- Jun 2, 2013
- 122
- 4
- 18,595
TL;DR
Using any kind of frame limitor in Cyberpunk 2077 induces micro stutter as long as you are hitting the monitor's set refresh rate. The easiest method I use to observe micro stutter is with a controller, then in game pan the camera slowly left/right, then with in 10-30 second I will notice at least 1 stutter in this space in this time frame.
Detailed Version
Lately I have been doing some extensive testing regarding frame time consistancy and PC latency. I've been doing this because with the latest DLC Phantom Liberty just recently coming out its re-invigerated me to play the game again. Not to mention I've also recently gotten a new monitor which looks fantastic with games.
Firstly I need to say I am extremely sensitive to bad frame time delivery and latency in games and I can almost always notice a single micro stutter. This is truly a curse - it drives me insane at times to troubleshoot issues which shouldn't be a problem but I still spend hours upon hours trying to troubleshoot and improve the delivery of fps in games. Truly a curse.
So now. Ofcourse I like to game at high fps and have the hardware to do so. However If I can't game at a high fps as desired, then I will sometimes cap the frame rate 1 frame lower, at a lower supported monitor refresh rate (this is important to note).
Every game I have ever tried and tested this method of capping frames below the monitor's refresh rate - with Gsync - works exceptionally well, because Gsync is doing what it's supposed to be doing - matching frame output to the monitor's refresh rate. Meaning a capped framerate of 59fps running @ a refresh rate or 60hz runs flawlessly smooth as well as (this is important) retains as low as possible PC latency, or lag.
This is the next important thing to note. From my own experience, I have noticed that most games in general have the least amount of latency when running very close to the refresh rate but not matching or exceeding the refresh rate. Simply put, I have observed that when fps is capable of exceeding the refresh rate there is a large amount of latency/delay. And when the fps is no where near able to match the refresh rate, there is almost the same amount of latency. The sweet spot is where you can exceed the refresh rate, cap the fps 1 frame below - which gives a smooth frame delivery because Gsync is doing it's thing, and latency is as low as it can be.
Results I have observed with Cyberpunk 2077 runinng at 60hz (all results +/- 10ms)
- (1) Vsync on via nVidia Control Panel (NCP) no other frame cap or in game vsync in place - approx 80ms+ latency
- (2) In game vsync on, NCP vsync off, no other frame cap in place - approx 100ms+ latency
- (3) Vsync on via NCP and fps capped to 60 via RTSS (Riva Tuner Statistic Server) - approx 55ms+ latency
- (4) Vsync on via NCP and fps capped to 59 via RTSS - approx 38ms+ latency (this is the best result)
With all this in mind, I am finally getting to the issue as suggested in the title of this thread. If my memory serves me well, in the early days when Cyberpunk 2077 was first released, this method of capping the fps under the monitor's refresh rate worked well as it should. As the game is today however, capping the fps 1 frame lower results in micro stutters. No matter what other method I try - trust me I've tried every possible combination of refresh rate, frame cap, vsync, gsync, fast sync, scanline sync, half resfresh rate, low latency, pre rendered frames that you could possibly think of, result no. 4 above has been the best result for latency but has introduced micro stutter now where before it never had - and shouldn't when using gsync. As the game is now, when I use any form of frame limitor including nVidia Reflex - as it has a frame limitor implemented into it as well I will get micro stutter (as long as the fps output from the graphics card is capable it hitting the refresh rate). If the fps is not is not hitting the refresh rate, then micro stutter does not not occur, but there is also a bigger latency penalty similar to result 3 above.
Also I have been using nVidia's OSD (activate by pressing ALT+R) to observe latency.
I'm curious to know others thoughts on this subject, how you personally play the game, any observations you might have came accross and whether anyone else has experienced the same?
Specs:
OS: Windows 11 Pro
MAINBOARD: Gigabyte Z790 Gaming X AX
CPU: i9 13900KF
CPU Cooler: Corsair H100i
RAM: 32GB Corsair Dominator Platinum 7200mhz
VIDEO CARD: Gigabyte Gaming 24gb OC RTX 4090
MOUSE: Logitech Gaming 300s
KB: Corsair K70 MK.2
Xbox Elite2 Gaming Controller
MONITOR: AW3423DW, 27GL850
HDD: Kingston KC3000 2TB
HDD: Samsung NVMe M.2 Evo Plus 970 1TB
Power: Silverstone ST1200w Platinum
Using any kind of frame limitor in Cyberpunk 2077 induces micro stutter as long as you are hitting the monitor's set refresh rate. The easiest method I use to observe micro stutter is with a controller, then in game pan the camera slowly left/right, then with in 10-30 second I will notice at least 1 stutter in this space in this time frame.
