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

[llabnip]

A port normally tries to give the same experience as the original, so any differences are errors.

I disagree. Many good ports attempt to give an even better experiece than the original. Extra modes, new soundtracks, more characters, hidden ships, artwork, additional difficulty levels, new levels, practice modes, hitbox viewing, auto-fire, etc. are changes that are intended to make a home port have even more features and potentially provide a better overall experience over an arcade original (witness the Saturn mode in Radiant Silvergun). In fact, we often come to expect these extra changes and are disapointed when improvements and new features are not added. Having an option to do away with the slowdown is an improvement to some of us (it certianly is for me, your milage may, in fact, vary). Having the ability to turn WAIT back on for the PSX DDP really shows some extra care on the part of the home port developers (SPS).

[BulletMagnet]

Meant "SPS", right?

Blah, some company with a 3-letter acronym for a name.

So sue me if I'm too lazy to look it up on GameFAQs.

[Recap]

Meant "SPS", right?

Blah, some company with a 3-letter acronym for a name.

They're actually Sharp (its division for the old X68000 software, to be more exact).

[undamned]

What I say about perfect Cave emulation, whether it's possible or not, doesn't matter.

Ok, that's pretty much all I needed to hear.

I'll tell you what, Recap, you've sparked my interest. I'm not in the know as to how the Guru, or whoever does their MAME studies, but if I can be guaranteed that my Dodonpachi PCB goes unharmed, I'm willing to lend it to a responsible party for such studies. I'm just curious what this will lead to in the way of Cave emulation.
-ud

[BulletMagnet]

This is starting to look like the "You don't DIE, you BORDER DOWN!" discussion I got myself tangled up in in another part of this forum, heh heh. In rememberance of that methinks I'd best just shut my yap, heh.

[LUNardei]

What I say about perfect Cave emulation, whether it's possible or not, doesn't matter. I'm not a programmer. I only know (and now, everybody here knows) that there are only some tiny clockspeed differences to get PERFECT Cave games emulation and that it can be fixed if a Cave PCB is send to MAME devs. Why? I asked them. Other people's opinion are *proof* to me depending on *who* they are. MAME devs are pretty reliable concerning these things to me.

This is actually false. I think you've misunderstood Guru's words. I've sopken with a guy of the italian MAME's forum, really involved in MAME's development. Guru has measured the clocks, and the result is...that MAME and pcb has identical hardware speed.
With a pentium 4 1,4 ghz (or maybe less) Cave's emulation is perfect.

Guru want a PCB to fix the problem of the flickering. In Cave's hardware the only thing not emulated is a the DMA driver of the video chip. This thing avoid the sprite limit. So the emulated game can generate infinite sprites without flickering...

The rest, imho, are opinions. Some people says slower, some says faster, but really our eyes can get a difference of...10%?
For me the discussion is over. I want to apologize if a can seem arrogant, but believe me, i cannot express myself perfectly in english..

[Recap]

I'll tell you what, Recap, you've sparked my interest. I'm not in the know as to how the Guru, or whoever does their MAME studies, but if I can be guaranteed that my Dodonpachi PCB goes unharmed, I'm willing to lend it to a responsible party for such studies. I'm just curious what this will lead to in the way of Cave emulation.
-ud

The thing is, you sparked PCB owner, that the PCB's integrity can't be fully guaranteed and that your own satisfaction is not really important. So if that's your prerrogative I don't think Guru wants your precious Do Donpachi PCB. Try it, if you want to, anyways.

This is actually false. I think you've misunderstood Guru's words. I've sopken with a guy of the italian MAME's forum, really involved in MAME's development. Guru has measured the clocks, and the result is...that MAME and pcb has identical hardware speed.
With a pentium 4 1,4 ghz (or maybe less) Cave's emulation is perfect.

Guru want a PCB to fix the problem of the flickering. In Cave's hardware the only thing not emulated is a the DMA driver of the video chip. This thing avoid the sprite limit. So the emulated game can generate infinite sprites without flickering...

Maybe I misunderstood, maybe the problem was not totally known when I asked, since some months have passed. The result is the same, anyhow: there are small emulation deficiencies which can be solved if a Cave PCB is sent to Guru.

[LUNardei]

Ok, this is sure
But it's not a real problem...the video now work in ideal mode, but actually in the PCB the number of sprites is limited, so when too much sprites appear on the screen...The pcb flicker, MAME don't.

[Neon]

No hard feelings, UD, raiden, I hope. I simply want the emulation to be as perfect as possible since the games are on the rare side and not everyone can have an original. Preservation too.

