Shmupmame 4.2: Lagless Mars Matrix, Strikers 1945, Galaga 88

[TVG]

Thanks, just wanted to make sure.

I tried 58hz and wow, silk smooth, it's awesome. The microstutter is about completely gone. Tried 57 and 59 just in case but it was not as good.
There's still tearing tho, a regular line that scans the screen about every two seconds but I think I can live with it. I don't think there's a way around that without input lag.

Actually scratch that, I get best results in all cave games (including donpachi) in 59hz. I know it's a bit OT but goddamn I'm overjoyed.

[TVG]

double post

[Lance Boyle]

Has anyone ever tried taking, say, four different versions of MAME, one of which is lagless, and then doing a blind test to see if they could actually tell which one out of the four has reduced input lag WITHOUT looking for background desynch?

I think a placebo effect is in place here.

[captpain]

Has anyone ever tried taking, say, four different versions of MAME, one of which is lagless, and then doing a blind test to see if they could actually tell which one out of the four has reduced input lag WITHOUT looking for background desynch?

I think a placebo effect is in place here.

It's actually a completely measurable effect. Take Batrider outside of the current version, and do a pause-frame advance test. Then, try it in Lagless v3. It can't be placebo because it's absolutely happening.

I'm confident I could tell you which version is which with no confusion at all. If you think there's a placebo effect then you need to Learn 2 Play, Noob.

[Lance Boyle]

It's actually a completely measurable effect. Take Batrider outside of the current version, and do a pause-frame advance test. Then, try it in Lagless v3.

I know about this. I'm just wondering if all the people screaming OMG IT'S SO MUCH MORE RESPONSIVE!! have actually tried taking a blind test.

I'd do it myself but I'm not sure how to arrange it. Probably with a .bat file or something.

[captpain]

It's actually a completely measurable effect. Take Batrider outside of the current version, and do a pause-frame advance test. Then, try it in Lagless v3.

I know about this. I'm just wondering if all the people screaming OMG IT'S SO MUCH MORE RESPONSIVE!! have actually tried taking a blind test.

I'd do it myself but I'm not sure how to arrange it. Probably with a .bat file or something.

I don't know what you're looking to prove. It's a pretty blatant difference, and if you suspect 'placebo' then I can guarantee you the problem is with your perception/reflexes.

edit: sorry if I'm looking like a jerk but we're talking like 60-90 ms extra response versus 18 ms or none (this depends on the game obviously)

[Lance Boyle]

It's a pretty blatant difference

The controls responding ONE SIXTIETH of a second quicker is not what I or most people would call "a pretty blatant difference".

I do notice it myself, but again, it's impossible to distinguish what I sense from placebo until a proper test is taken since the difference is so minute.

[captpain]

It's a pretty blatant difference

The controls responding ONE SIXTIETH of a second quicker is not what I or most people would call "a pretty blatant difference".

I do notice it myself, but again, it's impossible to distinguish what I sense from placebo until a proper test is taken since the difference is so minute.

Not all games run at 60 Hz. Also, lots of the games have lost more than a single frame of lag.

I would suggest you use strong players in your highly-exciting double-blind tests that prove what any capable player already knows.

[captpain]

I guess the reason I'm getting a bit bent out of shape by what you're saying is because it's an established fact that such a (seemingly) small change is a perceptible difference. It's not some kind of weird outer limit of human capability, and it doesn't need any testing. People skilled in an activity such as this can perceive the change.

If you introduced a 50 ms lag when I plucked a guitar string, I'd go nuts. The difference matters.

[nimitz]

You can actually take the test : Input lag ABX blind test

(thanks to moozooh for the link)


Everyone I know who took that test were able to tell the difference between 1 and 2 frames lag with close to 100% accuracy only after a few tries (each try is 20 test i think).

So yes, it is indeed noticeable and for some people (like me) it is instantly noticeable. I can try a game in mame and tell exactly how many frames of lag it has after only a few seconds.

[TVG]

Can you use the pause and advance frame method to see the input lag added by V-sync, triplebuffer etc?
I just used it right now on a couple games and mame (mameplus extended) cannot tell the difference. I tested it by pausing the game then doing shift+p while holding my stick in a direction.
In ketsui the ship moved on the 2nd frame regardless if I used these options (separately) or not. I'm a little confused.

[jonny5]

Can you use the pause and advance frame method to see the input lag added by V-sync, triplebuffer etc?
I just used it right now on a couple games and mame (mameplus extended) cannot tell the difference. I tested it by pausing the game then doing shift+p while holding my stick in a direction.
In ketsui the ship moved on the 2nd frame regardless if I used these options (separately) or not. I'm a little confused.

I've always said wait for v-sync didn't introduce lag. People always seem to think it does. Not sure why?

Does it say somewhere it does?

All it really does is kill the screen tearing, which is a good thing.

