Acceptable input lag in shmups?

[CDexWard]

After staying away from shmups successfully for about two years now, I cannot withstand the urge to get back into it right now. Currently I am setting up my tate setup via an old PC, 360 arcade stick and my precious gaming display. A couple of years ago I clearly had the wrong approach to shmupping, trying to play all Cave games at the same time. Now I want to really focus on a single game, Ketsui, play a couple of credits on my dedicated setup every day and get better at it.

By chance I stumbled upon this website today, which is already troubling me a bit in my preparations to start shmupping again: http://www.teyah.net/sticklag/results.html

I own two arcade sticks, the Madcatz TE Round 2 and the Madcatz FightStick Pro for Xbox 360, which are both ranked in the C Tier in this list and both with 6,6ms of lag. On top of that I will have another 16,6ms (= 1 frame) for using Vsync because i cannot bear the vertical tearing. The emulator will be Shmupmame 4.2. My display is the BenQ BL2411PT, which has another 5ms but is supposed to be lag-free otherwise.

That's 28,2ms of input lag to begin with on top of what my graphics card, directx and stuff I haven't thought about add to this. How severe is this amount of lag?

This is concerning me quite a bit. In the meantime I have been playing a lot of SNES and GBA games and never worried about input lag but I know that shmups require a lot faster reactions and twitch dodging and I would be highly demotivated to start heavily playing a game in which what I can achieve is limited by the delay my hardware brings to the table. I want to play this game for years now, be able to get as good as possible and experience everything, hunt for high scores etc. Is this still possible with the aforementioned amount of lag?

[iconoclast]

Your monitor and stick are adding less than a frame of input lag, it's irrelevant. Just play the game on 360 or PS3 if you want to hunt for high scores. Everyone on the leaderboard will be playing the exact same game with the exact same amount of inherent input lag, and the #1 scores on there are probably higher than any scores on here. There's nothing to worry about aside from external factors, which are fine in your case.

[Bananamatic]

use shmupmame v2.2 and always turn vsync and possibly any other filters off
either get used to the minor tearing or buy the ports tbh

[Shepardus]

I'd wager that unless it's really bad, input lag isn't going to be your bottleneck. Try this (first link, then the first link on that page which should say "Download") to see how sensitive you are to input lag. I personally can't tell the difference between 1 and 2 frames, 3 is noticeable and 4 is unmistakable.

[Xyga]

GroovyMAME, BGFX, autosync.

[Obiwanshinobi]

I personally can't tell the difference between 1 and 2 frames, 3 is noticeable and 4 is unmistakable.

Measuring input delay with step-frame test, played on a CRT* with a wired PlayStation controller plugged via USB adapter on a 32-bit WinXP system, WITHOUT v-sync, I did feel the difference between 1 and 2 frames, not so much between "0" and 1. And I sure as hell felt the lag in Zanac Neo when played on my PS2 (NOT in various 8-bit Zanac versions complementing it), CRT again. Bought PSX solely for the sake of playing Neo with lesser delay.

*) At 100, then 120 Hz refresh rates.

As a rule of thumb, shmups lagging 1 or even 2 frames (just step-frame test values, remember) on the aforementioned setup didn't feel any less playable than shmups I'd played on consoles at the time.

Can't remember how many frames Sengoku Ace lagged on "shmupMAME", but PS2 version (whether it's emulated on ported there; I've no idea) felt more responsive still.

Not something I'd worry about for as long as it's not distracting; a game must "feel" good enough for you to effectively practice hard. Oh, and if you MUST worry, you had better get used to screen tearing (v-syc adds a degree of delay no matter what). Of course, there are computer games that run all wrong if not v-synced or synced with "inappropriate" refresh rates, but I've yet to find it to be a problem with any emulator.

[OmegaFlareX]

I personally can't tell the difference between 1 and 2 frames, 3 is noticeable and 4 is unmistakable.

Came here to post exactly this.

Vsync adds 4+ frames by itself, so as some have said - get used to screen tearing or buy the ports.

[trap15]

Vsync adds 4+ frames by itself

lmfao, citation needed. It's 1 frame at most.

[Cagar]

-

[trap15]

No, at most. Vsync will wait to the end of the current frame to display the next frame. So 1 frame at most.

[Bananamatic]

then why does mushi on steam go from kinda shitty to unplayable with vsync on
is there a single frame breakpoint where everything goes to shit

the vsync patches in touhou also make a massive difference, no way that is one frame

[Obiwanshinobi]

I'm certain v-sync can be and is being applied in many different ways, depending on software. I know because GPU in my PC has a peculiar issue with the LCD TV it's hooked up currently - v-synced stuff typically assumes it's being refreshed at 50 Hz (!) and I have some fun & games with it: performance, lag, speed and whathave you differ WILDLY between games when tested with and without v-sync. I have specific examples just how wildly to give if anybody's interested.

P.S. It appears my GPU has no longer said problem with this TV of mine; maybe it was only a thing with the WinXP install from before switching monitors.

So I used to play Ys: The Oath in Felghana with older graphics card in 640x480 v-synced, but CRT refresh rate was 120 Hz back then and framerate would practically never drop below 60 fps. Under those circumstances, I think any delay v-sync brought to the table was "academic". This morning I just played it with a way more powerful GPU (a tad more potent PC overall, if still 32-bit WinXP system), except this LCD can't be set to refresh rate higher than 60 Hz. The framerate when not v-synced, however, constantly exceeds 100 fps. Therefore, although overall input lag must be higher than it was on CRT, I - again - think the impact v-sync has on response times here must be minuscule still (certainly nothing I would sensually detect).