I'm in the market for a PCB version of the near-perfectly-emulated Strikers 2. Idiocy? Perhaps...but I'm enough of a fanboy of the game to want the version absolute closest to perfect.

[raiden]

The thing is, you sparked PCB owner, that the PCB's integrity can't be fully guaranteed and that your own satisfaction is not really important. So if that's your prerrogative I don't think Guru wants your precious Do Donpachi PCB. Try it, if you want to, anyways.

I wonder how you´d treat him if he were the Campaign version owner.

No hard feelings, UD, raiden, I hope. I simply want the emulation to be as perfect as possible since the games are on the rare side and not everyone can have an original. Preservation too.

no prob at all, I´m thinking the same anyway.

The rest, imho, are opinions. Some people says slower, some says faster, but really our eyes can get a difference of...10%?

if you play a game often enough, it becomes extremely obvious and not something you have to search for. Especially if it´s not a constant speed difference, but a situational one.

Seems that indeed it IS a problem with PCB owners' ego here.
(...)
Maybe to show to people like you what a waste is to buy certain PCB's these days?

you´re getting funnier by the minute.

Do you play emulators with a small, not "tated" VGA monitor? Precision is very dependant on your screen's size.

like I said, the phenomenon is common to every game running on PC hardware. But I´d really like to know how screen size influences precision... Maybe that explains your strange conception of the word "perfect".

[rtw]

Interesting thread...

When doing a compare with MAME try to use the newest version, the core of MAME has been thoroughly rewritten lately so anything lower than 0.95 is inaccurate.

The problem with the MAME speed is in the blitter chip. If you go to

http://aarongiles.com/weblog/archives/2 ... blitt.html
http://aarongiles.com/weblog/archives/2 ... _blit.html

You can read all about blitters and their function. Guru needs a PCB so that he can try to determine the exact blitter behaviour.

Esprade in MAME is faster than the PCB and in addition you can see 'objects' which you cannot see on the PCB because the PCB limits the total number of sprites...

Our PCB's will eventually die one day, but emulation is forever

If you compare emulation from 1997 and today it's quite an improvement!

rtw

[highlandcattle]

he recap, since i don't have a xbox yet, how would you rate the video output? (i just bought daisenryaku 7 so i will get one someday)

i'm asking because i have the "pseudo-RGB" cable for my jp gamecube (you know, the one made from a VGA cable) and it's almost as bad as s-video. i heard the x-box video output is the worst of the 3, is that true, or does it depend on the country version?

Err doesnt nintendo sell rgb cables themselves (at least in europe I have one!)

[incognoscente]

European Gamecube RGB cables do not work on NTSC Gamecube systems. To get the same functionality on an NTSC Gamecube, a D-Terminal or Component Video cable must be modified.

[raiden]

a little afterthought: if nobody would buy old PCBs, there would be no emulation of them either, because arcade owners would simply throw them away after their arcade lifespan is over. I used to despise collectors for their "greed", but that was really short-sighted, I´ve come to realize. If nobody kept their games, I wouldn´t be able to buy Megadrive games nowadays. Emulation is the long-run perspective, but as this very thread shows, even after a game is emulated, the PCBs don´t become redundant, no matter how you look at it. So much for "waste of money".

[zaphod]

Repeat after me. SLOWDOW IN ARCADES IS INTENTIONAL.

Often, when developing a game, it's realised that it can't sustain full frame rate all the time. SO, knowing that, the exact times when slowdown will happen is carefully managed. That is what I mean by "it's intentional" the fact that it has to slow down is not, but once that is discovered, it is carefully tuned so that the slowdown and or stopping of it won't cause cheap deaths. Slowdown in arcade hooters is almost ALWAYS predictable, and this is NO accident.

I'm of the "real thing can't be replaced" camp. There's times where the "perfect emulation" sounds off to me (relies on capacitors on the output for sound) or there's too muhc input lag, or the game doesn't behave quite right. This is most important for shmups wich need very tight control to be playable. This is wh i will always prefer a good PORT to an emulation of a shooter, simply because there will be less controller latency.

Still mame is a great thing, and for most games, tit's good enough.

PS: one NON shmup game that is really hurt by emulation is Battletoads. I DEFY you to pass CLinger Wingers on anything but a real nintendo. you won't be able to do it. those milliseconds of imprecision will add up and you get smashed because the control just isn't tight enough on the emulation to beat it.

[undamned]

...Slowdown in arcade hooters...

...tit's good enough.

Where's your mind?
-ud

[system11]

Repeat after me. SLOWDOW IN ARCADES IS INTENTIONAL.

