GeDoSaTo Dynamic FPS Capping

Everything about latency. Tips, testing methods, mouse lag, display lag, game engine lag, network lag, whole input lag chain, VSYNC OFF vs VSYNC ON, and more! Input Lag Articles on Blur Busters.
mynm
Posts: 40
Joined: 23 May 2015, 06:36

Re: GeDoSaTo Dynamic FPS Capping

Post by mynm » 06 Jun 2015, 06:42

Hi, I have a problem when I cap fps, but I don’t know if it’s a problem or is normal. I see double image in motion. For example in games, when I move de camera around, when I move de camera with A and D in first person games, or with a lateral cam for example in cars.
I see that with all the programs I have probe to cap fps, included with the method of de thread.

I see other people with the same problem: http://www.svp-team.com/forum/viewtopic.php?id=1883 or http://www.avsforum.com/forum/40-oled-t ... bling.html

I have proved that in a crt and in some monitors and tvs and y see that in all of them.

For example I see that in the crt with 120hz and 60fps, 100hz-50fps, 60hz-30fps. I see double image and some blur. Buy I don’t see it with 46hz-fps. I can’t probe under that hzs.

I have been doing proves whit one of the lcds, that has CCFL backlighting, and with the brightness at 100% to avoid pwm effects in motion (http://www.tftcentral.co.uk/articles/pu ... lation.htm).

I have proved it at 24hz, 30hz, 42hz, 48hz, 60hz-30fps, 74hz-37fps, and I see that effect at 24hz, 30hz, 60hz-30fps, 74hz-37fps. But less at 42hz, 48hz. Whit less than 42 hz the mouse have some lag, but in some games increasing the DPIs or the sensibility of de mouse at the game, it almost disappears. With 42fps and 60hz I don’t see lag. Compared to the crt the lcd at 60hz is like the crt at 60fps-120hz, but with so much blur at the lcd, and the same at 30hz vs 30fps-60hz.

With the lcd less fps at 60hz or 74hz, or less hz, is more distance between the two images. I see the same at the crt, for example at 120hz and capping it to 96, 60, 48, 42, 30fps.

Whit the lag at less than 42hz but no whit 30fps or 42fps and 60hz, I understand that for example, at 30fps 60hz the game is refreshing at 16.6 ms, doubling the image, and the lag is cause of the less fps (less fps, less information, more reaction time), no because of the refreshing. 30hz has a 33.3 refresh, and that causes the lag. I’m wrong about that?.

I see that you all at the thread are experts in that thinks, I don’t see better place to do the question.

Edit: About the double image, the gsync do that: " doubling the refresh rate and inserting duplicate frames starting at around the 37 FPS mark" : http://www.pcper.com/reviews/Graphics-C ... ies-Differ. I don't have a gsync monitor, and I have a amd graphic card. Is the driver doing that?, or it may by the monitor?.

Thanks and regards.

Sparky
Posts: 682
Joined: 15 Jan 2014, 02:29

Re: GeDoSaTo Dynamic FPS Capping

Post by Sparky » 06 Jun 2015, 14:33

mynm wrote:Hi, I have a problem when I cap fps, but I don’t know if it’s a problem or is normal. I see double image in motion. For example in games, when I move de camera around, when I move de camera with A and D in first person games, or with a lateral cam for example in cars.
I see that with all the programs I have probe to cap fps, included with the method of de thread.

I see other people with the same problem: http://www.svp-team.com/forum/viewtopic.php?id=1883 or http://www.avsforum.com/forum/40-oled-t ... bling.html

I have proved that in a crt and in some monitors and tvs and y see that in all of them.

For example I see that in the crt with 120hz and 60fps, 100hz-50fps, 60hz-30fps. I see double image and some blur. Buy I don’t see it with 46hz-fps. I can’t probe under that hzs.

I have been doing proves whit one of the lcds, that has CCFL backlighting, and with the brightness at 100% to avoid pwm effects in motion (http://www.tftcentral.co.uk/articles/pu ... lation.htm).

I have proved it at 24hz, 30hz, 42hz, 48hz, 60hz-30fps, 74hz-37fps, and I see that effect at 24hz, 30hz, 60hz-30fps, 74hz-37fps. But less at 42hz, 48hz. Whit less than 42 hz the mouse have some lag, but in some games increasing the DPIs or the sensibility of de mouse at the game, it almost disappears. With 42fps and 60hz I don’t see lag. Compared to the crt the lcd at 60hz is like the crt at 60fps-120hz, but with so much blur at the lcd, and the same at 30hz vs 30fps-60hz.