[TrevHead (TVR)]

someone on that zeit2 thread mentioned that they do a waggle test as an easy way to test for lag. just move left and right very quickly if your ship starts to move right when you press left then you there you go.

[TVG]

I've always said wait for v-sync didn't introduce lag. People always seem to think it does. Not sure why?

Does it say somewhere it does?

There seems to be a consensus that it does. I did a few google searches and a lot of people say it.
It definitely does induce (noticeable) input lag in some FPS games, but I was never sure about mame because I cannot tell one frame of lag. I need at least two or more to start noticing it.

[SPYvsSPY]

It's a major difference without a doubt, just play the same game a couple times straight in the standard editions of mame,
then jump on the same game in laggless mame.

I thank you for the Laggless editions of mame nimitz I've played them all since their conceptions.

Maybe someone could find a way to implement the newer mame joystick program in to laggless, so that we could use our $150 Xbox 360 sticks? lol

[TVG]

someone on that zeit2 thread mentioned that they do a waggle test as an easy way to test for lag. just move left and right very quickly if your ship starts to move right when you press left then you there you go.

You'll be able to tell a massive lag that way maybe if you can't see it outright, but if you cannot tell one frame or two of lag otherwise, you won't be able to tell this way either IMO.

I know 2 or 3 frames is my threesold because I played quite a bit of fighting games on ggpo etc. Some people can tell one frame.

[jonny5]

Comparing PCB to shmupMAME at 15khz with direct draw at native resolution with wait for vsync enabled they are practically identical, both in appearance and control, for Cave stuff anyways.

If I sat you down and didn't tell you it was MAME, you wouldn't know.

[BPzeBanshee]

Allright, so I finally got to test this new version out, and I came across a few problems:

- For some reason on occaision Battle Garegga seems to say it has bad Text RAM. So far I've gotten this only once and when I enabled Service Mode and booted through that it disappeared.
- The Cave PGM games ran badly on my machine to begin with but now they're worse. I used to be able to get 100% speed with ESPGaluda with autoframeskip in what I believe was a leaked build of MAME .140u3 but it runs horribly now.
- Despite not having 'Increase Priority', on games that run poorly (the Cave games, Raiden Fighters always ran badly) the input starts lagging as if I had it set to high priority. Sometimes I'd move left a little bit, move up and then I'll just be moving up all the way to the top when I'm actually aiming completely somewhere else with the keyboard.
- I had problems figuring out how to set the aspect ratios and stuff for the games. It took me a while but the functions to set it to either 3:4 or 1:2 (like ESPGaluda) do exist in the release.
- Batrider's music sounds different compared to v2's release. Like the main instruments got weaker and the drums got louder. Is that a bug or bugfix? Either way I can't find how to change the audio tab settings like in newer versions of MAME.

On a good note:

- THUNDER FORCE AC RUNS FULLSPEED AT LAST YAY (would it be too difficult to ask for a lagless sega2.c driver for TFAC? It takes 3 frames for the ship to move in it.)
- Batrider really is much more responsive. So is Garegga but I still get owned in Garegga because of the invisible bullets and evil rank system.
- I dont get any general lag problems like what I had with v2.2 anymore. Once set up properly on the hard drive properly v3.0 is definitely the version to have for legacy machines.

[Clessx]

Thanks ! with this version 3 I can play all Cave pgm titles with no slowdown & lag and sound is also smooth. hiscore.dat doesnt seem to be updated to be compatible with new titles yet , but save states work just fine so no prob here. Is Batsugun sound emulated also included in v3 ? didnt obtain the new rom & test it yet.

I wonder if anyone has icons - snapshots - posters of pgm games for use with shmupmame ? without those the menu looks dull.

[BPzeBanshee]

Thanks ! with this version 3 I can play all Cave pgm titles with no slowdown & lag and sound is also smooth. hiscore.dat doesnt seem to be updated to be compatible with new titles yet , but save states work just fine so no prob here. Is Batsugun sound emulated also included in v3 ? didnt obtain the new rom & test it yet.

I wonder if anyone has icons - snapshots - posters of pgm games for use with shmupmame ? without those the menu looks dull.

You can always take screenshots of the game's menu yourself, that's what I do for my ShmupMAME setup.

[IseeThings]

It's actually a completely measurable effect. Take Batrider outside of the current version, and do a pause-frame advance test. Then, try it in Lagless v3.

I know about this. I'm just wondering if all the people screaming OMG IT'S SO MUCH MORE RESPONSIVE!! have actually tried taking a blind test.

I'd do it myself but I'm not sure how to arrange it. Probably with a .bat file or something.

Batrider doesn't buffer the sprites, even in regular MAME any buffering is being done by the game code.

