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

[Illyrian]

I have personally not had any problems with shmupmame 2.2 but wolfmame 0.99 won't run at all on windows 7 these days.

My suggestion would be to download a new set of base files/folders from a new version of mame and re-download the .exe file.

Also, do these problems only happen with MAME? You may want to reinstall your OS as that can have lots of positive effects on system stability.

[Ed Oscuro]

My suggestion would be to download a new set of base files/folders from a new version of mame and re-download the .exe file.

How's that going to help? MAME should launch no matter what folders you have alongside it - especially when you use those originally supplied with the executable.

Maybe we should just get a thread lock instead?

Udderdude isn't speaking for Nimitz. He may still deign to work on it some more - or not. Even so, where would all the posts go? There'd just be another thread.

[BPzeBanshee]

Maybe we should just get a thread lock instead?

Udderdude isn't speaking for Nimitz. He may still deign to work on it some more - or not. Even so, where would all the posts go? There'd just be another thread.

Good point, unfortunately. Having a dozen threads getting locked cause of this one is probably worse than a single thread with the same stuff repeated 9001 times over 50 pages - better to pick the lesser of two evils.

I don't suppose anyone's tried out WolfMAME Plus 0.99 on their problem machines to see if that worked or not? At one point it seemed like porting the Shmupmame drivers to that build may have been the solution but someone reported a different problem afterwards and everyone else seems to have ignored it.

[J_Taishu]

Sorry to just ask a question out of the blue, but does the slowdown I'm getting on Gunlock (Ie. Rayforce/Layer Section) on the original game, or is it being caused by SHMUP Mame?

Also, not a SHMUP but I had DaiMakaiMura suddenly reset on me during the 4th boss. Has anybody heard of this issue?

[Ed Oscuro]

Sorry to just ask a question out of the blue, but does the slowdown I'm getting on Gunlock (Ie. Rayforce/Layer Section) on the original game, or is it being caused by SHMUP Mame?

Also, not a SHMUP but I had DaiMakaiMura suddenly reset on me during the 4th boss. Has anybody heard of this issue?

There is a bulletproof way of checking, in MAME, whether the game is playing at "full speed." Hit F11. If the percentage shown drops below 100%, Layer Section is too much game for your PC. This seems unlikely, though. There is a remote possibility something in ShmupMame is not working right, but I doubt it...Layer Section is a popular enough game that I imagine hackles would have been raised and eggs thrown if the driver was added broken (or broken when added). The most likely possibility is that there could be some slowdown in the original game (I don't remember); it might help to say where you are getting the slowdown.

I would suggest checking out the newest EmuCR MAME Plus! available, just to see how it goes. The newer MAME releases are more demanding than the old ones, but you might be pleasantly surprised.

[BPzeBanshee]

There is indeed some slowdown in Rayforce as emulated in Shmupmame, particularly 1/3rd way through stage 2 when you hear the voice going 'turn left'. I've seen it in standard MAME builds but I believe there was a tag saying it might not have been PCB-accurate (in case anyone cares, I'd check MAWS but they're down).

[J_Taishu]

There is indeed some slowdown in Rayforce as emulated in Shmupmame, particularly 1/3rd way through stage 2 when you hear the voice going 'turn left'. I've seen it in standard MAME builds but I believe there was a tag saying it might not have been PCB-accurate (in case anyone cares, I'd check MAWS but they're down).

Yep, that exactly where I noticed it at first, although there is occasional slowdown throughout the game. Credit fed through the whole thing yesterday, but the % never dropped below 100, so I guess it's either the emu or the original and not my computer. I'll try testing it out on a few different versions of MAME.

[mjclark]

There is indeed some slowdown in Rayforce as emulated in Shmupmame, particularly 1/3rd way through stage 2 when you hear the voice going 'turn left'. I've seen it in standard MAME builds but I believe there was a tag saying it might not have been PCB-accurate (in case anyone cares, I'd check MAWS but they're down).

Yep, that exactly where I noticed it at first, although there is occasional slowdown throughout the game. Credit fed through the whole thing yesterday, but the % never dropped below 100, so I guess it's either the emu or the original and not my computer. I'll try testing it out on a few different versions of MAME.

Observer picked up on this problem last year:
http://shmups.system11.org/viewtopic.ph ... 54#p731554
It's only started occurring in the post-SH3 builds of MAME but due to all the furore about the CAVE emulation it seems to have just been overlooked...

[J_Taishu]

Found it:
http://mametesters.org/view.php?id=4493

I'll just find an older version to play it on.

[BPzeBanshee]

I'd check out Rayforce/Gunlock on MAME 0.145 again before you go and do that J_Taishu. I just ran it now and there was absolutely no stuttering or slowdowns AT ALL. I'm running it on a Core 2 Duo 2.8 Ghz under Windows XP using Gunlock (the parent).

