Binary multiple compares

[TVTED]

Hi
I'm looking for software to compare three or more {potentially corrupted} binary files, and find the similarities in the majority of the selected files, discarding the differences, and hopefully combining the results into a corrected file. As it is, I'm doing firmware forensics manually, and this is very time consuming for huge {> 2GB} files.
Anyone?

 

[COLGeek]

Where did these files come from?

This doesn't sound like any sort of application you are likely to find. What are you trying to achieve?

 

[little_me]

well, seemingly someone else has been looking for something similar too:
http://stackoverflow.com/questions/4559677/tool-for-3-way-binary-hex-file-comparison
The discussion mentions vim, which works on nix and according to post, also has gui version and windows versions.

But... I'm pretty sure that you know better what features you are looking for than I do, I don't do file comparing or merging for work.

 

[TVTED]

Where did these files come from?

This doesn't sound like any sort of application you are likely to find. What are you trying to achieve?

Forensic recovery of damaged firmware dumps.

 

[COLGeek]

So, you want to compare the hex dumps and then attempt to re-build a working load from the remnants?

 

[Hyboria]

Have you had a look at kdiff3

 

[TVTED]

So, you want to compare the hex dumps and then attempt to re-build a working load from the remnants?

OK
I'll be more specific. Say I have four EEPROM dumps, form the same unit, each uniquely corrupted. It stands to reason, any data in common with at least two of the dumps is probably good. So, if I could find a way to automate the comparison, and output a consolidated file based on the likelihood of good data, it would sure make me and a few friends really happy.

 

[COLGeek]

EEPROM dumps from what device(s)?

 

[Alabalcho]

Ask someone to write you a utility for that. Input parameter - three text files (with hex dumps), output file - another text file. Read a line from the three files, compare, pick the line present in at least two files. Half an hour max.

And then - boom, the three lines are all different, and you are back on square one.