Test Input lag: Megadrive vs MiSTer FPGA and also Megadrive vs RGB-Pi... Which one?

[oskarzer0]

Hi!
I've made a couple of videos to measure the response speed of mister and RGB-Pi versus the original system.
Megadrive vs MiSTer FPGA: MiSTer with Antonio Villena Scart´s and XBOX360 gamepad.Pressing the jump buttons at the same time. Same tv and same brightness and contrast settings.
¿Who is who? Place your bets! XD XD

You can put the video in slow motion and pause it if you do not see it clearly
https://youtu.be/jRUipmo6-Qo

This is more easy: Megadrive vs Raspberry con GPIO to SCART cable (http://www.rgb-pi.com/)
https://youtu.be/wdYboQyeYm4

Grettings!

[cacophony]

The first video shows identical results, but that's expected because the framerate of those video is only 30fps and you're bypassing the scaler lag by using analog out. The second video has the pi about 70ms slower to respond.

[Newsdee]

So... this means the USB handling of MiSTer is perfectly fine and not a source of input lag?

Good job on the comparison!

[cacophony]

So... this means the USB handling of MiSTer is perfectly fine and not a source of input lag?

Good job on the comparison!

It just means that the lag from USB handling is less than 2 frames, which is the granularity of the video.

[Newsdee]

Better than the RPi at least.

I use an 8bitdo wireless adapter so I expect lag, it shows up especially when batteries are lower. It's also possible there are crappy USB adapters out there, so some kind of testing suite would be good.

[vanfanel]

On the Pi tests, was max_swapchain=2 used on the RetroArch config? If you use max_swapchain=3 (default) you will get at least one additional frame of input lag.
Also, usbhid.jspoll=2 should be passed to the kernel on cmdline.txt.

That or the tests are not valid, I think.

[oskarzer0]

I don't know. I'm buy cable, download the sd image from their web, and play Alissia Dragon(USA) rom.

[Newsdee]

On the Pi tests, was max_swapchain=2 used on the RetroArch config? If you use max_swapchain=3 (default) you will get at least one additional frame of input lag.
Also, usbhid.jspoll=2 should be passed to the kernel on cmdline.txt.

That or the tests are not valid, I think.

To be fair, if it needs something like that it's quite obscure for the layman... most RPi setups may not be running with that probably.

[Sorgelig]

I more like the fact when at first they talk that 1 frame lag is crucial and then on another page they are talking how ~280ms lag is worse than ~200ms.

I always imagine the picture in this case: When cat has nothing to do it licks his balls..

[vanfanel]

Appart from the max_swapchain=2 used on the RetroArch config and usbhid.jspoll=2 in cmdline.txt, you also need to use the VC4 video stack. The default BCRM video stack stinks, and adds a frame more of latency. I don't know if the video stack in use is relevant for RGB-Pi: Does RGB-Pi rely on writing directly on the framebuffer? What framebuffer? Too many things that are ignored in this test, really.

@Newsdee: Well, something like Lakka could use these settings by default on the Pi, but I don't know/care if it does.

@Sorgelig: True thing

[cacophony]

Not sure where anyone was discussing 200ms vs 280ms, but 200ms of lag would make a lot of these older games impossible

[oskarzer0]

Other video:
MiSTer with simultaneous out HDMI and Scart at left, Megadrive Scart at right, video to 60HZ
https://youtu.be/oqI-gg0txHw

[cacophony]

Other video:
MiSTer with simultaneous out HDMI and Scart at left, Megadrive Scart at right, video to 60HZ
https://youtu.be/oqI-gg0txHw

You're using two controllers and attempting to press the buttons at the same time? (In certain cases the MiSTer CRT jump animation happens before the Megadrive CRT). If that's the case then it's not a very fair MiSTer vs Megadrive comparison. But assuming you had a single MiSTer hooked up to both the CRT and HDMI then that comparison is interesting and reveals a two frame difference. What HDMI monitor is that? Anyway, thanks for sharing the tests!

[oskarzer0]

Yes, i'm use two controllers, megadrive six buttons controller and xbox 360 gamepad, and I always try press at the same time. The ideal would be to be able to use a single controller connected to both systems at the same time, but I do not know how.

I try to show that between the Mister Scart and the original system there are no playable differences, but with hdmi if there are but very small differences. The HDMI monitor is a LG (https://www.lg.com/ae/monitors/lg-M2794DP) and Game mode (Response Time: 2ms (GTG))

[Locutus73]

(Response Time: 2ms (GTG))

Just FYI response time is the pixel response time and it doesn’t have anything to do with display added input lag.

