Can anyone explain to me what is going on in this thread and why it’s important?
I explained it in the second post in one sentence. Here is the long version.
Up to this point we only knew that OE PS3 and 360 lagged the same as each other. No one had ever tested cps3 for lag or used it as a baseline to compare the console ports to. Now we know that online edition does not have extra input lag that real arcade hardware didn’t have. PS2 lags an extra frame on all inputs.
This doesn’t mean they’re the same overall. Everyone who has spent a lot of time on cps3 feels that there’s a definite difference in speed and feel between cps3 and the console ports as well as the major emulators. There are a lot of things that can cause that that aren’t input lag. We’ve just eliminated one possibility.
We are going to test some of the other possibilities over the next couple weeks. There are also some possibilities that could cause a difference in how the game feels and responds that we have no good way to test. There was a thread about two years ago where various members of this community investigated some of those and came up with nothing conclusive. I’m sure it’s still around to read somewhere.
In a few week we will test FBA (the emulator that GGPO uses) and Shmupmame (a low lag alternative to MAME) and possibly Groovymame under the same conditions to see if there is any extra lag over cps3 there. Note that we will be using a CRT, editing the USB polling rates, and using windows XP which is better for lag than windows 7 or 8. Which is as close to ideal setup as we can get. Realize that whatever results we get for emulation are best case results, not what most people playing on GGPO are going to see. Windows versions with compositing, bad USB polling rates, bad video drivers, vsync, LCD monitors, all of these are common in your average computer setup and will adversely affect your input lag experience.
1 Like
In terms of real world applicability, these tests are useful for deciding which version of the game will give you the closest to arcade experience. Since arcade is the standard across most communities where 3s is thriving, players who are serious about the game have an interest in being as close to arcade as possible without spending 1000 or more to get a working arcade setup.
will you test just the older FBA version on GGPO or the more recent ones as well?
A good test i just thought about to compare combo timing on arcade and OE/FBA: the Necro downback HP loop.
Each 6 hits are 1 frame and whole combo use only one button which can be ‘led-ified’ to see any difference and is around ~250 frames long.
It’s not really easy (at least for me ^^) but definitely easier than daipan loop!
Anyone have access to a programmable controller? That’ll remove any human errors.
One aspect of OE that made me wonder about timing was the hit stun window. It could just be my perception, but it could explain the observation stated before that hit confirms are harder in OE. If OE is somehow quicker than arcade in that particular aspect it would result in many different timing issues.
Tried ShmupMame today and I must say it’s much much better experience than GGPO’s FBA 
Glad to see a thread about this.
Quick question and request if its not much of a hassle:
Can the Dreamcast version be tested? Thanks in advance
PRE-FUCKING-CISELY
GGPO’S fba is garbage. It’s running an outdated version and little interest from Ponder to update and improve the overall experience.
GroovyMame and ShmupMame both vastly outperform it.
You guys should test those madcatz arcade sticks, I have a feeling that the lag players are complaining about, is coming from the madcatz pcbs not the ps3.
Do they use much more resources? My current laptop is craptastic, dusting it would require tearing down the whole exterior and emulation is expensive.
If someone wants to send us a Dreamcast we could test I guess. but I think Dreamcast is already well established as being a poor (laggy, inaccurate) port with notable hitbox issues. So even if we find out exactly how much lag it has probably no one is going to use it.
With Madcatz, that’s another thing we don’t own. but isn’t the Madcatz TE PCB already known to add a frame of lag? pretty sure I saw that someone else tested that years ago. that’s we we use Hori!
It’d actually be interesting to see if there were input lag comparisons between different sticks on different systems. Thus far the only one I’ve seen is Toodles’ test of Q4 v. Cthulhu.
note to self/note out loud: checklist on what is left to test
lag tests
xbox 360
PS2 with PS2 pad
OE both consoles during online matches with delay at 0 on moderate ping connections
FBA (GGPO)
FBA (current version)
ShmupMAME
GroovyMAME
(test all emulators on Windows 7 first, then XP)
timing tests
Q SA1
hold up
programmable controller (input combos with strict timing on CPS3 and see if that timing works on OE and emulator)
Isotopez has been on vacation. we’ll get testing some of this stuff here in the next couple weeks. if I missed anything important or if you have other ideas for input lag/timing differences that need to be tested, let me know. also @d3v I’d like to request that the “About the offline delay” thread be de-sticky’d and this one be sticky’d to replace it.
Stickied.
360 version patch is coming Thursday.
Here’s hoping this can get tested ASAP.
It was mentioned in the 3s group that the dc version has input lag (3 frames), anything to that statement?
Sounds frivilous but, any chance of testing DC versus arcade? Ill test if I have too, just trying to save some time since op already has the set up to do proper testing.
Or is there some real data around about the dc version have 3 more frames inherent lag?
we’ll try and pick up a Dreamcast this week. right now our attention is on getting our cabinet to actually function correctly, but we’ll go back to testing this stuff after that and can do Dreamcast too.