Finding the emulation deficiencies - Vol. 1: CAVE games.

[Recap]

Many people believe that emulating an arcade game can't mimic the original PCB at all. Many of us know that, with the correct equipment (that is, with an Arcade VGA card and an RGB monitor, and, of course, a good joystick) it is actually possible, and the result is only dependant on how well coded the emulator is.

So many people play arcade games via MAME and other emulators instead of buying PCB's. That's why we're starting to speak of "MAME versions". Since some games don't have a perfect emulation status, their final behavior is indeed different for the player's eyes and hands.

Cave games are famous for this. MAME devs haven't achieved to reproduce the original hardware speed, hence the games plays differently. MAME devs are aware of this. I recently asked Guru (one of them) about the problem and he said they need a Cave PCB in order to fix the speed in all the Cave hardware-related games (which includes the likes of the Sailor Moon arcade games, Air Gallet or Metamoquester).

This thread has two purposes: Compiling accurately once and for all the emulation deficiencies of Cave's shooting games, since many people here have the PCB's and have compared both "versions" side-by-side. And secondly, checking how many of us are interested in buying a Cave (old) PCB for the MAME team in order to get perfect emulation of these games. Fortunately, there are not other glitches in them, so it wouldn't be a hard task for the driver's author.

Raine also emulates Cave games, so if anyone knows if there are differences with MAME, this is the place to tell.


DONPACHI: ?

DO DONPACHI: MAME acts as an overclocked PCB, hence there are less flickering sprites and slowdowns. Gameplay is affected.

ESPRADE: MAME acts as an overclocked PCB, hence there are less flickering sprites. Gameplay is almost intact (?).

DANGUN FEVERON: MAME runs it faster than the PCB (please, confirm).

GUWANGE: MAME acts as an overclocked PCB, hence there are less flickering sprites and slowdowns. Gameplay is affected.

AIR GALLET (Gazelle): "I can't see any difference between FBA (I don't know about MAME, my PC is too slow) and the PCB itself. Even the slow down at the end of level 1 when the space shuttle takes off looks pretty accurate." -Yojo!

PROGEAR NO ARASHI (CP-S II): "Slows down in MAME in more places than on the real board". Seems that MAME devs agree with this quote. If anyone knows how it is on Kawaks or Nebula compared to MAME, please, let us know.


Please, additions, corrections welcome. Be sure you're using a recent MAME version.

[system11]

Many people believe that emulating an arcade game can't mimic the original PCB at all. Many of us know that, with the correct equipment (that is, with an Arcade VGA card and an RGB monitor, and, of course, a good joystick) it is actually possible, and the result is only dependant on how well coded the emulator is.

No, it's not possible, sorry. Here's a few examples just to get started with:

Lag between game logic and soundboard. There's a delay.

Sampling of analogue sound.

Some games depending on quirks of the physical hardware - example Zaxxon, these -rely- on low grade capacitors leaking power to sound right.

Input lag. Your parallel port or USB input passing through an OS level driver is not as fast to react as a pair of TTL chips hooked right up to the CPU.

Emulation can be good, and it can be very acceptable as a substitute for racks of PCBs. It'll never be perfect though, and I'd say less than half a percent of people who say 'OK' on the MAME startup screen, are actually legally able to do so.

[Smraedis]

all of those listed were too slow in early mame versions, but have gone a little too quicker in newer versions, apart from Progear as it wasn't released back then, and its on different hardware too.

[neorichieb1971]

The biggest difference is price and space.

I mean lets face it, if your going to get an RGB monitor, a proper joystick and a VGA card or whatever, its almost as much pain as getting the real thing.

[Recap]

No, it's not possible, sorry. Here's a few examples just to get started with:

Lag between game logic and soundboard. There's a delay.

Sampling of analogue sound.

Some games depending on quirks of the physical hardware - example Zaxxon, these -rely- on low grade capacitors leaking power to sound right.

Input lag. Your parallel port or USB input passing through an OS level driver is not as fast to react as a pair of TTL chips hooked right up to the CPU.

I actually have made some tests side-by-side with some Neo-Geo games, for instance. There are no sound differences at all, much less input lags even with USB controllers (you don't need to use USB, thoe). Samples are used only by very very few games, on the other hand. The point is that perfect emulation of most games is possible, as old Cave games.

Emulation can be good, and it can be very acceptable as a substitute for racks of PCBs. It'll never be perfect though, and I'd say less than half a percent of people who say 'OK' on the MAME startup screen, are actually legally able to do so.

Sure. That would bring us to ask about which percentage of people here owning PCB's are actually legally able to do so. Coin-operated video-games are only legit if you have a licence, at least in my country. On the other hand, 2nd-hand Japanese games (and not only PCB's) are not legit, either. So yep, we can ask about legalities, if you want.

I mean lets face it, if your going to get an RGB monitor, a proper joystick and a VGA card or whatever, its almost as much pain as getting the real thing.

My RGB monitor is also my 29'' RGB TV, and my custom joysticks can be used with consoles, too. The "pain of the real thing" is media itself - hard to store, hard to keep it working.

