how to extract files from Sega Naomi gd rom by hb-x3 at 8:10 PM EDT on June 6, 2021
hi, im trying to extract sega naomi gd-rom verion of Capcom vs Snk 2. i am left with decypted .DAT file. i dont know how to open and extract its contents.
GD images are not too (structurally) different from CD/DVD ones. The .DAT file could be the GD image, except not directly mountable with disc imaging software, as block/sector sizes are not the same as aforementioned CD/DVD's. You may get some insight by examining .DAT file with hex editor and trying to figure contents. Data is segmented by 1 KiB-size "sectors" iirc (not sure).
hi, almendaz, thanks for the reply. I already extracted this gdrom. Its gives me the following three files : BDV.BIN , NAOMIGD.BIN , SNKGD_SL.BIN
SNKGD_SL.BIN is the main binary, and is 152 mb, so that's where i think all game files are located. im trying to unpack this file. So far I cant find anything to do so. Also, I decrypted the SNKGD_SL.BIN and ended with .dat file, which is the same size.
No need to bump, I answer on a daily basis, where possible - bump only after a week of silence :) Probably one of the small files or even the main .DAT could contain data table/TOC/something which could help with unpacking the decrypted .DAT file. Not guaranteed the presence of said "TOC" for any VFS. Worst case scenario, would have to inspect decrypted .DAT file and try to split it into recognizable "blobs"/chunks of data, so as to figure its contents. Look for some 4-byte "magic" text, offset values (L-Endian I think), RGB patters, (AD)PCM-like bytes or the like in .DAT usig hex editor. Otherwise, either "TOC" could be (moreover) encrypted/compressed, or there is no "TOC" & resort to manual splitting.
Hi, almendaz, thanks again for replying. So I opened the . dat file in a hex editor and there was a text string that caught my eye which was called Capcom vs Snk 2 USA. This is what I was looking to extract. Im a total noob, when i comes to this. What would you suggest I could do next ?
The found text could be indicative of unencrypted/uncompressed data. Try to find more text in the .dat file. There is no straight way to extract contents unless splitting the data manually & spending time. Otherwise, upload the .dat file and we'll take a look.
Hello - I could reserve a bit of time today. Some GD-ROM images have VFS attached, I can't tell why your upload does not have one, maybe non-dump or altered image packing (i.e. non .gdi), or just different revisions. I'm not into CD dumping methodology/scene. Sorry that I couldn't help you with your file. It's from GDL-0008. Unfortunately it's too much hassle without the TOC - as if TOC was really in another file (or ROM if guiding from dumpers terminology). So, I searched instead for another image dump* for the same game, T1 249 M, non "GDL" supposedly, and extracted its contents. Pswd is "w4wcvs2#df2k" (without quotes). Here Curiously this extract has more data that what the size of SNKGD_SL.BIN would tell. Hopefully they have the common files.
HI, almendaz. Would it help if I post .gdi files package from gdl-0008 ? The gdi is the Toc, which can be mounted in gdi explorer. or would you prefer a .cue file. I think the file you posted is for dreamcast verison of CVS2, not the naomi verison.
It could be of help, but I doubt it would. .gdi files are tiny "cue" sheets with low/high density distribution of sectors. By "TOC" (if I misunderstood the term) I meant some sort of file list, with offsets/filesizes included - this would have been the easiest way to extract the files. If you see the first part of the "track 3" (the high density part), you'll see the filenames corresponding to the extractable data that you have already found before (BDV.BIN, NAOMIGD.BIN, SNKGD_SL.BIN<encrypted>). But the ROM (*redump*) I found has a bigger file list, and gdi explorer made it pretty easy to extract the sub-files, it just took a manually edited .gdi file to include its "track 3" data, with bigger size to boot. As DATA is mainly ISO-extractable ("mode1"), I had no special constraints to create/borrow a .gdi text file. I got a hold of the gdl-0008 data from the web, and after decrypting process checksums are concordant with those of your files, so it would not make a difference.
<< I think the file you posted is for dreamcast verison of CVS2 >> Up until now, the only common file (subfile for 0008) for both naomi/dc versions is the AICA***.bin rom. But making parallel observations, it could be possible to associate files from the dreamcast version, with those from the naomi version, and that way subfiles can be extractable (and many would be identical between versions).