[J_Taishu]

It runs without slowdown in .145
Thanks.

[BPzeBanshee]

No worries, it was a huge surprise for me too. I honestly thought the slowdown featured with Rayforce in Shmupmame was like the real thing, but I guess not.

That being said though I've never heard any of the sound clipping stuff that Observer mentioned, ever. Not sure what the go is with that.

[Anonanon]

Is there a way to make the window scale exactly 2x, 3x, etc. of the game's natively resolution? For example, DoDonPachi is 240x320, I want it to start up at 480x640 exactly. It seems the prescale function doesn't do this... I guess I'm looking for a postscale function?

[Bee Cool]

So does anyone have an idea of how we can get a less laggy version of Batrider, Garegga, and Bakraid without having to get a machine that doesn't use windows 7?

It really blows that I can't run Shmupmame 3.0 anymore.

[Ed Oscuro]

To summarize the page replies previous to yours: Use MAME 145 or newer

So far nobody has brought evidence that new builds have more input lag than Shmupmame.

[moozooh]

ShmupMAME's lag removal is a hack, and Haze is very clearly against it, so it won't find its way into MAME even theoretically.

[BPzeBanshee]

To summarize the page replies previous to yours: Use MAME 145 or newer

So far nobody has brought evidence that new builds have more input lag than Shmupmame.

I've said this before and I'll say it again - I see no evidence that MAME 0.145 had less input lag compared to Shmupmame v3.0b nor have I noticed any improvements myself, and I noticed a pretty big difference playing ESPRade in new builds compared to even the first Lagless MAME release. Mahou Daisakusen still had more frames of delay than Shmupmame v3.0b using the frame-by-frame advance test when I tried it after you and that other guy mentioned it was 'improved'.

If you have other games you'd like me to try to confirm/deny your view, however, I'm all ears.

[Bee Cool]

To summarize the page replies previous to yours: Use MAME 145 or newer

So far nobody has brought evidence that new builds have more input lag than Shmupmame.

I just tried it, and you are completely wrong. Batrider still has 4 frames of input lag and Garegga has 5.

Edit: By some miracle Shmupmame 3.0 actually launched, but who knows if that will happen again. But I can definitely confirm that the input lag for Batrider and Garegga was 2 frames as it is supposed to be.

[Ed Oscuro]

To summarize the page replies previous to yours: Use MAME 145 or newer

So far nobody has brought evidence that new builds have more input lag than Shmupmame.

I just tried it, and you are completely wrong

If you bring some new evidence NOW, that doesn't make me wrong since I was talking about earlier. Remember that I can't launch Shmupmame 3 so I can't do a comparison myself.

RE Banshee: I am seeing the 5 frames of lag in Garegga 145; 3 frames for Shmupmame 2.2.

[Drachenherz]

Btw, what's wrong with shmupmame 2.2. For me, it runs garegga quite playable. Batrider on the other hand still seems to have some lag, at least it feels like this. And 2.2 runs completely stable with my win 7 64bit OS.

Edit: ed ninjad my post

So garegga has 1 more frame of lag on 2.2 compared to 3?

[Ed Oscuro]

Btw, what's wrong with shmupmame 2.2.

Missing / wrong ROMs for some games because it's too damn old.

Also:

Armed Police Batrider:
In Shmupmame 2.2, I get 4 frames of lag with Miyamoto.
In 145, I get 4 frames of lag with Miyamoto.

So far I'm only seeing Batrider as problematic. (Maybe I should trade in my Japanese APB for a Garegga..? )

[Drachenherz]

Btw, what's wrong with shmupmame 2.2.

Missing / wrong ROMs for some games because it's too damn old.

And ninjad my edit

[Ed Oscuro]

Btw, what's wrong with shmupmame 2.2.

Missing / wrong ROMs for some games because it's too damn old.

And ninjad my edit

Sorry

Update:
4 frame lag behavior for Earth Defense Force in both SM 2.2 and 145. (Actually, EmuCR-MAMEPlus-r4988-x64 which is 145u3.)

[Zaarock]

Armed Police Batrider:
In Shmupmame 2.2, I get 4 frames of lag with Miyamoto.
In 145, I get 4 frames of lag with Miyamoto.

afaik it took him until 3.0 to reduce lag on batrider because it was hard for some reason, look here http://shmupmame.wordpress.com/2011/01/ ... rsion-3-0/ and older posts in this thread

[Ed Oscuro]

Guardian Force: 145u3 is 8 frames lag, also completely unplayable. Shmupmame 2.2 is 7 frames lag. Both tagged with inaccurate sound and video.

