En video om Gsync/Freesync & FPS Caps

Såå... då var vi här igen hehe.. Detta ständigt omdebatterade ämne. HUR ska vi ha våra inställningar?
Alla har olika tips och råd. Så har det varit så länge jag kan minnas.
Vad säger ni om det hela? Vad har ni för inställningar?

Se videon och/läs summeringen under och ge mig era tankar.

This Setting is Ruining Your PC - Gsync/Freesync & FPS Caps
Over 90% of you watching right now are almost certainly using settings that either make your display look way worse, massively increase your latency, or both… Let’s talk about it

Now, before we get into this trust me favorite this video right now, because believe me you’re going to want to come back to this one time and time again as this video is my magnum opus of the over a year of latency and motion testing I’ve been doing on dozens of displays using both Nvidia and AMD GPUs with a 1000fps camera, and what I’ve found is going to shock you, as it turns out when set up correctly you can actually have far lower latency and a better image in motion all while using less PC resources.

How, well it all has to do with sending the fastest, most up to date frame your PC can produce without tearing, and unfortunately the default settings for GPUs and games don’t allow this to happen, and most advice on settings people give is also incorrect.

Now this journey all started with me testing various fps caps and Nvidia control panel settings, and what I found was that yes the old adage of Vsync adding tremendous latency in games was true, however at the time I also found that Gsyc+Vsync in the Nvidia control panel +Reflex in game could actually lead to far less input delay on many games and in fact was overall better than using an external cap in the driver or Riva Tuner Statistics Server to set a global limit.

Not only that, but you would also get a tear free experience using these settings and it could actually be even faster than no fps cap as when you are heavily GPU bound this can actually increase latency.

However, viewers then pointed out that external caps can have higher latency than in game caps, so I went to investigate this claim now using an AMD GPU and I found that they were actually correct. In-game caps led to far lower latency than any other method.

However I still hadn’t had a chance to revisit this on Nvidia GPUs to see if this is actually better than Gsync+Vsync+Reflex, so I recently did just that.

Now for these tests I ran 3 different configurations on a 138Hz OLED monitor.

Gsync+Vsync+ULLM
Gsync+Vsync+Reflex
Gsync+Vsync+133fps cap
Gsync+120fps cap as in my testing it turns out that VRR used in Gsync and Freesync monitors actually needs 1ms of time to fully work.

What I mean by this is for you to get 0 perceivable tearing you must add an additional ms of latency to give the display more time to process.

This is done by first getting the latency of your monitor which for me as an example would be to take 1000ms / 138fps to get 7.246ms of total latency.

You then add 1ms to this number which would give us 8.246ms of total latency, and to understand what frame rate that is, we simply take 1000ms again / this new latency which gives us 121.27fps, however I then round down to the nearest even number to give it a little headroom, which means that actually the max framerate we can run on a 138Hz display is 120fps.

Other common refresh rates would look like this

240Hz = 192fps
144Hz = 124fps
120Hz = 106fps

When I was testing AMD GPUs I found that If you don’t do this you will likely see slight tears at the bottom of your screen which can make the display look weird and throw you off when you're playing a competitive game.

However on an Nvidia GPU I found that if you combine Gsync+Vsync you actually only need to add 0.3ms of latency for VRR to work correctly

240Hz = 223fps
144Hz = 138fps
120Hz = 115fps

so in theory unless AMD can also do the same thing, Nvidia may be able to deliver higher untorn framerates, but that’s only if the latency is still good, so let’s find out.