-

[StarCreator]

There are lots of reports of turning aero off reducing input lag.
There aren't any reports of turning aero on reducing input lag.

It's logical to try turning it off first. And another reason to recommend to try to turn it off: it's on by default. Off is the only other state you have to try.

Why do you care? What could it possibly hurt to try?

Of course it would be interesting to see a scientific analysis of how aero influences input lag, but I don't see anyone doing that. Also as said, it's entirely possible aero input lag is influenced by other factors.

So basically, if Google reports to you that you can see the road from your car better if you remove your car's windshield, is your default reaction to remove the windshield?

You keep saying because people on the Internet say it is so, then it must be true. Is it too much to ask you to think for yourself rather than putting blind faith in people who may or may not be qualified to make such judgements and parroting whatever you happen to read unprompted? Why is it so wrong to question a statement that has no concrete evidence backing it?

[ZellSF]

There are lots of reports of turning aero off reducing input lag.
There aren't any reports of turning aero on reducing input lag.

It's logical to try turning it off first. And another reason to recommend to try to turn it off: it's on by default. Off is the only other state you have to try.

Why do you care? What could it possibly hurt to try?

Of course it would be interesting to see a scientific analysis of how aero influences input lag, but I don't see anyone doing that. Also as said, it's entirely possible aero input lag is influenced by other factors.

So basically, if Google reports to you that you can see the road from your car better if you remove your car's windshield, is your default reaction to remove the windshield?

You keep saying because people on the Internet say it is so, then it must be true. Is it too much to ask you to think for yourself rather than putting blind faith in people who may or may not be qualified to make such judgements and parroting whatever you happen to read unprompted? Why is it so wrong to question a statement that has no concrete evidence backing it?

That's exactly what I'm asking people to do. Try it. Don't assume it makes or doesn't make a difference, test both settings and figure it out yourself.

Yes, I said it pretty definitely, but so did the people recommending changing prerendered frames or running the game in windowed mode, and none of them have a scientific proof of why that works either. Assuming what everyone says is true might be a bad idea, but ignoring what everyone says is much worse.

Oh, and car analogies? Really?

[Keade]

For AMD users, the max prerendered frames setting mentioned a few times here is available in ATI Tray Tools as "flip queue size". (...)

Thank you very much, I can't wait to try this

I have just tested this using RadeonPro in fullscreen : far from perfect but better for sure. I get even less input lag in fullscreen using frame limiting, but this causes awful tearing all over the place.
Going to stick to windowed until a fix is available...

[Ghegs]

I seem to belong to the lucky "no noticeable input lag"-group. Played for a while (USB Saturn pad via XPadder, fullscreen) and it felt nice 'n accurate the whole time. I haven't messed with any of the game's settings, either.

It is a bit silly to have to unplug my 360 controller just so I can play it with the keyboard-emulating-XPadder. Hopefully that gets patched along with the other things.

[e_tank]

Desktop composition causing input lag is well known.

If you don't like tearing, force vsync on in your graphic card drivers rather than relying on your desktop compositor to do it. It leads to some weird issues, especially with emulators.

Known by whom? How long ago? Under which versions of Windows? Where is the proof?

i'm tired and my brain is fried but let me to try explain this real quick..
ZellSF is correct, by design a compositing window manager can only increase an applications response time (from what it would be without it), it can't reduce it any. the compositor adds a layer of complexity between the application and your os's graphics stack. instead of writing directly to the display buffer, under window composition the application will write to a buffer provided by the compositor, which (when running in windowed mode) will then composite all the buffers it has together before writing its result back to the display buffer.

for an application that would otherwise be run with vsync enabled, under ideal circumstances it's possible for the compositor not to add to its response time, but it's def not the norm. usually the compositor will either consistently or occasionally skip/miss frames. in the end the severity of any increase in response time is dependant on your particular system/setup.

Is it too much to ask you to think for yourself rather than putting blind faith in people who may or may not be qualified to make such judgements and parroting whatever you happen to read unprompted?

if you can't take john carmack's word on the matter, who can you trust?

[StarCreator]

i'm tired and my brain is fried but let me to try explain this real quick..
ZellSF is correct, by design a compositing window manager can only increase an applications response time (from what it would be without it), it can't reduce it any. the compositor adds a layer of complexity between the application and your os's graphics stack. instead of writing directly to the display buffer, under window composition the application will write to a buffer provided by the compositor, which (when running in windowed mode) will then composite all the buffers it has together before writing its result back to the display buffer.

for an application that would otherwise be run with vsync enabled, under ideal circumstances it's possible for the compositor not to add to its response time, but it's def not the norm. usually the compositor will either consistently or occasionally skip/miss frames. in the end the severity of any increase in response time is dependant on your particular system/setup.

Yes, that's how I thought it worked. In practice though, I've never noticed this process creating any more lag than in-game vsync, nor has anyone I personally know who has tried it under Windows 7. (I've heard worse things about Windows Vista's implementation, but I don't think any of us are still running Windows Vista.)

if you can't take john carmack's word on the matter, who can you trust?

