Trickiest games to emulate?

Stef · Post by **Stef** » Sat Nov 24, 2012 11:39 pm

Ex-Mutants
TimeKiller
SeaQuest
I got troubles with these 2 games, due to good V Int flag emulation or timing...

Gargoyles requires good TAS instruction emulation.
On Sega Genesis 1 and 2, the 68k can't lock bus so the instruction do not set the semaphore, this game require emulation of that feature.

Nemesis · Post by **Nemesis** » Sun Nov 25, 2012 12:30 am

I just remembered I've had trouble with ultimate mortal kombat 3, with the AI characters just sitting there and blocking instead of fighting, and issues with the wrong game mode being entered from the menu. Anyone know what that might be related to? It's come and gone a few times while I've been developing my emulator, and I've got the same problems again now, but it's a hard one to debug.

Huge · Post by **Huge** » Sun Nov 25, 2012 5:21 pm

Nemesis wrote:Really? I've never heard of that. I'm interested in doing tests on that to figure out what could cause it. Where does it lock up, partway through the level? Do you know if the lockup was occurring on a US or European system? Was there a particular ROM you were using? I can't run the game right now but when I'm back at my computer I'll give it a go. I own the cartridge of this game so I can easily test it on the hardware.

It's at the difficulty where all players are "Officer xy", at which point the game is nearly unplayably hard anyway. This is the level beyond the last 5th level, and if you check the level # under options it just says "ddddd", and if you get wrecked or win a course, the game (sometimes? always?) locks up on the cutscene.

Have some savestates:
http://www.sendspace.com/file/mq0gmx

One of them has you finishing the last course of Level 5 at 1st place, you can still brake and then finish at a later place so it doesn't skip to the next glitched level (write down the password under options and use them on a real cart, then finish level 5 manually and check the glitched level). The other two save seems to be corrupted, I dunno why. I took these many years ago back on the gens/kega forums, and I have a bunch of older states dating back to 2001 with this same glitched levels.

I'd be interested to know what causes this, and if there is anything past the glitched level in Road Rash II. It always bugged me how it crashed on the last levels.

Eke · Post by **Eke** » Sun Nov 25, 2012 7:27 pm

Actually it crashes on real hardware too, i just tested it with an everdrive and game genie codes to access the 6th level (tested on PAL MD2 switched to 60hz).

It appears to be a "secret" level so i wouldn't be surprised if it was simply a game bug and there is no real end.

PS: I couldn't see any DDDDD near the level #, just a few weird signs which i assume are supposed to represent bikers (?)
See screenshot below taken from my emulator, which display (and crash) the same

Huge · Post by **Huge** » Sun Nov 25, 2012 8:54 pm

Odd how the game doesn't actually end. I wonder how the heck they intended that to work.

I got dddddd on Kega Fusion, using... 7 year old savestates (but I found a few 11 year old ones on my hdd as well, which oddly enough still work). Obviously that's not the most accurate. But I was wondering about the source of that lockup since at least that long.

panzeroceania · Post by **panzeroceania** » Wed Dec 26, 2012 6:32 pm

it's not THE hardest, but getting all the sprites to render properly in Thunder Force IV can be a challenge for some emulators.

Charles MacDonald · Post by **Charles MacDonald** » Wed Dec 26, 2012 9:58 pm

panzeroceania wrote:it's not THE hardest, but getting all the sprites to render properly in Thunder Force IV can be a challenge for some emulators.

That reminded me, Contra Hard Corps and the Thunder Force IV intro use VRAM fill in a way that is good for testing an emulator's implementation.

Also Aleste is one of the very few games that uses VRAM to VRAM copy.

I can't recall the name, but there's one title that accidentally switches into mode 4 during initialization, sets up some registers beyond $0A (which should be ignored) and then switches back. If you don't emulate how the higher registers can't be accessed in mode 4, the game has a corrupt "SEGA" logo and intro screen. Maybe somebody remembers.

I think it may have the dubious honor of being the only licensed game that uses Mode 4, if only for a few milliseconds.

Slightly off topic but I think the _easiest_ game to emulate is Hard Drivin. It doesn't even use DMA and pretty much all the graphics are from the background layer with minimal sprite use. Maybe the only quirk is that it runs in user mode instead of supervisor but that's not a big issue.

Stef · Post by **Stef** » Wed Dec 26, 2012 10:11 pm

James Pond 3 (if i remember correctly) also used VRAM fill for the platform texture... got a bug in Gens with that for sometime.

Eke · Post by **Eke** » Sun Dec 30, 2012 9:05 pm

Charles MacDonald wrote:
I can't recall the name, but there's one title that accidentally switches into mode 4 during initialization, sets up some registers beyond $0A (which should be ignored) and then switches back. If you don't emulate how the higher registers can't be accessed in mode 4, the game has a corrupt "SEGA" logo and intro screen. Maybe somebody remembers.

I think it may have the dubious honor of being the only licensed game that uses Mode 4, if only for a few milliseconds.

yes, i forgot about this one, i found it some years ago while debugging games
and there are actually two of them doing this accidentally (both enable mode 4 then change the screen width, which has no effects on real VDP but screws the display on emulators): "Bass Master Classics Pro Edition" (screwed logo) and one of the two revisions of "Captain Planet & the Avengers" (screwed in-game graphics)

Eke · Post by **Eke** » Sat Jan 19, 2013 5:56 pm

"Sunset Riders" does not display its logo effect correctly if you don't emulate the rarely used HVC latch bit (there is a thread about that one somewhere on the forum).

Another game relying on VDP FIFO emulation (VRAM writes delay) is "Sol-Deace", which will display the spinning earth intro with a black bar if VRAM writes are not accurately delayed.

I also found out recently that the game "Clue" probably relies on accurate VRAM write timings as well: just before displaying the difficulty level screen, it will clear a VRAM portion during active display by using normal VRAM writes from CPU and somehow get interrupted by VINT, which itself writes VDP registers and screws up later VRAM access (code/address register being resetted).

Depending on how much VRAM writes were done before the interrupt occurs, the difficulty selection screen got partially corrupted with ASCII chars. Happens in both Regen & Genesis Plus GX, not in Kega. Off course, it works fine in Gens or any emulator that does not emulate VDP FIFO and processes all VRAM writes without delay, which is why I think it needs very accurate VRAM timings emulation... although I am still not 100% sure (still investigating)