I noticed this when looking at it, because it requires a specific case to handle in the driver, which to me makes it seem wrong because ALL the Toaplan / Raizing games in toaplan2.c use the same video chip. It's possible there is a CPU triggered buffer behaviour tho.

ThunderForce has no sprite buffering in the video hardware either.

Heaven help you for the PGM glitches, there are years worth of fixes to backport for that, I'm not even sure it had *zoom* support in 0.99

The reason it's faster for most stuff is that the emulation is less complete.

[ZOM]

APB lagless in mame? I came.

[Dave_K.]

Comparing PCB to shmupMAME at 15khz with direct draw at native resolution with wait for vsync enabled they are practically identical, both in appearance and control, for Cave stuff anyways.

If I sat you down and didn't tell you it was MAME, you wouldn't know.

Can you post your hardware and OS configuration?

I'm pretty sure I could tell the difference with DDP, as I've yet to see any Mame output exactly 57.550645 Hz on an arcade monitor at 15khz without a slight ripple (especially noticeable when lasering bosses that flash at 1/2 of that refresh). Best you can do is speed up the game to 60hz to play with an ArcadeVGA...but even then the speed increase is noticeable.

[PROMETHEUS]

It's actually a completely measurable effect. Take Batrider outside of the current version, and do a pause-frame advance test. Then, try it in Lagless v3.

I know about this. I'm just wondering if all the people screaming OMG IT'S SO MUCH MORE RESPONSIVE!! have actually tried taking a blind test.

I'd do it myself but I'm not sure how to arrange it. Probably with a .bat file or something.

Im pretty sure I could tell the difference with a blind test cause I definitely felt a nice difference in DDP the second I started MAME lagless, but yeah would be interesting to test.

[captpain]

It's actually a completely measurable effect. Take Batrider outside of the current version, and do a pause-frame advance test. Then, try it in Lagless v3.

I know about this. I'm just wondering if all the people screaming OMG IT'S SO MUCH MORE RESPONSIVE!! have actually tried taking a blind test.

I'd do it myself but I'm not sure how to arrange it. Probably with a .bat file or something.

Batrider doesn't buffer the sprites, even in regular MAME any buffering is being done by the game code.

I noticed this when looking at it, because it requires a specific case to handle in the driver, which to me makes it seem wrong because ALL the Toaplan / Raizing games in toaplan2.c use the same video chip. It's possible there is a CPU triggered buffer behaviour tho.

ThunderForce has no sprite buffering in the video hardware either.

Heaven help you for the PGM glitches, there are years worth of fixes to backport for that, I'm not even sure it had *zoom* support in 0.99

The reason it's faster for most stuff is that the emulation is less complete.

So Batrider in ShmupMAME v3 responds to player inputs faster than it does on the PCB? Is this for sure?

[nimitz]

So Batrider in ShmupMAME v3 responds to player inputs faster than it does on the PCB? Is this for sure?

Nope, because the way mame handles the video update on these games is wrong and it's impossible with the current way mame operates to make it be handled correctly.

So mame adds a frame of lag on these games, I managed to make it behave "normally" with some trickery. I will post detailed info for this fix on the blog soon.

also, the video driver for the Cave PGM is now working properly (zooming and all) and will be added to 3.0b.

[Zeron]

also, the video driver for the Cave PGM is now working properly (zooming and all) and will be added to 3.0b.

Awesome , but will the problem of 360 sticks not being detected fixed?

[BPzeBanshee]

also, the video driver for the Cave PGM is now working properly (zooming and all) and will be added to 3.0b.

Awesome , but will the problem of 360 sticks not being detected fixed?

Are you referring to the arcade-style 360 sticks or 360 controllers in general? I ask because my wired 360 controller using XBCD drivers works just fine with all versions of MAME including this one, and as I recall MAME used the official drivers fine when I first tried it too so investigation into custom driver support for 360 arcade sticks would be a better solution I would think.

[Zeron]

Are you referring to the arcade-style 360 sticks or 360 controllers in general? I ask because my wired 360 controller using XBCD drivers works just fine with all versions of MAME including this one, and as I recall MAME used the official drivers fine when I first tried it too so investigation into custom driver support for 360 arcade sticks would be a better solution I would think.

I got a SFIV TE stick with the newest official drivers installed, but il try the unofficial XBCD one, I just hope having two drivers installed at the same time wont create some drivers conflict.

[joeblade]

Are you referring to the arcade-style 360 sticks or 360 controllers in general? I ask because my wired 360 controller using XBCD drivers works just fine with all versions of MAME including this one, and as I recall MAME used the official drivers fine when I first tried it too so investigation into custom driver support for 360 arcade sticks would be a better solution I would think.

I got a SFIV TE stick with the newest official drivers installed, but il try the unofficial XBCD one, I just hope having two drivers installed at the same time wont create some drivers conflict.

I've got the tekken6 wireless and no joy so i'll be interested in your findings zeron!