Edge of Nowhere Oculus Rift mystery file type?

[tyrannojones]

Hey everyone and anyone looking at this! Maybe someone can help me figure this one out...

Edge of Nowhere is a great looking survival horror on the Oculus, based on Lovecraft's "AT THE MOUNTAINS OF MADNESS" and it's been built using the Unreal engine (4 I think), which means I could possibly take a look at the files on the inside. The problem however is that when I try to access the files, they're packed into some kind of unknown file type. The core asset files are two 3gb and one 1.5gb files, but there are two others, much smaller that I'm assuming are in the same file format, which I've attached here.

Anyone able to help?

SAMPLE 1: TOC.???
SAMPLE 2: DAG.???

[aluigi]

Can you list all the toc and dag files?
full list with original names.

[aluigi]

In the meantime the format doesn't seem to have much sense.
The structure of toc and dag is the same, toc doesn't seem to have clear references to dag and dag has a long list of filenames which takes most of the file but is not covered (doesn't have sense to have all these names without files).
The script extracts all the referenced data:

Code: Select all

comtype zlib_noerror    # otherwise it gives error

set SIGN1 binary "\xaf\x12\xaf\x77"
set SIGN2 binary "\xaf\x77\x1f\x89"

getdstring SIGN 4
get SIZE long
if SIGN == SIGN1
    savepos OFFSET
    get ZSIZE asize
    math ZSIZE - OFFSET
elif SIGN == SIGN2
    get ZSIZE long
    savepos OFFSET
else
    print "Error: unknown magic value"
    cleanexit
endif
clog MEMORY_FILE OFFSET ZSIZE SIZE

idstring MEMORY_FILE "1TAD"
callfunction GET_ENTRY 1
math ENTRIES = SIZE
for i = 0 < ENTRIES
    callfunction GET_ENTRY 1
    log "" OFFSET SIZE MEMORY_FILE
next i

startfunction GET_ENTRY
    get DUMMY long MEMORY_FILE
    get OFFSET long MEMORY_FILE
    get SIZE long MEMORY_FILE
endfunction

[tyrannojones]

That's literally all there is in the assets folder, just those two and the 3 much much larger files.

oc1.JPG

oc2.JPG

Here are cuts taken from the first archive using your filecutter.bms
Cut 1
Cut 2

[aluigi]

Ok, this is my solution:
http://aluigi.org/bms/edge_of_nowhere.bms

It should be able to dump and automatically decompress ALL the archived files.
The extracted files are probably just DAT1/1TAD files that are a sort of container of resources, they can be extracted but the extracted data is not so useful and the part which is not extracted contains information about the names and types of the resources... basically it's a bit messy so the script will extract the files "as-is". Optionally you can launch the script on some of the extracted *.1ta files if you want to see what's the output.

"toc" contained all the information while "dag" is totally useless (you can extract the dag file with the script in any case).
For the curious people, the names inside "dag" are NOT referred to the files in g00s00* they are 19000 against over 43000 files so let's close a possible annoying discussion about names vs files before it starts.

Let us know if all the extracted files will have the 1ta extension or there are also other types of files.

[tyrannojones]

Thanks mate! I'll take a gander at things now and see what comes up! It's an interesting format... I don't actually own an oculus myself but've watched videos of this game in particular and wanted to see what I could make of its guts... I might have to see if other OR games are the same as this as far as this kind of packing goes...

[tyrannojones]

Just had a look through and extracted everything -- looks like they're mainly dat1/1tad but there are apparently a small number of shockwave files and WAV files in there as well. But because I'm still kinda learning all this reverse engineering stuff I have no idea what to do with any of this really...

[aluigi]

Maybe try to use the script on one of the biggest *.1ta files just to check what gets extracted from them, maybe there is something useful.

[tyrannojones]

Okay, it just spat a couple of new .dat files at me, but neither script worked on it following that... https://www.sendspace.com/file/zghsxe

[aluigi]

Yeah, that extracted data is pretty useless.

[tyrannojones]

HAHAHAH! I tried a few others and they spat out similar files -- different counts, and sizes, but more .dats for sure.