing as my hand cranked device, they say that if the disk you buy dont work proper, then
bring it back and they will [for 12 months] give you a free spin clean. Maybe if there are similar style shops in your
area, it might be worth enquiring if they have such a device. [I only tried it once after failed attempt at hand polish, it worked for me].

[Ghitulescu]

Thanks, but I asked for a software solution, not a "cleaning" one.

The disc is not scratched, it's a bad burn.
Usually I tested every disc, but after a while I considered it to be a waste of time, the only commoditiy I don't have plenty. In 20 years of burning I had only a couple of media that went defective straight off the burner - they were mostly the result of the buffer underrun, long time ago when the burners needed to have a constant flow of data, and windows 95 always interferred with. Burnproof technlogy was the greatest invention ever after the invention of the CDR.
And it worked, but after some 10 years, it's no longer readable.

It was a bad burn that is a burn that was close to the limits of readability already from the beginning. And I did not tested it. This is why goddess Fortuna hit me. It was a CMCMAG by the way.

I understand that the number of sectors of the imgae is stored, this is why I needed the sources, to know the offset and change it in a hex editor. Or the author can do this in the new version of the software.

[jpsdr]

From what i've heard, 10 years is the average lifespan of burned media, as they... i don't know how to translate properly... "decay"... over the time.
Pressed (not sure proper translation) don't have this issue. "Bad" disc may have other issue link ink on the printed face diffuse (again translation) and damage data layer. Otherwise estimated lifespan of pressed media is a lot higher (80 years..?).

So if your disc is around or more than 10 years old, it's "just" the end of his lifespan.
After loosing a lot of discs 10-15 years ago because of this fact i didn't know, now i always put on my disc the year it's burned.
And i redo them after 7 years.
Lost a lot of things, especialy family VHS film/record i've digitalised and put on DVDs around the begining of the first DVD burning, (i've done this at the end of the 90s if i remember). Now they are unreadable, and i'm not able anymore to digitalise the VHS source, should they still be avaible.

[Ghitulescu]

Yes, I know these, I burn since the first burners roared in the computer shops.

I did extensive tests before, and after a while I did tests only for each new brand, never had issues. I have CDRs from 95-96 that are as happy as bacj then, insignificant more C2s than fresh, and DVDRs from 2003-04 also with almost unchanged PI/POs.

It has to be a bad disc as none of its brothers from the same splidle exhibit such fatal error (the inner section, usually this happens at the outer rim section).

I know it's my fault that I did not test the DVDR for quality, but I was tricked into thinking it should be the same quality as the first disc of the spindle. This won't happen again :)

Now, this doesn't help me to extract the unique data (yes, another mistake).
I might end up with a totaly unrecoverable DVDR, which is highly plausible now, after some 50 drives of all makes (except BTC, which I have sold few years ago).
If this will happen, I might need to recreate the DVD again from the original tapes (capturing, editing, titling etc.).

Thanks for interests

[SeeMoreDigital]

My problem is that I have a DVD that is badly scratched.

Smokers toothpaste [and a little water] is sometimes recommended.
Either round and round in little circles [as for bulling military boots], OR, out from center, lines radiating from center, not following rotary direction of tracks [scratches radiating from center can be recovered by error correction, scratches in direction of tracks can defeat error correction]

Thanks, but I asked for a software solution, not a "cleaning" one.

The disc is not scratched, it's a bad burn.But, but, but... What? [rhetorical].

[Ghitulescu]

It is not scratched - I thought it is, but when I took it out and looked at it it wasn't.
It was too late to change the subject.
As I said, I rarely have bad burns so I wrote it was scratched as it was the most plausible explanation.

[ReinerSchweinlin]

dvdisaster follows a similar approach, maybe it is able to recover some of your stuff:
https://www.heise.de/download/product/dvdisaster-22039