That's a great point! Ya it is interesting to consider how much lag is too much. Strikers II is obviously an abomination of lag that should be avoided. I tend to think of it as lag relative to the other games on the console. Like on PC and 360 I would consider 2-3 to be best case scenarios. For whatever reason the Switch really struggles with input lag in general, so for it 4-5 would be considered a responsive game.EmperorIng wrote:Thanks for setting all this up! It's a good short-hand to figure out which ports to avoid, or (in my case with Strikes 2 on Switch) confirming what I could already feel.
Makes me wonder what amount is "acceptable." I know you mentioned in your podcasts that Ikaruga and Danmaku Unlimited 3 are near the top for you in regards to the lack of input lag.
That's a good point, I'll update the controllers and double check to see if that makes a difference. I don't think it could shave more than a frame off the readings, but it's definitely something I'm going to look into just in case.BrianC wrote:Thanks for the input lag table!
I remember reading that Dual Shock 4 controller in wireless mode has very little input lag. I'm not sure which controller on the switch has the least amount of lag, but make sure the Pro Controllers and Joycons are updated. I definitely noticed less lag with the Pro in Vs. SMB after updating the controller's firmware (I had a hard time getting past 6-3 consistently before it was updated). VX-SA is known for having low input lag, so that may be part of the reason some games are coming up with less lag compared to games on the Pro Controller.
Ha ya I had no idea what to put in that column, got it fixed nowmikejmoffitt wrote:This is good data to collate.
Just a small nitpick, but for things running on PCB, you should really indicate that vsync is "on" - virtually every one of these PCB shmup titles runs in lockstep with vsync, and is timed to that interval.
I wrote: On (Hardware)
Yeah I was thinking the same thing. My opinion is that the game's input lag problems don't just stem from the raw lag, so to say, but also how the game is all over the place with its slowdown. I think that's what makes it so frustrating. It basically hits you with unexpected spikes of input lag rather than just being consistently laggy. The port is a hot mess ha.CloudyMusic wrote:I knew that Strikers II on Switch felt bad, but no idea it was that bad.
Once again, it's interesting to see that SDOJ apparently has basically the same lag as some other Cave 360 ports. Maybe it's just other aspects of the game that contribute to it feeling unresponsive beyond just raw input lag.
Oh I should explain that. I am using the VX-SA on PGM in conjunction with the undamned USB decoder, which doesn't add any additional lag (I checked on other platforms).Xer Xian wrote:Thanks for sharing, very interesting!
The Switch looks bad overall, but PS4 games without a lag reduction feature (guess it means vsync off?) probably would as well.
Btw, isn't there an error on the last entry - listing the VX-SA as input for an arcade board?
I did not check the other Darius games, I should have. I basically tested Gaiden because it's my favorite ha. Sadly I am not a baller, so I don't have any other PCB's to test at this moment. However, I do plan on testing a bunch of Saturn games soon, so we'll be able to see how the Saturn games stack up to the Switch ports.GFoyle wrote:This is great as I have been really curious about the input lag on Switch especially. It's clearly a bit more than on other platforms based on those results, but I think it's a bit up to debate how much +1/+2 frames on 60fps really matter in the end. I guess it really depends on the game really.
Would be great to have some PCB comparison for those Psikyo games especially, are those originally in the 2 frames range similar to Ketsui or more, just to see how much Switch + Zerodiv's emulator adds to it. That +1 extra frame compared to others Psikyo games in Strikers 1945 II is a bit strange, so I wonder could it be that the PCB also had a bit worse input lag than the others Psikyo games you tested? I have played those psikyo ports on Switch quite a bit and yeah, that extra frame makes it feel more sluggish compared to the others.
M2 are usually quite a perfectionists so interesting to see 4 frames for Thunder Force but 5 for Darius collection, why the difference (did the original Darius games have more input lag as well?).
I've tested five different hdmi to vga converters with a Leo Bodnar lag tester (sensitivity to the tenth of a millisecond) and none of them showed any added latency.komatik wrote:There's no such thing as "lagless" HDMI -> VGA: it's not a simple pin conversion like HDMI -> DVI, the chip in that cable needs to decode the digital signal and reencode it into analog and that process is not instantaneous.
This is news to me. Not saying it's impossibile, but in my experience (which is also commonly shared) a typical CRT PC monitor is lagless (or as close to lagless as it can be).A typical '90s multiscan VGA computer monitor is not 100% guaranteed to be lagless either: some crappy screens were known to introduce a frame of delay depending on what resolution and refresh rate combo you fed them.
OK well I mean I've tested some converters too and they do. It's entirely up to who made the converter and what input format you're converting to what output format. (1080i vs 720p vs whatever -> 1024x768@60 etc). My point is that if you're talking numbers down to 1/60th of a second, just saying "I tested it" without explaining how it was tested is glossing over important details. I'm not saying the specific HDMI>VGA converter he's using is guaranteed to introduce extra lag, just that some converters can, and he should cover all the bases by explaining his methodology completely. There's obviously a big difference between testing with something like a Leo Bodnar vs testing by eyeballing it with LEDs and a slowmo camera or whatever.Xer Xian wrote:I've tested five different hdmi to vga converters with a Leo Bodnar lag tester (sensitivity to the tenth of a millisecond) and none of them showed any added latency.