I was thinking that nobody could make claims or arguments quite as ridiculous as Recap has managed. This is why I've stopped bothering to reply since he clearly doesn't want to listen to someone who has probably repaired more PCBs than he's ever played, and owned many more than that. I actually stopped dumping ROMs from my own boards because of people whining about PCB collectors as seen above. Surely the bottom of the barrel was being scraped.

Alas, for I was wrong.

On a few choice games specifically by Cave, they appear to have coded in a purposeful drop in speed to save face when the hardware can't drive enough bullets. I wouldn't call this intentional, more controlled - coping with something rather than designing it out in the first place.

However, the vast majority of slowdown in arcades, is not intentional or controlled at all. It's because the hardware can't cope, and it's not graceful at all.

Big up to Neon for being a voice of reason and using the phrase 'as perfect as possible'. This is a worthy goal, and a realistic view of the world.

[TWITCHDOCTOR]

...Slowdown in arcade hooters...

...tit's good enough.

Where's your mind?
-ud

Thats awesome!!! I would have never noticed those typo's...well maybe, but not today, thats for sure.

[Recap]

Repeat after me. SLOWDOW IN ARCADES IS INTENTIONAL.

I was thinking that nobody could make claims or arguments quite as ridiculous as Recap has managed. This is why I've stopped bothering to reply since he clearly doesn't want to listen to someone who has probably repaired more PCBs than he's ever played, and owned many more than that. I actually stopped dumping ROMs from my own boards because of people whining about PCB collectors as seen above. Surely the bottom of the barrel was being scraped.

Alas, for I was wrong.

I was trying to pass over certain alluding posts here in order to not fuck even a bit more one thread which I think is interesting enough, but seems that some super ego PCB repairers just persist on using insulting or despising words against me without actually adding anything constructive at the end. Bah.

Your experience with PCB's would indeed have been useful here, but you're too prejudiced and you make statements based on dated emulators. Try Cave emulation with today's MAME. Use a good motherboard, CPU, an Arcade VGA and a good controller interface. And form again your opinion.

Perfect emulation is possible for a player's eyes, ears and hands. No matter how expert the player is. Only the correct equipment is needed. That's what I said. What this thread has shown is that all the people claiming that MAME's Cave games emulation runs the games way too slow to the real hardware, or even with true gameplay differentiations, were wrong. They're ALMOST the same today. They might be THE SAME soon.

Relax your ego. For your own good.

[system11]

Your experience with PCB's would indeed have been useful here, but you're too prejudiced and you make statements based on dated emulators. Try Cave emulation with today's MAME. Use a good motherboard, CPU, an Arcade VGA and a good controller interface. And form again your opinion.

Is a Supermicro motherboard and twin 2.8ghz Xeons good enough for you? For input I've tried so far: oldskool PC digital joystick on sound card port, USB PC pad, other USB PC pads (as all drivers are not created equal), and a PS2 pad into a USB adapter (actually the worst response, but the best controller..) The display card is a Radeon 9700 Pro, sound from a Creative SB Live. It is not a slow machine. I've also tried MAME of various versions under Windows and Linux. I've tried out the commercial emulator based multicabs, several varieties at coinop shows and distributor offices. I use MAME (current versions) at the moment to play Donpachi until I can source the PCB. I frequently use MAME to try out games I remember as being ok, before coughing up the cash to buy them. You're making statements based on assumption.

What myself, Cigs, and others have said is the truth - the most obvious fault is that the input just isn't right. Actually since the arcade boards are all flawed anyway, you'll never see all their precise quirks and faults emulated. In fact you often find the same game in slightly different board variants, one of which runs slightly differently. Perfect emulation is a pipedream, in fact the timing can never be exact either, since the odd crystal speeds driving some arcade boards do not fall on precise CPU clock cycle boundaries, being as they are, analogue components. Clock speed aside, PCs in general are a total mess of timings and interrupts, where actually to behave like the real thing, you're after something as simple as a lightswitch for the player pushing left.

For your records, on my machine DDP runs too quickly, and Battle Bakraid (yes, Raizing) is missing the odd stuttering slowdown the PCB suffers, and the bullets don't flicker either. For Guwange, the sound is too raspy, less treble on the PCBs. All of them feel like the controls are wrapped in a sock, compared to the arcade versions. I've owned all of them (I kept Bakraid).

I'd suggest you check your own ego - crusading along a path while ignoring entirely what people are saying. Like a virtual case of LALALALAL I CANT HEAR YOU vented in text.

[Recap]

In fact you often find the same game in slightly different board variants, one of which runs slightly differently.

Just what I said. That's exactly the reason for me to infer that *perfect* emulation is possible.

