First of all thanks for your help. I checked the drive built into my notebook and found out that it's capable of reading all sectors in the first audio track's pregap, effectively returning the first sector full with scrambled bytes and the second one scrambled partly only. Determining the correct factory write offset shouldn't be any problem now. I will submit the results in the next days when I have some more spare time, before that I'll cross-check everything again as I didn't use that drive for any dumping purposes before.
Regarding your suggestion of posting a .sub file created by CloneCD I have an interesting story to tell...
In my opinion (feel free to convince me of being wrong if that's the case) CloneCD can't read more information from a CD than CDTool can, it's all about interpreting the results which come from the drive. Manic Karts uses Sony style pregaps which start at (-)XX.XX and end at (-)00.00 in contrast to Philips style pregaps which end at (-)00.01. The first sector returning pregap subchannel data for Track 2 contains 02.73, the last one 00.00, so it's obvious that pregap length is 02.74.
There exist subchannel data read offsets (READ CD on sector X returns subchannel data from sector X+Y, where Y is the offset) and when using my primary drive it has a subchannel read offset of +2 on Mode 1 sectors which turns into +1 when entering Audio sectors, effectively returning the very same subchannel data for two subsequent sectors. The drive in my notebook behaves the same but instead has +1 on Mode 1 and zero on Audio sectors.
The last time I tested CloneCD it was unable to compensate that: The subchannel data which is returned twice was kept in the .sub file and as CloneCD entered the first sector of Track 2 it decided to realign subchannel data, effectively dropping the last pregap sector's subchannel data.
So in effect the .sub file would contain the following relative positions in Q channel for the pregap:
02.73, 02.72, 02.72, 02.71 ... 00.03, 00.02, 00.01
Just for helping remeber, my drive returns this data:
02.73, 02.72, 02.72, 02.71 ... 00.03, 00.02, 00.01, 00.00
The problem here is that there isn't a unique method to both interpret .sub file contents and drive responses correctly. Counting number of sectors which contain pregap subchannel data leads to 03.00 for my drive and 02.74 for .sub file. Measuring the difference between the minimum and maximum relative position in Q channel data leads to 02.74 for my drive and 02.73 for the .sub file.
I've posted some more details on CDFreaks:
http://club.cdfreaks.com/f52/clonecd-en … ds-245875/
Anyway, I hope that I could (at least a little bit) convince you of CDTool used by a discreet person is really sufficient for securely detecting pregap lengths (and style, which is unluckily not preserved in a CUE sheet file btw). 