KryoFlux software (dtc.exe) - problem?

[rygar]

My friend ZeeWolf discovered something interesting. It's seems dtc.exe have some problems with creating proper ADFs from dumps. All detailed described in attached PDF:
PAFP_ Case study 1_2024_fix.pdf
Files to analyze in TheZone

[dlfrsilver]

Guess where it's wrong...... remove the ".raw" part of the folder..... you're doing something that windows doesn't like.....

A folder can't have the same name+extension as any other files in the root folder where the game KF folder is.

It's not the DTC.exe problem, it's a windows rule problem.

And you need to make (you or any other guy out there dumping) proper disk dumps. the DTC or the HxC are not magical tools that can sort out physical errors due to disks not clean with mould.

A bad dump is a bad dump, isolate the bad track(s), clean with a floppy cleaner then redump and check that there are no error.

From what i see in the PDF, it's due to an error during reading.

"Summary/advice
1. Dtc.exe from the KryoFlux package is the worst at decoding the flux stream among
those tested here, without reporting errors"

Of course, it's stream, there is no verification done XD !

You need to analyse and check yourself that the disk you dumped is good, and not bad (mould requires clean up, flux displaced due to agressive clean up, etc).

2. To convert from RAW to ADF, it is better to use a tool other than dtc.exe

This function works correctly only if the KF dump is good/correct. it won't output good files from bad dumps.

3. For currently ripped ADFs that could potentially be exposed to this case, it is
recommended to re-convert from RAW->ADF using another tool, or if RAWs were
not taken - re-dump"

I do not agree with this way of proceeding. A track is bad, a dump is bad, the first move to do is to use a floppy cleaner, and examine the disk surface low and high for mould spots. Then clean them before reading the disk again.

Conclusion of all this : you try to make the tool compensate for the bad dumps issued. This is not the correct way to operate.

The good way of doing dumping is this :

1) before reading any disk, check the disk surface with floppy cleaner.
Never put a disk of unknown condition inside a drive !

2) if you find mould or any suspect traces or spot, use alcohol 70% with a precision cotton bud (not those for the ears!) and wet the spot of mould as much as possible in order to remove it without damaging the disk surface
and displace the flux on the disk surface.

3) once the check/clean up is done, you can do the dump

4) if everything went correctly, the dump is fine.

5) export the ADF from any tool you want, there will be no surprises.

[uicr0Bee]

Hi, let me add some details:
1. The disk in question is ABC-SOFT Disk 14 - a German PD disk.
2. The disk surface is clean and shiny. No mould, no dirt.
3. The dumps were made to .raw and .adf formats in one pass, with "-i0" and "-i5" parameters together (KF official manual).
4. stdout and stderr saved to .log and .err files respectively - they are included in the .zip archive uploaded to TheZone by rygar.
5. First dump made on Jan 11, with no errors in .log or .err files
6. On Jan 16, I forgot that the disk was dumped earlier and I did it again. Once again, no errors in .err or .log files. While ejecting the disk from the drive I noticed that it had been marked as dumped before.
7. Just out of curiosity, I compared the Jan 11th and 16th .log files and found them identical, but .adf files were slightly different. I found it strange, so I did several additional dumps with the same method. All dumps from Jan 16 gave identical .adf files.
8. I sent the dumps to Zeewolf (both .raw and .adf files) and he discovered that .adf files converted from Jan-11 .raw dumps using gw.exe, disk-analyse.exe and hxc.exe (2024) are identical to .adf converted from Jan-16 .raw files, which is not the case for hxc.exe (2021) and dtc.exe.

We thought someone may find it interested too and maybe investigate deeper. Also that you cannot tell whether a dump is OK, only based on stdout and stderr outcome, while using "-i0 and -i5" parameters together, or simply "-i5" alone for standard unprotected disks to get .adf only, without .raw (unless you run the same disk a couple of times to compare the .adf files).

That's all. Take care