In theory it should be, but I have seen lesser-quality monitors that do weird stuff when presented with certain signals and for all intents and purposes end up with what's effectively a 1 frame buffer due to unnecessarily bouncing the signal around internally or something. Although to be fair I believe this is more common on later-model monitors that have an OSD for changing settings vs older monitors with physical dials. (Note that I'm specifically only talking about crappy PC desktop monitors here, not arcade CRTs or other stuff).Xer Xian wrote:This is news to me. Not saying it's impossibile, but in my experience (which is also commonly shared) a typical CRT PC monitor is lagless (or as close to lagless as it can be).
Hey man thanks for the feedback. So for both the Switch and the PS4 I used both the HD Fury cable, as well as a low latency ASUS gaming monitor to double check my results. I am aware that the Gamerfury must add some small amount of lag, but not enough to change the results compared to native VGA. I did a test comparing the two and did not see any changes in my readings.komatik wrote:As someone who works in AV and IT, you really REALLY need to post waaaaaay more information about your PC's drivers, your monitor, and how you tested that HDMI-VGA converter. There's no such thing as "lagless" HDMI -> VGA: it's not a simple pin conversion like HDMI -> DVI, the chip in that cable needs to decode the digital signal and reencode it into analog and that process is not instantaneous. A typical '90s multiscan VGA computer monitor is not 100% guaranteed to be lagless either: some crappy screens were known to introduce a frame of delay depending on what resolution and refresh rate combo you fed them. Different games can invoke video drivers in different ways and force things like triple buffering without you knowing it. USB has inherent polling lag you need to pay attention to.
I'm not saying this to defend the Switch's terrible numbers (I've personally noticed it being bad when playing platformers) or any particular game, it's just that whenever people compare lag they always end up throwing in extra variables they don't entirely control for and make lots of assumptions that don't hold true nearly as often as folks think they should. This is a good first step but you really need to do your due diligence and cover all the details, because when people start arguing about 1/30th of a second differences the details become make-or-break.
I would also strongly encourage you to find a TV that all these systems can connect to natively and double check your numbers. Obviously if the TV has its own internal lag all the numbers will be higher but I wouldn't be surprised to see some of them not rise as much as others.
Yes the pro controller certainly is a sticking point for the Switch. I do plan on testing more input methods on Switch in the future to see if they could possibly improve the results. But as the pro controller is a popular input method, and really the only one I had access to at a reasonable price at the time (I'm not dropping $200 for the hori fightstick just yet ha) I think it's a good starting place.BrianC wrote:I heard that the VX-SA is adds very little input lag (rated 0.0 at Teyah.net), the fighting commander PS4 adds about 4.52ms, and the Switch Pro could add over a frame of input lag (not sure if the tests I read were done with an updated controller, where there was definitely a noticeable difference in input lag). Then again, I think the test for the Switch controller I read combined it with the input lag of the tv screen. I did hear that the Psyiko ports are laggy, but some of the numbers seem slightly high compared to the numbers on PC.
Is it possible to directly compare some of the Cave PC ports to the 360 versions? I'm curious if they add any extra lag compared to the console ports and how they fare with V-Sync on and off.
Tonight I am going to test Shmuparch/Retroarch games on switch, so I'm super super curious how that will compare to shmuparch tests on PC. I think these results can really help us figure out if the delay is somewhere in the hardware itself, or if its a software issue.GFoyle wrote:I think terrible is such strong word to use in this context and I would hope people wouldn't use this easily and label the problem really much worse than it really is. Even in the Strikers 1945 2 on Switch, it doesn't make it unplayable and there are a lot of people who really haven't noticed anything strange at all after playing it a lot.. Of course less input lag, the more fluid and nice it feels, but I still can go and play some original DoDonPachi in local arcade or Ketsui on my PS4 and then some Strikers on Switch and still enjoy it. I do wish it would be better though.
There are much worse examples though, like Sega Classics Collection on Switch which has really bad input lag on some/most games. I guess everyone has their own threshold where they start to be bothered.
Anyway, based on those numbers it looks like Switch would have couple of frames more input lag than other setups by default and then any extra is more dependent on game or emulator running it.
OK good, that removes some variables from the equation. You should copy/paste this paragraph into your page where you talk about your setup and the cable.Mark_MSX wrote:So here's how I tested the HD fury cable: my computer has HDMI and native VGA output. So I connected my CRT monitor to the native VGA, did a frame test on a shmup (ZeroRanger), and then swapped over to the HDMI input using the HD Fury cable and got the same result as the native VGA, no additional latency.