[system11]

I actually have made some tests side-by-side with some Neo-Geo games, for instance. There are no sound differences at all, much less input lags even with USB controllers (you don't need to use USB, thoe). Samples are used only by very very few games, on the other hand. The point is that perfect emulation of most games is possible, as old Cave games.

No, it's not possible. Get over it. Emulation is not exactly the same as the real thing, it never will be until someone has a magnificently powerful supercomputer that emulates each TTL chip individually with precise timing. Even TTL chips by different makers have different performance too. Digital approximation of analogue sound (example, faking a tone generator) is not exactly like the real thing either. Every damn part of it is an approximation. Even the clock timing is an approximation, some of the clock speeds used by PCBs were .. interesting to say the least.

Sure. That would bring us to ask about which percentage of people here owning PCB's are actually legally able to do so. Coin-operated video-games are only legit if you have a licence, at least in my country. On the other hand, 2nd-hand Japanese games (and not only PCB's) are not legit, either. So yep, we can ask about legalities, if you want.

I was only having a little fun with that line, don't be so defensive. I personally used to use emulation as a way to test broken rom images and other quirks against a possibly or known faulty PCB, I used to repair quite a lot of boards in fact. Just don't claim emulation can be perfect. It can be very good emulation, certainly good enough to be playable and look the same - nobody disputes this. Claims of perfection are the deranged ramblings of emukiddies.

[TWITCHDOCTOR]

Kawa-X runs Neo Geo titles on my Xbox VERY well. I can't honestly say I could tell the difference. I've played Nates Neo cab with Strikers 45 PLus(maybe a slight difference here) but none that I could notice with Blazing Star and Pulstar. (Pulstar has a minor sound issue when the score is being counted at the end of level, but thats it)

[BulletMagnet]

Air Gallet? Is that a Cave game?

In addition, I remember people saying that Dangun Fever On ran faster in MAME than the PCB does, though I can't confirm that myself.

[sffan]

Just don't claim emulation can be perfect.

It can be, in the future I'm sure. So don't claim emulation CAN'T be perfect. Nya nya.

[incognoscente]

Air Gallet is by Gazelle, released by Banpresto. It runs on Cave hardware (the hardware that powered Donpachi-Guwange, not the modified PGM, etc. hardware Cave has used since).

[Recap]

I was only having a little fun with that line, don't be so defensive.

Not really defensive; I actually buy everything I play in brand new condition if it's released for the home market. Just answering to a misplaced "line".

I personally used to use emulation as a way to test broken rom images and other quirks against a possibly or known faulty PCB, I used to repair quite a lot of boards in fact. Just don't claim emulation can be perfect. It can be very good emulation, certainly good enough to be playable and look the same - nobody disputes this. Claims of perfection are the deranged ramblings of emukiddies.

In this case, claims of perfection comes from someone who has made side-by-side tests with real hardware, but if you want, I can ask to the MAME devs and let you know the thoughts of these "emukiddies". They'll probably say that, indeed, you also can find speed differences between PCB's of the same game. So yep, PERFECTION is, most likely, impossible. Emulation without noticeable differences for even the most expert player IS possible and is a reality today.

Or, if you prefer, I can say it this way: ANY arcade to console port has more behaviour differences than a big part of today's emulated games. Even the ones from Naomi to DC and other analogies.

Kawa-X runs Neo Geo titles on my Xbox VERY well.

I'm sorry but nope. The Xbox can't output true low resolutions. That's a big difference from the very beginning.

Air Gallet? Is that a Cave game?

Believe me; almost.

In addition, I remember people saying that Dangun Fever On ran faster in MAME than the PCB does, though I can't confirm that myself.

Thanks. I'll update with this. Do you know if it's "constant" faster speed or just less slowdowns?

[Neon]

If it's still not resolved by then, I'd be willing to purchase a Cave PCB and send it to the MAME team sometime this summer (when I have a job and money). Whichever is cheapest; I'm assuming Donpachi would be.

Can you provide specifics? What address, whom should I email/irc/aim/whatever

[Recap]

If it's still not resolved by then, I'd be willing to purchase a Cave PCB and send it to the MAME team sometime this summer (when I have a job and money). Whichever is cheapest; I'm assuming Donpachi would be.

Can you provide specifics? What address, whom should I email/irc/aim/whatever

Guru: http://unemulated.emuunlim.com/index.html

But you don't need to purchase it alone. I made this thread to see how many of us want perfect Cave emulation and want to contribute. Give this thread a few weeks.

[BulletMagnet]

Do you know if it's "constant" faster speed or just less slowdowns?

I believe it's "constant," as I think someone said it ran 20 percent faster than the arcade game, or something like that...I can't confirm that though, as I'm only going off what I (think I) remember someone saying.

[LUNardei]

DO DONPACHI: MAME runs it too slow (about 80% of the original speed?).

80% of the original speed? No. It's IMPOSSIBLE. I've seen it with with my eyes. 80% it's a noticeable difference, don't you think?

[LRa]

If it's still not resolved by then, I'd be willing to purchase a Cave PCB and send it to the MAME team sometime this summer (when I have a job and money). Whichever is cheapest; I'm assuming Donpachi would be.