Detailed Version
Lately I have been doing some extensive testing regarding frame time consistancy and PC latency. I've been doing this because with the latest DLC Phantom Liberty just recently coming out its re-invigerated me to play the game again. Not to mention I've also recently gotten a new monitor which looks fantastic with games.
Firstly I need to say I am extremely sensitive to bad frame time delivery and latency in games and I can almost always notice a single micro stutter. This is truly a curse - it drives me insane at times to troubleshoot issues which shouldn't be a problem but I still spend hours upon hours trying to troubleshoot and improve the delivery of fps in games. Truly a curse.
So now. Ofcourse I like to game at high fps and have the hardware to do so. However If I can't game at a high fps as desired, then I will sometimes cap the frame rate 1 frame lower, at a lower supported monitor refresh rate (this is important to note).
Every game I have ever tried and tested this method of capping frames below the monitor's refresh rate - with Gsync - works exceptionally well, because Gsync is doing what it's supposed to be doing - matching frame output to the monitor's refresh rate. Meaning a capped framerate of 59fps running @ a refresh rate or 60hz runs flawlessly smooth as well as (this is important) retains as low as possible PC latency, or lag.
This is the next important thing to note. From my own experience, I have noticed that most games in general have the least amount of latency when running very close to the refresh rate but not matching or exceeding the refresh rate. Simply put, I have observed that when fps is capable of exceeding the refresh rate there is a large amount of latency/delay. And when the fps is no where near able to match the refresh rate, there is almost the same amount of latency. The sweet spot is where you can exceed the refresh rate, cap the fps 1 frame below - which gives a smooth frame delivery because Gsync is doing it's thing, and latency is as low as it can be.
Results I have observed with Cyberpunk 2077 runinng at 60hz (all results +/- 10ms)
- (1) Vsync on via nVidia Control Panel (NCP) no other frame cap or in game vsync in place - approx 80ms+ latency
- (2) In game vsync on, NCP vsync off, no other frame cap in place - approx 100ms+ latency
- (3) Vsync on via NCP and fps capped to 60 via RTSS (Riva Tuner Statistic Server) - approx 55ms+ latency
- (4) Vsync on via NCP and fps capped to 59 via RTSS - approx 38ms+ latency (this is the best result)
With all this in mind, I am finally getting to the issue as suggested in the title of this thread. If my memory serves me well, in the early days when Cyberpunk 2077 was first released, this method of capping the fps under the monitor's refresh rate worked well as it should. As the game is today however, capping the fps 1 frame lower results in micro stutters. No matter what other method I try - trust me I've tried every possible combination of refresh rate, frame cap, vsync, gsync, fast sync, scanline sync, half resfresh rate, low latency, pre rendered frames that you could possibly think of, result no. 4 above has been the best result for latency but has introduced micro stutter now where before it never had - and shouldn't when using gsync. As the game is now, when I use any form of frame limitor including nVidia Reflex - as it has a frame limitor implemented into it as well I will get micro stutter (as long as the fps output from the graphics card is capable it hitting the refresh rate). If the fps is not is not hitting the refresh rate, then micro stutter does not not occur, but there is also a bigger latency penalty similar to result 3 above.
Also I have been using nVidia's OSD (activate by pressing ALT+R) to observe latency.
I'm curious to know others thoughts on this subject, how you personally play the game, any observations you might have came accross and whether anyone else has experienced the same?
Specs:
OS: Windows 11 Pro
MAINBOARD: Gigabyte Z790 Gaming X AX
CPU: i9 13900KF
CPU Cooler: Corsair H100i
RAM: 32GB Corsair Dominator Platinum 7200mhz
VIDEO CARD: Gigabyte Gaming 24gb OC RTX 4090
MOUSE: Logitech Gaming 300s
KB: Corsair K70 MK.2
Xbox Elite2 Gaming Controller
MONITOR: AW3423DW, 27GL850
HDD: Kingston KC3000 2TB
HDD: Samsung NVMe M.2 Evo Plus 970 1TB
Power: Silverstone ST1200w Platinum
Facebook
Twitter
Instagram