With the lcd less fps at 60hz or 74hz, or less hz, is more distance between the two images. I see the same at the crt, for example at 120hz and capping it to 96, 60, 48, 42, 30fps.

Whit the lag at less than 42hz but no whit 30fps or 42fps and 60hz, I understand that for example, at 30fps 60hz the game is refreshing at 16.6 ms, doubling the image, and the lag is cause of the less fps (less fps, less information, more reaction time), no because of the refreshing. 30hz has a 33.3 refresh, and that causes the lag. I’m wrong about that?.

I see that you all at the thread are experts in that thinks, I don’t see better place to do the question.

Edit: About the double image, the gsync do that: " doubling the refresh rate and inserting duplicate frames starting at around the 37 FPS mark" : http://www.pcper.com/reviews/Graphics-C ... ies-Differ. I don't have a gsync monitor, and I have a amd graphic card. Is the driver doing that?, or it may by the monitor?.

Thanks and regards.
That would be low persistence vs a moving eye. If your eye is tracking something on screen, and it's displayed in the same place twice(because framerate is lower than refresh rate), then you'll see two images of it. Higher persistence means motion blur. PWM artifacts can look similar, as the backlight pulsing isn't synchronized with the framerate.

mynm
Posts: 40
Joined: 23 May 2015, 06:36

Re: GeDoSaTo Dynamic FPS Capping

Post by mynm » 06 Jun 2015, 17:08

Sparky wrote:That would be low persistence vs a moving eye. If your eye is tracking something on screen, and it's displayed in the same place twice(because framerate is lower than refresh rate), then you'll see two images of it. Higher persistence means motion blur. PWM artifacts can look similar, as the backlight pulsing isn't synchronized with the framerate.
Thanks for the reply.

Yes, I have readed that even whit low persistence displays even beyond 75hz it shows stroboscopic effects, if I have understood what you have said and that: forums.blurbusters.com/viewtopic.php?f=7&t=333 . But I'm not sure if is that stroboscopic effects or if it's some thing like 3:2 pulldown like in movies. With the crt I don't see stroboscopic effects at 46hz.
And for example whit 24fps and 74hz in the lcd (edited:that's wrong is only duplicated in LCDs) 40fps 120hz in the crt I see three times the image.

Other thing is that lcds are sample and old and it don't have flickering, if it's missing a photogram every 16.6ms like shows the image of the first post in the capped line, it may have flickering, or it olds the frame all the 33.3ms with out refreshing it every 16.6ms?.

Excuse me if I din't had understood what you have said of the low persistence, I'm not an expert in that technical things and I'm not English native speaker.
Last edited by mynm on 09 Jun 2015, 07:00, edited 2 times in total.

Sparky
Posts: 682
Joined: 15 Jan 2014, 02:29

Re: GeDoSaTo Dynamic FPS Capping

Post by Sparky » 06 Jun 2015, 18:23

mynm wrote:
Sparky wrote:That would be low persistence vs a moving eye. If your eye is tracking something on screen, and it's displayed in the same place twice(because framerate is lower than refresh rate), then you'll see two images of it. Higher persistence means motion blur. PWM artifacts can look similar, as the backlight pulsing isn't synchronized with the framerate.
Thanks for the reply.

Yes, I have readed that low persistence may cause that under 75hz, if I have understood that: forums.blurbusters.com/viewtopic.php?f=7&t=333 . But I'm not suar if is that persistence or if it's some thing like 3:2 pulldown like in movies. In the crt I don't see low persistence at 46hz, I don't know if the low persistence occurs only for lcds. And for example whit 24fps and 74hz in the lcd or 40fps 120hz in the crt I see three times the image.