Yeah that doesn't surprise me in the least. I've yet to see any LCD screen that can match a CRT. Honestly I'm surprised it was even that low.Mark_MSX wrote:Compared to the CRT, the monitor did appear to be adding a slightly higher bit of lag, about half a frame, so using the HD Fury cable is faster than using a low latency HDMI monitor.
Honestly it's probably not if my personal experience is any evidence. My friend has a Switch and I noticed immediately that it had really severe lag as soon as I picked up the controller.Mark_MSX wrote:I don't think the extra delay in the Switch results is due to my equipment at all.
Is there any way to get a wired controller into the Switch or is it strictly bluetooth only?Mark_MSX wrote:Though it might be possible that the Switch results could be improved with a different controller or something. That's something I'm looking into right now actually.
OK wait I'm confused, that's the LCD. Did you actually use that for anything because your tests only talk about the CRT.Mark_MSX wrote:My monitor is the ASUS MX279H.
PLEASE PLEASE PLEASE know what you're doing when configuring your settings. All of the different frame-delay and runahead stuff can act very differently on different hardware and mix in a bunch of variables that make any comparisons useless. (Unless you're specifically going for a lowest-lag-possible setup... but that should be sorted separately from the rest of the results).Mark_MSX wrote:Tonight I am going to test Shmuparch/Retroarch games on switch, so I'm super super curious how that will compare to shmuparch tests on PC.
Apparently, the Pro controller in wireless mode has less lag than wired USB mode, and about the same as a Gamecube controller (via USB adapter: so the fact that the Pro controller is slower in USB mode can't be attributed to slower USB polling rates on the console end).Mark_MSX wrote:Yes the pro controller certainly is a sticking point for the Switch. I do plan on testing more input methods on Switch in the future to see if they could possibly improve the results. But as the pro controller is a popular input method, and really the only one I had access to at a reasonable price at the time (I'm not dropping $200 for the hori fightstick just yet ha) I think it's a good starting place.
On a no-compromise setup yeah, or at least with only a tiny mere ms difference remaining.donluca wrote:Also, please have a GroovyMAME setup up and running as it will give you the closest result possible to the original PCB.
If your hardware is strong enough, on older games you can use a high frame delay and get the exact same lag as the PCB, provided you use a CRT and a low lag input device.
Strikers1945guy wrote:"Do we....eat chicken balls?!"
Speaking about the xbox 360 numbers being a frame too high, this might be possible. It might be possible that all my results could be a frame too slow for some reason based on the way the light is wired or something. I'm not exactly sure how that could be possible, but I'm not closed off to the idea. One thing that I am certain of, though, is that the 360 port of Ketsui is a frame slower than Retroarch and PGM. So it could be possible that the PGM is only 1 frame and the 360 port is 2 frames. If this is the case in some way, the good news is that numbers are still consistent and can be used to compare the lag across the games I've tested.Xyga wrote:Something is bugging me, looking at the 360 figures in particular, they're like 1 frame more than it feels in practice.
Testing lag isn't easy in that additional ms quickly add up, stick, firmware, adapter, drivers/polling, reference location on the screen, and there can be several more surprises.
Such database is very useful but to keep it healthy it's good to keep questioning the method, hardware and per-game results over and over until no doubt remains.
Anyway nothing prevents you from polishing it up over time.
On a no-compromise setup yeah, or at least with only a tiny mere ms difference remaining.donluca wrote:Also, please have a GroovyMAME setup up and running as it will give you the closest result possible to the original PCB.
If your hardware is strong enough, on older games you can use a high frame delay and get the exact same lag as the PCB, provided you use a CRT and a low lag input device.
The CPU requirements both from MAME and Groovy have been going up lately though, for a number of systems that are now better emulated, but also iirc Groovy that's currently in the process of getting more accurate (WIP).
This requires at least a quad-core that can clock over 4GHz easily (and a not-too-shitty AMD GPU)
PS: a LCD is fine too as long as it's lagless, and there are quite a number out there despite what people believe, a lot of still don't get the difference between 'input lag' proper and the variables at different measurement positions/conditions and a small number of factors that reviews websites don't really advertise or explain enough in details.
Yes I've heard this as well! I did my initial batch of tests with the Pro controller plugged into the dock. However, I'm going to retest some games in wireless mode and if there is in improvement, I'll add the extra results to the Switch games.Xer Xian wrote:Apparently, the Pro controller in wireless mode has less lag than wired USB mode, and about the same as a Gamecube controller (via USB adapter: so the fact that the Pro controller is slower in USB mode can't be attributed to slower USB polling rates on the console end).Mark_MSX wrote:Yes the pro controller certainly is a sticking point for the Switch. I do plan on testing more input methods on Switch in the future to see if they could possibly improve the results. But as the pro controller is a popular input method, and really the only one I had access to at a reasonable price at the time (I'm not dropping $200 for the hori fightstick just yet ha) I think it's a good starting place.
