RTSS vs in-engine framerate capping

[RealNC]

I don't know why the keyboard should be able to know anything about the monitor and vice versa

Things are measured differently, obviously. I assume the keyboard's display is fed by something like RTSS or similar software which measures the game's output, while the monitor measures the g-sync module's scan rate. These are very different things that are measured and will never match. Just because the game outputs a frame doesn't mean the driver will send it to the monitor at that exact time, or that the g-sync module will show it at that time.

Measurements done at different stages of the output chain will give you different results.

[tygeezy]

I don't know why the keyboard should be able to know anything about the monitor and vice versa

Things are measured differently, obviously. I assume the keyboard's display is fed by something like RTSS or similar software which measures the game's output, while the monitor measures the g-sync module's scan rate. These are very different things that are measured and will never match. Just because the game outputs a frame doesn't mean the driver will send it to the monitor at that exact time, or that the g-sync module will show it at that time.

Measurements done at different stages of the output chain will give you different results.

im actually using the same software for both. MSI afterburner plus rtss. In the afterburner settings there is an option to display on osd and keyboard and I have both enabled. If I get some time I'll use shadow play to record a sessions n and then do a similar run with my camera pointed at keyboard so I can show you guys what's going on.

Perhaps there's a better and more accurate way to show frametimes? I heard newest afterburner plus rts has a graph.

[Chief Blur Buster]

Unfortunately, they tend to be very laggy modes, even worse during 1/3 and 1/4.[

Our preferred technique is frame capping instead.

Capping will result in stutter (on non-GSYNC/FreeSync displays), even if the cap is half the Hz. These vsync modes are the only way to get smooth animation.

Note that there's no input lag increase with 1/2 vsync at 120Hz compared to normal vsync at 60Hz. In fact, it's a bit lower due to the faster scanout speed.

And G-Sync is not an option with 3D vision. It's one of the other. If you want 30FPS with 3D vision, 1/2 vsync is the only way to do it without microstutter.

All of this be true (also added note, for clarification).

[RealNC]

It was a response to this post:

How about 1/4 vsync and 1/3 vsync? I see they have that for NVidia inspector. Do those work? Or should you really only use full vsync and half? Playing games at 15 and 20 fps would be balls, but I can see those options being useful on high refresh rate monitors that don't have gsync.

I have a 144 hz gsync monitor by the way, but I also use the 3d vision on my monitor, so I still need to mess with vsync if I use 3d vision. And I hate input lag, so I've been looking for the best option to have minimal input lag while also having a smooth stutter free image.

So the topic is about non-gsync vsync

[jorimt]

Source engine framerate cap seems to be particularly inaccurate. If I cap with rtss at 140 I get a lock 7.14 ms pretty much locked. With the in game capper at 140 the frametimes bounce between 9 and 11 with it never going below 9 ms.

Dishonored 2 has a great framerate capper. It doesn't have as many options as other cappers, but the rates it can set seem to be more accurate. I settled in on 75 fps with some graphic tweak to hit that rate at all times and it's locked at 13.33 ms frametimes.

As I said in my initial reply, frametime consistency isn't necessarily the ultimate goal with an in-game limiter. Again, just because RTSS has steadier frametimes, doesn't make it superior, it is simply a side-effect of its functionality.

Yes, CS:GO has a quirky limiter, but it's still lower latency than RTSS. As for Dishonored 2's in-game limiter, while it isn't as flexible, it should also be superior to RTSS in input latency levels.

Also, don't get caught on perfect frametime output, at least if it is at or fluctuating above (lower frametimes) your target framerate. Let's say you have a 60 FPS in-game limit, and you watch the frametime meter fluctuate between 16.6ms (the target frametime), and, say, 12ms, 14ms, 8ms, 16.6ms, etc. Frametime is separate from scanout time. Frametime dictates how long it takes a frame to be rendered and ready for scanning in. It could theoretically sit there forever until the display decides to refresh and start scanning it in.

In other words, as long as the fluctuating frametimes you are seeing are lower than your set framerate limit, they will be displayed in the desired intervals regardless if they finish earlier or not. The only thing lower fluctuating frametimes with in-game limiters can cause is less input lag.

@jorimt Would frame times that are faster than your monitors refresh rate cause tearing if you have gsync on vsync off and input lag if you were gsync on vsync on?

Have you read my article in full? It has all of those answers in detail.

But short answer, yes, G-SYNC + V-SYNC "Off" can tear if frametimes exceed the max refresh rate target, but it can also tear if the framerate drops suddenly with that combination. This is why you need G-SYNC + V-SYNC "On," which compensates for sudden shift in frametime. The only "lag" it adds over the other combo is lack of tearing.

So the topic is about non-gsync vsync

I was confused about that myself, but I think the OP is switching between both.