Other thing is that lcds are sample and old and it don't have flickering, if it's missing a photogram every 16.6ms like shows the image of the first post in the capped line, it may have flickering, or it olds the frame all the 33.3ms with out refreshing it every 16.6ms?.
CRTs are typically very low persistence. Sample and hold is 100% persistence, and results in motion blur equivalent to the distance your eye moves across the screen in the duration of 1 frame. Here's a demonstration of the effect that you can try on different monitors: http://www.testufo.com/#test=eyetracking

As for PWM artifacts on a sample and hold monitor, instead of blur, you get several discrete images, as the backlight is pulsed on the same image while your eye is moving across it.

mynm
Posts: 40
Joined: 23 May 2015, 06:36

Re: GeDoSaTo Dynamic FPS Capping

Post by mynm » 07 Jun 2015, 06:26

Sparky wrote:That would be low persistence vs a moving eye. If your eye is tracking something on screen, and it's displayed in the same place twice(because framerate is lower than refresh rate), then you'll see two images of it. Higher persistence means motion blur. PWM artifacts can look similar, as the backlight pulsing isn't synchronized with the framerate.
Yes it’s some thing like this test: http://www.testufo.com/#test=eyetracking&pattern=lines2, but it isn’t it because in games at 60hz or 50hz I don’t see that effect (I don’t move in games the mouse or with A and D the camera, as fast as that images moves), but at 60fps and 120hz or 50fps-100hz, I see double imege, and at 24fps -72hz triple image, (edited:that's wrong is only duplicated in LCDs).

It isn’t PWM artifacts because I put the back light to 100%, if y put it for example at 20% I see like five repetitions of the images two times xD. At 100% I see the image two times and blurred.

Doing proves whit the crt, at that test I see double image with 46hz, I see two ufos and lines, I don’t know if it’s a failure of the test. Is like stuttering but the test says that the sync is ok.

And Mark Rejhon says at de post # 42 of that thread http://www.avsforum.com/forum/40-oled-t ... ing-2.html

That:
As the eye tracks across the screen during video game motion, the repeated frame, creates a doubling effect. I also see the doubling effect during fast pans on old 35mm films at the movie theaters, since old 35mm projectors flashed the projector light at 48Hz. Pans in scenery have the same 'doubling' effect.
So it is a doubling effect, caused by a repeated frame.

So I would like some technical explanation of what is causing that effect, if it’s the monitor, or the graphic card. Because I don’t see any article that explains that effect, I only see posts saying that it’s true, and other saying it isn’t. And in other forums that I have done the question no one see it.
Last edited by mynm on 09 Jun 2015, 07:01, edited 1 time in total.

Sparky
Posts: 682
Joined: 15 Jan 2014, 02:29

Re: GeDoSaTo Dynamic FPS Capping

Post by Sparky » 07 Jun 2015, 11:43

mynm wrote:
Sparky wrote:That would be low persistence vs a moving eye. If your eye is tracking something on screen, and it's displayed in the same place twice(because framerate is lower than refresh rate), then you'll see two images of it. Higher persistence means motion blur. PWM artifacts can look similar, as the backlight pulsing isn't synchronized with the framerate.
Yes it’s some thing like this test: http://www.testufo.com/#test=eyetracking&pattern=lines2, but it isn’t it because in games at 60hz or 50hz I don’t see that effect (I don’t move in games the mouse or with A and D the camera, as fast as that images moves), but at 60fps and 120hz or 50fps-100hz, I see double imege, and at 24fps -72hz triple image.

It isn’t PWM artifacts because I put the back light to 100%, if y put it for example at 20% I see like five repetitions of the images two times xD. At 100% I see the image two times and blurred.

Doing proves whit the crt, at that test I see double image with 46hz, I see two ufos and lines, I don’t know if it’s a failure of the test. Is like stuttering but the test says that the sync is ok.

And Mark Rejhon says at de post # 42 of that thread http://www.avsforum.com/forum/40-oled-t ... ing-2.html

That:
As the eye tracks across the screen during video game motion, the repeated frame, creates a doubling effect. I also see the doubling effect during fast pans on old 35mm films at the movie theaters, since old 35mm projectors flashed the projector light at 48Hz. Pans in scenery have the same 'doubling' effect.
So it is a doubling effect, caused by a repeated frame.