Regards.
Locutus73

[Sorgelig]

this is human's lag:
https://www.humanbenchmark.com/tests/re ... statistics

so, before talk about 1-2 frames lag, fix your own lag first
I think this discussion can be closed already.

the whole discussion is pointless, as real meaning of this discussion is when you already know when you have to react. For example you see moving enemy and you predict the time of collision. But if you don't know when to react, the whole human reaction lag is added which is around 250-300ms.
But you don't take in account of brain automatic training, so predictable reaction is reduced by brain after couple tries. Input controller lag, scaler lag and even human lag - all this is eliminated by brain after short time of play, so you even won't notice it.

[oskarzer0]

But the frame diference in MiSTer with simultaneous out HDMI and Scart is because the hdmi monitor have lag, no?

[Locutus73]

My bad old body is laggy... very laggy.
But the discussion isn’t totally pointless, because input lag is additive and some old games were designed on CRT and quick response time: beating Mike Tyson in NES Punch Out (no predictable movements, just pure reflexes) is a good example of how input lag can make a game unplayable (you can’t actually beat Mike on some emulator setups).

Regards.
Locutus73

[oskarzer0]

Look in attachment, in the original system and in Mister for Scart, Sonic is blue ball and in the led (MiSTer simultaneous HDMI) not yet (1 frame - 2 frame lag)

[Locutus73]

But the frame diference in MiSTer with simultaneous out HDMI and Scart is because the hdmi monitor have lag, no?

Yes plus the added lag by the MiSTer scaler if unused in the VGA out. Maybe if you set vga_scaler=1 in MiSTer.ini you can have a good measure of how is laggy your HDTV compared to a CRT, but Sorgelig should confirm this hypothesis.

Regards.
Locutus73

[Sorgelig]

Look in attachment, in the original system and in Mister for Scart, Sonic is blue ball and in the led (MiSTer simultaneous HDMI) not yet (1 frame - 2 frame lag)

sure. scaler uses tripple buffering to make sure playback is smooth. So, HDMI has 1-2 frames lag. Otherwise it will have tearing.

Maybe if you set vga_scaler=1 in MiSTer.ini you can have a good measure of how is laggy your HDTV compared to a CRT, but Sorgelig should confirm this hypothesis.

yes, it will output the same HDMI video through VGA output, so you can measure the lag of TV itself.

[paulbnl]

this is human's lag:
https://www.humanbenchmark.com/tests/re ... statistics

so, before talk about 1-2 frames lag, fix your own lag first
I think this discussion can be closed already.

the whole discussion is pointless, as real meaning of this discussion is when you already know when you have to react. For example you see moving enemy and you predict the time of collision. But if you don't know when to react, the whole human reaction lag is added which is around 250-300ms.
But you don't take in account of brain automatic training, so predictable reaction is reduced by brain after couple tries. Input controller lag, scaler lag and even human lag - all this is eliminated by brain after short time of play, so you even won't notice it.

The fact is total reaction time = brain lag + total input lag.

The brain only tricks you into perceiving less input lag then there really is. The real reaction time is still longer with higher input lag and games which needs short reaction time like Punch Out will be much more difficult.

Everyone that plays retro games knows that low input lag is needed.

[Sorgelig]

The brain only tricks you into perceiving less input lag then there really is. The real reaction time is still longer with higher input lag and games which needs short reaction time like Punch Out will be much more difficult.

it looks like you didn't read what you've quoted.
In regular predictable reaction brain compensates the WHOLE chain of lags.
In unpredictable reactions adding 2-3 frames of lag on top of ~250ms of human lag makes no difference.

[Sorgelig]

Everyone that plays retro games knows that low input lag is needed.

Yes. And i've played a lot of games on MiSTer using USB wireless pad with HDMI output and don't notice the lag at all.
So, i know what i'm talking about.

I like healthy discussion but i REALLY hate this crappy lag topic which is based only on imagination of hypothetical situation.
Users with lag obsession may go to look for another system. No one keeps you using the MiSTer which has so bad lags making games unplayable for you (and only for you).

[oskarzer0]

Look in attachment, in the original system and in Mister for Scart, Sonic is blue ball and in the led (MiSTer simultaneous HDMI) not yet (1 frame - 2 frame lag)

sure. scaler uses tripple buffering to make sure playback is smooth. So, HDMI has 1-2 frames lag. Otherwise it will have tearing.

Maybe if you set vga_scaler=1 in MiSTer.ini you can have a good measure of how is laggy your HDTV compared to a CRT, but Sorgelig should confirm this hypothesis.

yes, it will output the same HDMI video through VGA output, so you can measure the lag of TV itself.

Ok! Thanks to everyone!!!