Or to explain it a bit better, the differences are becoming too unnoticeable to speak about *imperfection*. And I NEVER had input lags, even with USB adaptors, side-by-side with the real hardware and fighting games, which do also suffer a lot from input defiencies, I repeat.

[undamned]

In fact you often find the same game in slightly different board variants, one of which runs slightly differently.

Just what I said. That's exactly the reason for me to infer that *perfect* emulation is possible.

That makes zero sense! Bloodflowers just said that PCB's can be different amongst themselves, so how do you perfectly reproduce that which is inconsistant? How do you make an exact copy of the undefined? Unless your point is that, via emulation, we can make a perfect game, i.e. better that the original PCB? If that is what you are saying, you should start another thread, as your point has changed since the start of this thread.

LALALALAL I CANT HEAR YOU

-ud

[Recap]

In fact you often find the same game in slightly different board variants, one of which runs slightly differently.

Just what I said. That's exactly the reason for me to infer that *perfect* emulation is possible.

That makes zero sense! Bloodflowers just said that PCB's can be different amongst themselves, so how do you perfectly reproduce that which is inconsistant? How do you make an exact copy of the undefined? Unless your point is that, via emulation, we can make a perfect game, i.e. better that the original PCB? If that is what you are saying, you should start another thread, as your point has changed since the start of this thread.
-ud

I just edited my previous post before reading this to explain. PCB's of the same game are not always equal but nobody says any is "imperfect". It's that what I meant with *perfection* - differences are unnoticeable even for the most expert player.

[black mariah]

What myself, Cigs, and others have said is the truth - the most obvious fault is that the input just isn't right.

Agreed. I suck at shmups and readily admit as much. Even I can notice the discrepancies in input performance at times. One tap sends you five pixels over, the next sends you fifteen over. It's incredibly annoying when you die because of something you KNOW you didn't do.

I just edited my previous post before reading this to explain. PCB's of the same game are not always equal but nobody says any is "imperfect". It's that what I meant with *perfection* - differences are unnoticeable even for the most expert player.

Agreed as well. All this bullshit about leaky capacitors makes my ass hurt. I still heartily recommend Metamucil for most visitors to this forum.

[Dave_K.]

I just got my DoDonPachi PCB in from excellentcom today, and right away noticed the timing is completely different from MAME.

The boss battles (and some mid-boss battles) are NOTICEABLY slower on the PCB! At first I thought maybe my board was having problems, but then whenever the amount of enemy bullets lightens up, things speed back up. In fact, I'd almost say that from the begining of the first stage, things seems a tad faster than mame.

Anyway, with the timing all different, its like starting over again, except I'm better at the boss battles now!

I also got ESPRade PCB, but haven't playing that one as much in Mame to make a good comparison.

[gameoverDude]

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.

Just as Smraedis said, I recall older MAME versions being tremendously slower on Cave games.

Look at Giga Wing on Dreamcast- it's got an appreciable speed difference from CPS2.

I think it's best to have the original slowdown ported over from the arcade game as a "WAIT" option. I can live with slowdown, especially if the arcade game was that way. Frame drops are an atrocity on the other hand.

[Venom]

Going back to Dangun, I think there was some speed difference playing in MAME. Then again I was playing the PCB on a 13" commodore and MAME on a 19" flatscreen. I felt is was faster, and in that game a little more speed hurts. The rumor I heard was 54hz pcb vs 58 hz MAME. All I've seen are anecdotes, is there any proof?

[futurespa]

In the latest mame versions, all titles like 1944, Raiden Figthers Jet, DDP etc.. etc.. are exactly the same as the PCB, in other words you own the arcade for free. And that is what stings most of the PCB buyers here and therefor they will try to defend their hobby.

[rtw]

In the latest mame versions, all titles like 1944, Raiden Figthers Jet, DDP etc.. etc.. are exactly the same as the PCB, in other words

Exactly the same ? In MAME 0.96 the music in RFJET is far from perfect!

you own the arcade for free. And that is what stings most of the PCB buyers here and therefor they will try to defend their hobby.

What do you mean stings ? Who do you think does the dumps so that people can enjoy them in emulators ? I dumped the ASIA Metrotainment version of RF2 f.ex

Mind you, I know that the PCB's won't last forever, but emulation will. So it's in my self interest to preserve these gems. And before the flame war starts, I am not comfortable with emulating games which are less than 5 years old.

Now let us get this topic back on it's rails, the issue was how can we supply the Guru with a PCB that he can use to measure the timing in the blitter chip.

rtw

[futurespa]

That's all very true. And the music in RFjet is not correct (yet) indeed.

Furthermore all of the good mame emulations are identical to the PCB.

Nuff said.