Since the naomi version of cvs2 boots on flycast, Could I use a debugger version of the flycast emulator To dump the contents of the SNKGD_SL.dat file Some how ?
I do not know about flycast or emulator debugger, but from experience you would not get filenames, maybe byte blobs that represent assets/resources. For all of the subfiles, a great work would take place.
I did a quick study on the DAT files (a MANUAL study, so it's not truly quick!). In order to compare the possible file structure of both game images, I converted "track 03" .bin to .iso from "dreamcast" image (bin2iso used), and compared byte chunks with decrypted "SNKGD_SL2.DAT" from "naomi" image. Let's call both of these files as "iso" and "dat" for sake of simplicity. I recurred to GDI explorer to get the LBA's from the H.D. section in "track 03" using this customized .gdi text: ### 3 1 0 4 2352 T01.bin 0 2 450 0 2352 T02.raw 0 3 45000 4 2352 T03.bin 0 ### (Tracks were renamed from "Track ##" to simply T##; AUDIO file renamed from .bin to .raw just like in the original .gdi obtained with chdman with gdl-0008's CHD). I selected portions of "dat" file and compared presence in "iso", and took note of their offsets. As "track 03" starts at LBA 45000, then the LBA listed for each file in GDI explorer is calculated: [LBA in "GDI explorer"] := [(0x)offset found in "iso"]/0x800 + 45000, and thus the filenames were found. With these findings I made a table that correlates bytes between "dreamcast" and "naomi" images.
----------------------------------------------------------------------- filename @ offsets(0x#) [iso] vs. [dat] : filesize(0x#) ----------------------------------------------------------------------- AICADRV.BIN @ [35FEC000] vs. [800000] : F7C8 //contained within 0x800 bytes CD-block (DC.ver) FONT_E.BIN @ [3601E000] vs. [80F7E0] : 20020 ????? @ [3601E000] vs. [82F800](copy?): 20020(?) NOT-SAME FILES? DEBUG.BIN @ [35FFB800] vs. [84F820] : 1940 S20RM04.BIN @ [3603E800] vs. [851160] : DD200
? 377AA81C --- 965380 ? 37774C1C ---
MESJ_WIN.BIN @ [37773800] vs. [92E360] : 57600 || 57604(?) NOT-SAME FILES MESJ_DM.BIN @ [3776E800] vs. [985980] : 4D84 MESE_WIN.BIN @ [37768000] vs. [98A720] : 644C MESE_DM.BIN @ [37763800] vs. [990B80] : 4294 SPR_TTL.BIN @ [362B9800] vs. [994E20] : 253C SPR_TTLE.BIN @ [362BC000] vs. [997360] : 25B4 EFCTPOL.BIN @ [37707000] vs. [999920] : 115B0 EFCTTEX.BIN @ [37718800] vs. [9AAEE0] : 4AE98 DM00FAC.BIN @ [368AC800] vs. [9F5D80] : 34380 ... @ [--------] vs. [A2A100] : ---- -----------------------------------------------------------------------
I interrupted this tedious process. Then, I figured if these values are found in the "dat" file, and I was right. {offset}{size} values pairs were found at [0x14A5A4], and end at [0x14B12C]. 369 registers of 0x4 + 0x4 bytes each. For that chunk of data, a script could be made to extract "subfiles" equivalent to those from the "dreamcast" image, but unfortunately without filenames; then after trimming last "FF FF FF FF" bytes from resulting files, they would be compared against the files extracted from the "dreamcast" image, so as to get "filenames".
Here is the unpacked .DAT in case you're interested. Extracted them using VGM_toolbox. I renamed the files to their corresponding DC version; the "---<name>.BIN" files are aproximately equal to the "<name>.BIN" from DC version; the "_[###]" files are SPSD audio data for naomi, while they are ADX in the DC version.
Hi, almendaz. thanks for your help. I want convert the dreamcast JAP version of CVS2 into English. When I tried to merge your renamed files into the dreamcast version it froze.
Hi, almendaz. thanks for your help. I want convert the dreamcast JAP version of CVS2 into English, using the USA naomi version. When I tried to merge your renamed files into the dreamcast version it froze.
I do not think you can simply swap files between versions, first one would have to understand the data format for the files containing the dialogs/text for both consoles.