[tygeezy]

Source engine framerate cap seems to be particularly inaccurate. If I cap with rtss at 140 I get a lock 7.14 ms pretty much locked. With the in game capper at 140 the frametimes bounce between 9 and 11 with it never going below 9 ms.

Dishonored 2 has a great framerate capper. It doesn't have as many options as other cappers, but the rates it can set seem to be more accurate. I settled in on 75 fps with some graphic tweak to hit that rate at all times and it's locked at 13.33 ms frametimes.

As I said in my initial reply, frametime consistency isn't necessarily the ultimate goal with an in-game limiter. Again, just because RTSS has steadier frametimes, doesn't make it superior, it is simply a side-effect of its functionality.

Yes, CS:GO has a quirky limiter, but it's still lower latency than RTSS. As for Dishonored 2's in-game limiter, while it isn't as flexible, it should also be superior to RTSS in input latency levels.

Also, don't get caught on perfect frametime output, at least if it is at or fluctuating above (lower frametimes) your target framerate. Let's say you have a 60 FPS in-game limit, and you watch the frametime meter fluctuate between 16.6ms (the target frametime), and, say, 12ms, 14ms, 8ms, 16.6ms, etc. Frametime is separate from scanout time. Frametime dictates how long it takes a frame to be rendered and ready for scanning in. It could theoretically sit there forever until the display decides to refresh and start scanning it in.

In other words, as long as the fluctuating frametimes you are seeing are lower than your set framerate limit, they will be displayed in the desired intervals regardless if they finish earlier or not. The only thing lower fluctuating frametimes with in-game limiters can cause is less input lag.

@jorimt Would frame times that are faster than your monitors refresh rate cause tearing if you have gsync on vsync off and input lag if you were gsync on vsync on?

Have you read my article in full? It has all of those answers in detail.

But short answer, yes, G-SYNC + V-SYNC "Off" can tear if frametimes exceed the max refresh rate target, but it can also tear if the framerate drops suddenly with that combination. This is why you need G-SYNC + V-SYNC "On," which compensates for sudden shift in frametime. The only "lag" it adds over the other combo is lack of tearing.

So the topic is about non-gsync vsync

I was confused about that myself, but I think the OP is switching between both.

got it, in using gsync + vsync for what it's worth. I've read a lot of that article but admittedly not cover to cover; that's my bad.


Did you see my question regarding the discrepancy in frametimes between my keyboard LCD and monitor display from rtss?

If I can sneak away and get some time I'll try and record the footage. Essentially I'm seeing some large frametime spikes that my LCD display picks up, but the osd for my monitor doesn't. I'll hit a 33.33 ms frametime spike that my keyboard displays but my monitor will say 7-8 ms.

[jorimt]

All that matters is what you see/feel. Are you seeing a short halt or pause during these 33.3ms spikes or no? Again, you shouldn't put much stock in frametime numbers, regardless of the readout you're looking at.

You'll know when you hit a frametime spike bye eye.

[tygeezy]

All that matters is what you see/feel. Are you seeing a short halt or pause during these 33.3ms spikes or no? Again, you shouldn't put much stock in frametime numbers, regardless of the readout you're looking at.

You'll know when you hit a frametime spike bye eye.

i do feel some of them. Cs go has given me random microstutters dating back to my last system. My i7 860 and has continued on in my 7700 k build. The on screen frametime display doesnt pock it up buy my keyboard does.

[jorimt]

As I stated in the "Range" section of my G-SYNC 101 series, frametime spikes (sudden, very high frametimes) are virtually unavoidable, regardless of your system capabilities or how well you have things setup up. An optimal setup and proper settings will only reduce/shorten their occurrence.

It's likely just CS:GO (level/asset loading, etc; this can also be exacerbated by heavily modded maps).

As for frametime readings, short of FCAT, they are usually hit or miss accuracy-wise.

[tygeezy]

As I stated in the "Range" section of my G-SYNC 101 series, frametime spikes (sudden, very high frametimes) are virtually unavoidable, regardless of your system capabilities or how well you have things setup up. An optimal setup and proper settings will only reduce/shorten their occurrence.

It's likely just CS:GO (level/asset loading, etc; this can also be exacerbated by heavily modded maps).

As for frametime readings, short of FCAT, they are usually hit or miss accuracy-wise.

Okay, I should really go through the entire article from cover to cover before I come back with more questions. Thanks for answering everything I've thrown your guys way and I appreciate all the work you guys have out in. There are tons of false information that seems to float around forums and reddit regarding vsync, gsync ect.

I'm glad it's not just me that gets these random frametime spikes. Although cs go (source engine) appears to be a particularly bad offender. Overwatch is really a lot smoother and more consistent with a better framerate capper. It is newer, so I guess that's to be expected.