So I would like some technical explanation of what is causing that effect, if it’s the monitor, or the graphic card. Because I don’t see any article that explains that effect, I only see posts saying that it’s true, and other saying it isn’t. And in other forums that I have done the question no one see it.
If refresh rate is higher than frame rate, frames are being repeated. The alternative would be some really awful flicker.

mynm
Posts: 40
Joined: 23 May 2015, 06:36

Re: GeDoSaTo Dynamic FPS Capping

Post by mynm » 07 Jun 2015, 12:06

Sparky wrote:If refresh rate is higher than frame rate, frames are being repeated. The alternative would be some really awful flicker.
No I play with vsync, and if at 50fps-100hz or 60fps-120hz with the crt, with no repeated frames, the flicker will be like at 50hz or 60hz, and it isn't.

Glide
Posts: 280
Joined: 24 Mar 2015, 20:33

Re: GeDoSaTo Dynamic FPS Capping

Post by Glide » 07 Jun 2015, 13:26

mynm wrote:
Sparky wrote:If refresh rate is higher than frame rate, frames are being repeated. The alternative would be some really awful flicker.
No I play with vsync, and if at 50fps-100hz or 60fps-120hz with the crt, with no repeated frames, the flicker will be like at 50hz or 60hz, and it isn't.
If you are displaying 50 FPS at 100Hz, 60 FPS at 120Hz, or 30 FPS at 60Hz etc. each frame is being displayed twice.

With a flicker-free display, this should make no difference.
With a display which has PWM flicker or a low-persistence display, you will see double-images.

mynm
Posts: 40
Joined: 23 May 2015, 06:36

Re: GeDoSaTo Dynamic FPS Capping

Post by mynm » 07 Jun 2015, 15:44

Glide wrote:If you are displaying 50 FPS at 100Hz, 60 FPS at 120Hz, or 30 FPS at 60Hz etc. each frame is being displayed twice.

With a flicker-free display, this should make no difference.
With a display which has PWM flicker or a low-persistence display, you will see double-images.
Thanks for the reply.

Yes to all except that with the back light brightness at 100% my monitors don't flick, and I see the double image. It isn't caused by the back light. I don't know what is duplicating the frames to avoid the flicker, the monitor or the graphic card?, what is the technical explanation of it?, do you have any article that explains it?.

Edited: 50 FPS at 100Hz, 60 FPS at 120Hz are at CRT I don't have a LCD of 120hz it only haves 74hz.
At the LCD I see the same double image at 30hz and at 30fps at 60hz, but with more blur and lag at 30hz.
And I see less double image at 42hz at the LCD (no duplicated frames but some double image because of the less responsiveness) than at 50 FPS at 100Hz or 60 FPS at 120Hz at the CRT (duplicated frames and no blur).

Edited:
Some examples of points in games where I see It very clear are:

In Mirror's edge at the intro chapter when you start the game, in the lines of that wall: http://i.imgur.com/DQSr4mG.png
Or at the lines at the ground: http://i.imgur.com/MjQQzpP.png
Moving the camera only with A and D, with out the mouse.

Or in GTAIV at this point: http://i.imgur.com/hBkxDpQ.png http://i.imgur.com/byqGEsy.png
Moving the camera with a game pad or the mouse (more with the pad I don't know why, with the mouse I see it better at around 1000 DPIs) I see that smokestack duplicated (50fps 100hz), triplicated (40fps 120hz) or quadruplicated (25fps 100hz).

Glide
Posts: 280
Joined: 24 Mar 2015, 20:33

Re: GeDoSaTo Dynamic FPS Capping

Post by Glide » 08 Jun 2015, 17:27

Looking at the Mirror's Edge example (because I own that game) I'd suggest that if you're sure that your display is 100% flicker-free, it may be response time issues on the LCD panel or simply the fact that 30 FPS is inadequate for the speed of motion with that high contrast scene.

If there's no flicker at all, 30 FPS at 60Hz should be identical to 30 FPS at 30Hz.

I will say that I did see judder when testing in the training sequence of Mirror's Edge at 30 FPS, but this got me to do some further investigating, and while my display does not use PWM, it turns out that there is a subtle flicker at 60Hz - which is enough to reduce persistence to about 12ms using this test.

I was able to detect this flicker using TFT Central's camera test where you take a photograph at a fixed shutter speed and see if there are any divisions in the line. At 1/30s there were two lines which are pure white, so the flicker frequency is 60Hz and not PWM. (and in other modes on the TV it flickers at 120/240Hz - again, not PWM)


Unless you're using a G-Sync display (and perhaps even then?) anything less than your refresh rate is generally going to be a bad experience for one reason or another.

Post Reply