[Bee Cool]

If you bring some new evidence NOW, that doesn't make me wrong since I was talking about earlier. Remember that I can't launch Shmupmame 3 so I can't do a comparison myself.

RE Banshee: I am seeing the 5 frames of lag in Garegga 145; 3 frames for Shmupmame 2.2.

Wait what? Shmupmame 3.0 was created to reduce lag in Raizing's games, so of course Shmupmame 2.2 will be laggier thats the whole point. MAME has never had lower input lag than shmupmame 3.0 regarding Raizing's games, so you're just making stuff up if you claim that. I don't know why you ever thought shmupmame didn't actually reduce input lag on the games it claims it does.

[BPzeBanshee]

5 frames for MAME .145 in Battle Garegga and 3 in Shmupmame v2.2? Jesus I don't remember it being that bad in the builds.

The main reason I don't support v2.2 is because it's a hog comared to v2.0 in overall performance never mind frame delay. I'll have a go at Garegga with these now and compare.

EDIT: Garegga's weird in .145, using Miyamoto and doing the frame-by-frame advance results in some iffy responses, sometimes one frame of pressing LEFT results in it doing nothing and others it starts moving left 2 frames in for three steps and sprites/backgrounds moving separately and stuff. In Shmupmame v2.2 there is only one frame of lag with Miyamoto moving on the second frame after pressing LEFT (frame 1 I press LEFT and skip, nothing, frame 2 press nothing and skip, Miyamoto moves), and behaving the same way each time. In Shmupmame v3.0b it's even better (frame 1 I press LEFT and skip, he moves left).

Maybe it's just me but Garegga looked more broken with the normal build than in ShmupMAME - the response is certainly worse.

[Ed Oscuro]

MAME has never had lower input lag than shmupmame 3.0 regarding Raizing's games, so you're just making stuff up if you claim that.

It's a good thing I didn't claim that!

I don't know why you ever thought shmupmame didn't actually reduce input lag on the games it claims it does.

I didn't claim this either. However, you'll note that Shmupmame is being mooted as a care-free replacement for other builds of MAME. I'm interested in all games, including the games it's intended to fix for lag - but I'm interested in games individually, and those that are equal or bettered by newer MAME builds in accuracy are not getting my vote for Shmupmame use.

EDIT: Garegga's weird in .145, using Miyamoto

At the risk of sounding like I'm abandoning my position above (since obviously many people use Miyamoto), is that a different situation than e.g. Flying Baron or one of the other standard ships?

and doing the frame-by-frame advance results in some iffy responses, sometimes one frame of pressing LEFT results in it doing nothing and others it starts moving left 2 frames in for three steps and sprites/backgrounds moving separately and stuff. In Shmupmame v2.2 there is only one frame of lag with Miyamoto moving on the second frame after pressing LEFT (frame 1 I press LEFT and skip, nothing, frame 2 press nothing and skip, Miyamoto moves), and behaving the same way each time. In Shmupmame v3.0b it's even better (frame 1 I press LEFT and skip, he moves left).

Again, it's not bad to know about unusual cases, but the method I use to count lag frames is simply to start moving in a direction for a number of frames, then reverse, and count frames until it registers the new movement direction. This gives pretty objective results for the most important scenario, which is reversing direction. I would be surprised if the movement from idle to full motion differed, but if it did that would be pretty game-specific.

[BPzeBanshee]

At the risk of sounding like I'm abandoning my position above (since obviously many people use Miyamoto), is that a different situation than e.g. Flying Baron or one of the other standard ships?

I'm not sure what you mean here. I only picked Miyamoto because I noticed you picked that one for Batrider so I picked it for both games when I tried them out. I suppose the results *could* vary with other ships but when I tried this last time I used different ships and got the same results.

Again, it's not bad to know about unusual cases, but the method I use to count lag frames is simply to start moving in a direction for a number of frames, then reverse, and count frames until it registers the new movement direction. This gives pretty objective results for the most important scenario, which is reversing direction. I would be surprised if the movement from idle to full motion differed, but if it did that would be pretty game-specific.

Point taken - actually it's a good idea indeed. I'll use this method for future tests - I remember the last time there was comparisons going on in the thread everyone went from idle to moving left/right. Do you personally count the frame on which you first inputted the reverse or the frame after you've let go?

You're right about such motions being game-specific too - that would be a function programmed in for improvement in tap-dodging, but if done wrong it can also be perceived as frame delay or even inertia. Xeno Fighters R does this as of v0.11 and does it right but I've seen it in other games too, possibly Baryon: Future Assault.

[dmauro]

I'm having a lot of trouble getting v3.0b working on my cab. It wants to run at the desktop resolution despite going into default game options and using either "auto" or explicitly setting a resolution. I would love some help from anyone that has managed to get these looking arcade perfect in a cabinet.