Can you provide specifics? What address, whom should I email/irc/aim/whatever

Guru: http://unemulated.emuunlim.com/index.html

But you don't need to purchase it alone. I made this thread to see how many of us want perfect Cave emulation and want to contribute. Give this thread a few weeks.

Yeah, but if he purchases it alone, the board will eventually come back to him.

[Recap]

DO DONPACHI: MAME runs it too slow (about 80% of the original speed?).

80% of the original speed? No. It's IMPOSSIBLE. I've seen it with with my eyes. 80% it's a noticeable difference, don't you think?

It's noticeable but it seems that the driver's author of MAME team didn't have a chance to check the game in its original form when he added the game. I don't really know if it's 80%; I've just stolen the info, that's why I put a question mark.

Yeah, but if he purchases it alone, the board will eventually come back to him.

Alrighty, then, but I fear that the interest in collaborating may fade out if someone claims that he'll do it alone. What if he cannot purchase it at the end?

[Neon]

Alrighty, then, but I fear that the interest in collaborating may fade out if someone claims that he'll do it alone. What if he cannot purchase it at the end?

Understandable. However I should have about $170 in a week with the money I now have and the money CMoon is sending for Saturn SFZ3. So if someone can point me to a site or private seller that has a Cave PCB in this price range, including shipping, we can do it sooner rather than later.

[yojo!]

Air Gallet is by Gazelle, released by Banpresto. It runs on Cave hardware (the hardware that powered Donpachi-Guwange, not the modified PGM, etc. hardware Cave has used since).

I used to play it all the time via FBA , which made me buy the PCB. Quite frankly I can't see any difference between FBA (I don't know about MAME, my PC is too slow) and the PCB itself. Even the slow down at the end of level 1 when the space shuttle takes off looks pretty accurate.

[Recap]

However I should have about $170 in a week with the money I now have and the money CMoon is sending for Saturn SFZ3. So if someone can point me to a site or private seller that has a Cave PCB in this price range, including shipping, we can do it sooner rather than later.

I'll start to look for a cheap PCB and I'll ask Guru again in MAME boards to be sure they still need it. Your game preferences?

[Neon]

I'd prefer DDP or Dangun, but any of them will be fine as long as they get the job done.

[Recap]

I used to play it all the time via FBA , which made me buy the PCB. Quite frankly I can't see any difference between FBA (I don't know about MAME, my PC is too slow) and the PCB itself. Even the slow down at the end of level 1 when the space shuttle takes off looks pretty accurate.

Updated, thanks.

[BR1]

I completely agree with LUNardei.
We have played Dodonpachi both on PCB and on MAME: well we haven't noticed any difference!
IF there's a difference it isn't of the 20%, FOR SURE.
I don't know who could give you such a wrong piece of information....

[Recap]

I completely agree with LUNardei.
We have played Dodonpachi both on PCB and on MAME: well we haven't noticed any difference!
IF there's a difference it isn't of the 20%, FOR SURE.
I don't know who could give you such a wrong piece of information....

Well. This is getting interesting. That info is still "given" in these boards, so I hope the people claiming that appear here sooner or later. It's possible that they used a pretty old MAME version. I'll await for their opinion.

[BR1]

I completely agree with LUNardei.
We have played Dodonpachi both on PCB and on MAME: well we haven't noticed any difference!
IF there's a difference it isn't of the 20%, FOR SURE.
I don't know who could give you such a wrong piece of information....

Well. This is getting interesting. That info is still "given" in these boards, so I hope the people claiming that appear here sooner or later. It's possible that they used a pretty old MAME version. I'll await for their opinion.

This is probable: older MAME versions had some problems with Cave games.

[Smraedis]

it was slower in old versions of mame, didn't anyone read my post up there?

[tehkao]

Even if it is 10% or 20% slower on MAME, if one has never played the original version and can't tell the difference anyway, who the heck cares?

In the end it only matters to the diehard hardcore elitists/purists.

[it290]

Well, it matters for score purposes. I thought I was doing pretty well at improving my DDP skills playing in MAME, until I started playing the Saturn version again and got my ass handed to me.. just for comparison, my MAME high score is around 25 million, whereas on the Saturn I haven't broken 15 mil yet. BTW, I was using the latest versions of both xmame and mame32, so I don't think it's just an issue with older versions, unless the PCB is also noticably slower than the Saturn version.

[yojo!]

BTW, I was using the latest versions of both xmame and mame32, so I don't think it's just an issue with older versions, unless the PCB is also noticably slower than the Saturn version.

It may be due to the refresh rate of the PCB versus the console version. DDP Pcb displays at 58fps and the saturn more likely runs at 60fps. I noticed the same thing with R-Type PCB vs PSX. the PCB runs at 55fps making it easier to dodge bullet in level 7 & 8. When I started playing the PSX I noticed the difference in speed right away and got my ass kicked.

[Recap]

it was slower in old versions of mame, didn't anyone read my post up there?

Yep, but you don't say how much slower. The question is if they're still noticable slower than the PCB.


I'm going to update with BR1's comment until someone says the opposite, but before I'd like to know if he has played the PCB enough.