Redump Forum

Small fix for option /rc and /raw being impossible to choose...
Fixed bug with option /be not selectable

https://github.com/saramibreak/DiscImag … 441a5c7ba1
https://github.com/saramibreak/DiscImag … 8dd5992bda
^-^

sarami wrote:

F1ReB4LL wrote:https://www.sendspace.com/file/4nvir6 -- misdetected index in "separate tracks" cue, in the img cue the index is correct. Weird, because there's only 1 track.
https://redump.info/disc/37761/ -- not my dump, so I'm not sure if the latest DIC version was used.

I have never possessed the disc that is the only 1 track and exists index. I have already understood the point to fix the bug, but please wait until I get this disc to test.

Fixed.

mediafire no longer works for me, can you just upload the binaries to github as packages for download.

I think I found an undesired bahavior in descrambling code path. In function "DescrambleMainChannel" if "IsValidMainDataHeader" check fails the sector won't be descrambled. But I think it should be descrambled regardless of the result of the check. I have a CD that has an intentional C2 error in first 12 bytes of a sector and the contents of the sector don't match with /be mode.

EccEdc does not detect such sector as bad but IMHO it should. The end result is that there's scrambled data in output image.
Also if Combined Offset is < 0 such errorneous sector will not be reported to mainError file. The code in 'CheckAndFixMainHeader' does not seem to handle cases when Combined Offset < 0.

What do you think?

reentrant wrote:I have a CD that has an intentional C2 error in first 12 bytes of a sector and the contents of the sector don't match with /be mode.

If you read these sectors with IsoBuster - do you get proper descrambled data or some kind of read error? I believe, that proper sync is mandatory for descrambling.

After some deeper tests it seems that /d8 and /be return different data. Audio tracks in /be mode are shifted by 'combined offset' vs /d8 mode. I need to do some further tests but it looks like I'll be redumping everything...

Nexy wrote:mediafire no longer works for me, can you just upload the binaries to github as packages for download.

Uploaded in github.

reentrant wrote:I think I found an undesired bahavior in descrambling code path. In function "DescrambleMainChannel" if "IsValidMainDataHeader" check fails the sector won't be descrambled. But I think it should be descrambled regardless of the result of the check. I have a CD that has an intentional C2 error in first 12 bytes of a sector and the contents of the sector don't match with /be mode.

I want you to dump the disc using IsoBuster too, and I want to know the result.
Yet if an C2 error exists in first 12 bytes of a sector, I think you need to dump the other disc of the same titile to confirm whether the C2 error is intensional or not.
Or if you have other plextor drive, could you try to dump in that. If you haven't, I recommend a Premium2, Premium, PX-W5224.

reentrant wrote:Also if Combined Offset is < 0 such errorneous sector will not be reported to mainError file. The code in 'CheckAndFixMainHeader' does not seem to handle cases when Combined Offset < 0.

I'll add it before long.

After some even more deeper tests here are my results:
1) Audio tracks are properly offset corrected in D8 mode. In BE mode audio tracks are not offset corrected.
2) Data tracks: DIC seems to have problem with VOB Protect CD and D8 mode. I also tested a SafeDisc 1.00.026 and tracks are correctly dumped using both D8 and BE. For VOB discs I'm reripping data tracks in BE mode and the results are CloneCD compatible which I think is desired. In D8 mode VOB protected sectors have corrupted sync and header areas and DIC is not able to descramble them. OTOH in BE mode such sectors are replaced with 55 pattern. Maybe DIC should also replace such sectors with scrambled 55 pattern?

I have a Settlers IV CD that seems to expose another issue with D8 command. Some sectors (variable count) after C2 sector are not read correctly. I'll be investigating the issue later. But I think the conclusion is that D8 ripping is not reliable in case of C2 errors.

Ok the problem with VOB is that DIC is not able to recognize the protection and any C2 errors that are part of protection are not replaced with 0x55. Could you add a command line switch to replace any C2 sector with 0x55?

At least we can defeat VOB this way. Right now it's impossible to come up with proper dump without manually editing the track.

EDIT: Please also put EdcEcc to github. I made some patches to it to fix a VOB protected track.

reentrant wrote:Ok the problem with VOB is that DIC is not able to recognize the protection and any C2 errors that are part of protection are not replaced with 0x55. Could you add a command line switch to replace any C2 sector with 0x55?

Sorry, I have no time to code it for a while because I'm busy with other thing...

reentrant wrote:EDIT: Please also put EdcEcc to github. I made some patches to it to fix a VOB protected track.

Added.