Often there is a similar argument online stating that if your FPS can only reach 120, there is no need to buy a 240Hz monitor.
However, theoretically, a shorter refresh cycle also represents a lower display latency. If Vsync is turned off, then tearing should occur randomly on the screen. Considering visibility, on average, it should be detected by the human eye when scanning halfway. Excluding issues such as frame queuing, the latency is the Frametime plus half the scan-out time.
Vertical sync enabled means rendering has to wait for the scan out to finish, thus introducing a delay of 1-2 refresh cycles
full latency is frametime plus 1.5 to 2.5 scan-out time.
G-sync can eliminate the delay in frame waiting, but frame tearing is not allowed, so on average, it will be half a scan more than vsync. Therefore, the total latency is one scan-out time.
Since latency is always related to the scan-out time, a higher refresh rate will undoubtedly reduce latency, even if the frame rate does not meet standard FPS.
Additionally, a higher refresh rate represents a larger VRR range and less tearing (even at lower frame rates), so a higher refresh rate is always beneficial, is my understanding correct?
Is a refresh rate higher than the frame rate meaningful?
-
lann
- Posts: 11
- Joined: 16 Jun 2024, 03:02
Is a refresh rate higher than the frame rate meaningful?
Last edited by lann on 05 Dec 2024, 06:00, edited 1 time in total.
- RealNC
- Site Admin
- Posts: 4589
- Joined: 24 Dec 2013, 18:32
- Contact:
Re: Is a refresh rate higher than the frame rate meaningful?
Yes, but there's diminishing returns. Going from 60Hz to 120Hz provides a good improvement (up to 8ms), but going from 120 to 240 only has a minor impact (it's up to 4ms.) When going from 240 to 480 it's even smaller (just 2ms.)
The biggest latency improvement comes from capping your FPS. On a 144Hz display, if you cap FPS to 120, you'll get a huge improvement. You need two things to have the lowest latency possible with g-sync + vsync: a) Prevent FPS from going near max Hz, and b) prevent the GPU from getting fully saturated. If the game has Nvidia Reflex implemented, then it takes care of both of those automatically. Without Reflex, you need to take care of it yourself.
These are much more important than having a higher refresh rate.
The biggest latency improvement comes from capping your FPS. On a 144Hz display, if you cap FPS to 120, you'll get a huge improvement. You need two things to have the lowest latency possible with g-sync + vsync: a) Prevent FPS from going near max Hz, and b) prevent the GPU from getting fully saturated. If the game has Nvidia Reflex implemented, then it takes care of both of those automatically. Without Reflex, you need to take care of it yourself.
These are much more important than having a higher refresh rate.
Steam • GitHub • Stack Overflow
The views and opinions expressed in my posts are my own and do not necessarily reflect the official policy or position of Blur Busters.
The views and opinions expressed in my posts are my own and do not necessarily reflect the official policy or position of Blur Busters.
-
lann
- Posts: 11
- Joined: 16 Jun 2024, 03:02
Re: Is a refresh rate higher than the frame rate meaningful?
I still have a few questions.
After turning off Vsync, assuming the refresh rate matches the frame rate, and if the tear line appears at the center of the screen, the upper half would have a delay equal to one refresh cycle, while the lower half would have no delay. On average, this results in a display delay of half a refresh cycle. What happens if the frame rate is double or even triple the refresh rate? The tear lines would appear more frequently in this case. Would the display delay still average out to half a refresh cycle?
Another question is about Vsync. Theoretically, Vsync makes the GPU wait before completing a frame swap until the display finishes refreshing. If the GPU’s rendering time is shorter than the refresh cycle, the delay seems closer to Gsync, which is one refresh cycle. However, if the rendering time exceeds the refresh cycle, a frame swap would be missed, resulting in an additional refresh cycle of delay. So, this means Vsync would cause a minimum delay of one refresh cycle and a maximum of two refresh cycles, right? Additionally, since it limits frame rendering, there seems to be rendering delay related to the refresh rate (e.g., 60Hz locked at 60FPS). Could this also involve frame queue accumulation or similar issues? I’m curious about this.
Lastly, regarding frame capping and Reflex: since Reflex minimizes latency by clearing the frame queue, and frame capping also prevents excessive frame accumulation, why not simply set the frame queue to zero? If this would cause stuttering, why doesn’t Reflex encounter similar stuttering issues?
After turning off Vsync, assuming the refresh rate matches the frame rate, and if the tear line appears at the center of the screen, the upper half would have a delay equal to one refresh cycle, while the lower half would have no delay. On average, this results in a display delay of half a refresh cycle. What happens if the frame rate is double or even triple the refresh rate? The tear lines would appear more frequently in this case. Would the display delay still average out to half a refresh cycle?
Another question is about Vsync. Theoretically, Vsync makes the GPU wait before completing a frame swap until the display finishes refreshing. If the GPU’s rendering time is shorter than the refresh cycle, the delay seems closer to Gsync, which is one refresh cycle. However, if the rendering time exceeds the refresh cycle, a frame swap would be missed, resulting in an additional refresh cycle of delay. So, this means Vsync would cause a minimum delay of one refresh cycle and a maximum of two refresh cycles, right? Additionally, since it limits frame rendering, there seems to be rendering delay related to the refresh rate (e.g., 60Hz locked at 60FPS). Could this also involve frame queue accumulation or similar issues? I’m curious about this.
Lastly, regarding frame capping and Reflex: since Reflex minimizes latency by clearing the frame queue, and frame capping also prevents excessive frame accumulation, why not simply set the frame queue to zero? If this would cause stuttering, why doesn’t Reflex encounter similar stuttering issues?
RealNC wrote: ↑05 Dec 2024, 05:58Yes, but there's diminishing returns. Going from 60Hz to 120Hz provides a good improvement (up to 8ms), but going from 120 to 240 only has a minor impact (it's up to 4ms.) When going from 240 to 480 it's even smaller (just 2ms.)
The biggest latency improvement comes from capping your FPS. On a 144Hz display, if you cap FPS to 120, you'll get a huge improvement. You need two things to have the lowest latency possible with g-sync + vsync: a) Prevent FPS from going near max Hz, and b) prevent the GPU from getting fully saturated. If the game has Nvidia Reflex implemented, then it takes care of both of those automatically. Without Reflex, you need to take care of it yourself.
These are much more important than having a higher refresh rate.