Quickly searching that entry he doesn't actually state the difference between a vsynced output and an windowed-Aero output, only that he disabled compositing (AND vsync, presumably) for absolute minimum latency. That was never what I attempting to compare; I'm not going to give up vsync entirely until there's an acceptable alternative (GSync).

[katzinator]

Thanks a million for your contributions and clarifications to this thread, ZellSF and e_tank. I'm looking forward to giving your suggestions a shot when I get home from work tonight.

[ratikal]

I seem to belong to the lucky "no noticeable input lag"-group. Played for a while (USB Saturn pad via XPadder, fullscreen) and it felt nice 'n accurate the whole time. I haven't messed with any of the game's settings, either.

It is a bit silly to have to unplug my 360 controller just so I can play it with the keyboard-emulating-XPadder. Hopefully that gets patched along with the other things.

I'll have to try out Xpadder. I'm not entirely sure if it was lag I was feeling when I was playing with my Saturn Pad w/ Joy2Key. I didn't notice any lag at all using the keyboard and mouse.

By the way, keyboard and mouse works out really well in this game. I'm actually kind of impressed.

[P_HAT]

Just wanted to say thanks for anyone who helped(one way or another) with lnput lag case.

Looks like "MPF 1" works for me.

[blufang]

1. I'm not experiencing input lag, but I'm no expert. Everything instantly responds, unlike many fighting games online which use input lag based netcode. Fightsticks work too, as long as they are for the XBOX360 (heard people complain about this on steam I think). I'm using a Madcatz SFxT TE stick.

2. Anyways I'm thrilled, I've never owned the game. However a friend had it for dreamcast, and I played it a lot way back then. I usually wait for steam sales, but I'm paying full price cause I really want to support this. I look forward to Treasure releasing their other games such as Radiant Silvergun, Bangai0, Guardian Heroes as well as the supposed 'new game for pc'.

3. I think other than Sine Mora and Jamestown there haven't been many big name shmups, ones that are reviewed by many publications get a lot of press, etc. Hope Ikaruga sells millions of digital copies.

[Raytrace]

while I didn't really have many complaints (except for my own lack of skill :p) - the maximum frames prerender 1 setting does seem to make a noticeable difference, I even managed to lose no lives on the first Stage!

[Zaarock]

I seem to belong to the lucky "no noticeable input lag"-group. Played for a while (USB Saturn pad via XPadder, fullscreen) and it felt nice 'n accurate the whole time. I haven't messed with any of the game's settings, either.

It is a bit silly to have to unplug my 360 controller just so I can play it with the keyboard-emulating-XPadder. Hopefully that gets patched along with the other things.

I'll have to try out Xpadder. I'm not entirely sure if it was lag I was feeling when I was playing with my Saturn Pad w/ Joy2Key. I didn't notice any lag at all using the keyboard and mouse.

By the way, keyboard and mouse works out really well in this game. I'm actually kind of impressed.

You can also try x360ce . No idea if it introduces any less lag than keymappers though.. it just emulates a regular 360 controller instead.

[gs68]

In an act of masochism, stupidity, and for the purpose of Twitter accounts dedicated to dumb forum quotes, I decided to have a look at the Ikaruga Steam forum.

Someone seems to be throwing a huge fit over the lack of online multiplayer.

Ah, video game players as usual.

[Formless God]

Is there a way to quickly reset the game instead of having to go back to the title screen -> select the game mode -> watch the pretty launch sequence? My restartitis can't take this.

[hail good sir]

Enable easy menu, I think it's in the extra settings. It should save your selections and make it only a couple button presses.

[Formless God]

Yeah, that's better. Thanks!

[BazookaBen]

Playing at 1728x2304 on a 21" Lacie CRT. So basically aliasing-free and perfect color in fullscreen 3:4. Treasure truly delivered on this version... well, after they patch offline and input lag bugs.

Used Radeon Pro to force my pre-rendered frames (called flip-queue in this program) to 1, then 0. Both settings seemed to have no lag, though I'll have to load up the gamecube version tomorrow to compare.

[Agentunknown]

This is great news. But I will wait for a Steam sale and this will be $4.99

[ZellSF]

Playing at 1728x2304 on a 21" Lacie CRT. So basically aliasing-free and perfect color in fullscreen 3:4. Treasure truly delivered on this version... well, after they patch offline and input lag bugs.

Used Radeon Pro to force my pre-rendered frames (called flip-queue in this program) to 1, then 0. Both settings seemed to have no lag, though I'll have to load up the gamecube version tomorrow to compare.

I doubt there's much to patch when it comes to input lag beyond what you can fix yourself.

Of course even if every PC gamer should know how to force it, Treasure should still try to force prerender queue off and add options for framelimiting and disabling vsync.

oh, if you are a PC gamer remember that you should not be forcing triple buffering for Ikaruga, it does not need it and it might increase input lag (if you're forcing triple buffering, you'll know it, everyone else ignore this).

[Keade]

^ Well, I sure hope Treasure improves responsiveness in fullscreen, because I could not get fullscreen to work good (AMD HD 6870 + RadeonPro). Unless I have missed some important setting in RadeonPro, it looks like I have to choose between terrible input lag or okay reponsiveness but horrid 50-pixels-tall-band-tearing.
It would be less of a problem if windowed could go up to 1080p with small window borders, but windowed display is currently limited to 720p with big borders.

