redump.info Forum

About the timestamps: https://www.epochconverter.com/ldap

DiscImageCreator also decodes the filestamps from DMI and SS, see outputScsiCmdLogforDVD.cpp
and from XDVDFS: execScsiCmdforFileSystem.cpp

Example from https://redump.info/disc/6024/:

DMI:
2005/03/30 20:54:46

SS:
Timestamp of authoring: 2005/03/29 21:07:14
Timestamp of unknown: 2005/03/29 18:38:08

volDesc:
Image creation time: 2005/03/30 02:38:16


I guess it's strange that the image creation date is later than the other timestamps. But this would give a time range of ~8 hours that we could try for brute forcing (if this is even possible).

@LedZeppelin68 are you still planning to investigate the second seed type?

first type seed is only 32bit, 256^4 combinations

second type seed is 128bit, 256^16 combinations, impossible to bruteforce, forever

Ok but my questions about the timestamp? Is there no way to predict the seed?

Jackal wrote:Ok but my questions about the timestamp? Is there no way to predict the seed?

i can say - no

the seed value for second type is 128 bit, which consists of two 64 values:
1: system time and date (we can reduce this value to year, or maybe a month) +
2: and time from last reset (2^64 combos)

it is a very huge number to bruteforce

I think we must save, what can be saved, and store al possible info

I remade xbox_rnd tool, now it doesn't remove the junk from ISO, but only brutes the seed of first type
saves it to log, and it works much faster

P.S. drag all ISOS on the EXE and wait

Did you make any new research on this, Led?

diego-rbb-93 wrote:Did you make any new research on this, Led?

LedZeppelin68 wrote:first type seed is only 32bit, 256^4 combinations

second type seed is 128bit, 256^16 combinations, impossible to bruteforce, forever

So, the tool is useless for most discs.

diego-rbb-93 wrote:Did you make any new research on this, Led?

I would continue to research type 2 algo, maybe some statistical analysis will help

for now, all i can offer to save space
1) for a little amount of isos with type 1 seed (maybe few hundreds) - xbox shrinker
2) for big games with regional clones, which data occupies most of the iso - use merger

Sorry, this is all i can offer

Any chance you would consider open sourcing this?

Thanks for your work.