[ZellSF]

^ Well, I sure hope Treasure improves responsiveness in fullscreen, because I could not get fullscreen to work good (AMD HD 6870 + RadeonPro). Unless I have missed some important setting in RadeonPro, it looks like I have to choose between terrible input lag or okay reponsiveness but horrid 50-pixels-tall-band-tearing.
It would be less of a problem if windowed could go up to 1080p with small window borders, but windowed display is currently limited to 720p with big borders.

You're looking for "dynamic framerate control" in RadeonPro. Limit framerate to 60. Framerate limiting is important when running vsynced. The setting might only be available in preview versions of RadeonPro. Look here.

[Keade]

IIRC I had tried framelimiting and this is what caused the tearing. I am not 100% sure "dynamic framerate control" was checked though, will have to check ...

@gs68: people have grown to like co-op and multiplayer, I do not see anything to be surprised or appalled about (except that Ikaruga in co-op is a nightmare imho)

[ZellSF]

IIRC I had tried framelimiting and this is what caused the tearing. I am not 100% sure "dynamic framerate control" was checked though, will have to check ...

Framerate limiting should not cause tearing at all. Remember vsync should still be on. Make sure that dynamic vsync or whatever AMD calls their equivalent is off.

[AntiFritz]

Another thing I've noticed that's missing from the xbla port is the dot eater leaderboards.

[BazookaBen]

Ok guys, I jus set up my two CRT's, one for the steam version, one for the gamecube version. I switched back and forth, playing small segments of level one, doing figure 8's to get a sense for the responsiveness. I feel like even with pre-rendered frames set to 1, there was still a very slight delay on the PC, like 1 frame max. It was so miniscule that I thought it could just be the different d-pads throwing me off.

Anybody else wanna try a side by side and see if you get the same results?

[katzinator]

Ok guys, I jus set up my two CRT's, one for the steam version, one for the gamecube version. I switched back and forth, playing small segments of level one, doing figure 8's to get a sense for the responsiveness. I feel like even with pre-rendered frames set to 1, there was still a very slight delay on the PC, like 1 frame max. It was so miniscule that I thought it could just be the different d-pads throwing me off.

Anybody else wanna try a side by side and see if you get the same results?

You are fantastic for giving this a shot and I sincerely appreciate it. I feel like the best way would be to have someone with a hacked gamepad or multi-console joystick controlling both simultaneously to be absolutely sure. Still, your test is definitely a good start.

If I get some spare time next week I'm going to give running it simultaneously with a DC/GC emulator and capture/compare any differences (assuming said emulator(s) are lag-free :\)

[ZellSF]

Ok guys, I jus set up my two CRT's, one for the steam version, one for the gamecube version. I switched back and forth, playing small segments of level one, doing figure 8's to get a sense for the responsiveness. I feel like even with pre-rendered frames set to 1, there was still a very slight delay on the PC, like 1 frame max. It was so miniscule that I thought it could just be the different d-pads throwing me off.

Anybody else wanna try a side by side and see if you get the same results?

You are fantastic for giving this a shot and I sincerely appreciate it. I feel like the best way would be to have someone with a hacked gamepad or multi-console joystick controlling both simultaneously to be absolutely sure. Still, your test is definitely a good start.

If I get some spare time next week I'm going to give running it simultaneously with a DC/GC emulator and capture/compare any differences (assuming said emulator(s) are lag-free :\)

I really, really doubt those emulators are lag-free.

Considering how much more abstraction there is on PC, only 1 frame delay over the GC version is much better than I would have expected.

[StarCreator]

You are fantastic for giving this a shot and I sincerely appreciate it. I feel like the best way would be to have someone with a hacked gamepad or multi-console joystick controlling both simultaneously to be absolutely sure. Still, your test is definitely a good start.

It's more likely differences in slowdown between the two ports would cause too much desync to make it a worthwhile test.

[El_Spank]

I can't comment on input lag with any authority since I haven't really touched the game in about 8 years but I am running PC Ika with XB360CE and a PS2 Hori Fighting Stick via a Rockfire PS2-USB adapter.

If there is any lag, it's not noticeable by me. Then again, bear in mind what I said in my first sentence.

God I haven't been on this site in years. I pretty much quit playing videogames for about 5 years (other than Eve Online). Hope everyone is well ♥

[Keade]

I had forgotten that the game relied so much on memorization for simple chaining, I get owned as soon as level 2.
Also, the leaderboards have been populated with a lot of high scores, such that my day one scores now "officially" suck (everyone in my friends list suck more, 'though)

+1 against emulators measurements, there are not lag free so cannot provide an accurate reference point imho.

EDIT:

Framerate limiting should not cause tearing at all. Remember vsync should still be on. Make sure that dynamic vsync or whatever AMD calls their equivalent is off.

The game inputs are laggy as long as I keep don't force Vsync off (Vsync control = Always off) + Lock frame rate up to monitor's refresh rate.
This seems to give good results: seemingly no tearing, and lower input lag \o/