Blur Busters Forums

Oh yes, 240fps camera.

One of us should research if the Samsung Galaxy S9 or other smartphones that records only 0.2 to 0.4 seconds of highspeed video, can trigger its super-slo-mo on the LED-turn-on event. That would in theory allow super-slo-mo of the new to work with game lag measurement tests, despite the very short high speed recording times.

I purchased a Sony Xperia XZ Premium for less than $200 off eBay; it's going to be my 720p 960fps high-def camera for TestUFO scanout tests, like for this article. That's 10x the resolution of my old 64-pixel-tall "thin strip" high speed Casio camera I originally used for the LightBoost video several years ago.

Here's a list of cheap high-def 1000fps cameras.

The main problem is these cameras records for only 0.2 to 0.4 seconds of high definition 1000fps of 960fps video. So the motion-trigger should be configured to the LED turn-on (there's often a selectable area on the screen to detect motion on), to begin automatically recording high speed video from that moment.

<Blur Busters Idea>
An Android app could theoretically be developed for these smartphones (Samsung Galaxy S9/S9+ and Sony XPeria Premium XZ/XZ2) to automatically count the frames from the LED turn-on to a screen response behaviour (measure timing of change at a specified pixel coordinate, to timing of change at a different pixel coordinate). If one of you readers are an Android programmer with one of these 960fps phones, I want to hear from you about a possible collaboration -- [email protected] -- I have Android and iPhone programming experience and can help with my Sony Xperia XZ Premium phone too but I would like to collaborate with someone on this.
</Blur Busters Idea>

Joker18 wrote:As far as I know RTTS limiter ads a bit of inputlag compared to ingame limiter and this is one of the reasons it can achieve a much smoother experience.

The reason for the smoother frame pacing is just the accuracy of RTSS. Has nothing to do with latency. Most in-game limiters don't try to be sub-millisecond accurate. In fact, sometimes they don't even try to be 1ms accurate. They can vary by several milliseconds. This is why RTSS gives lower input lag with vsync. In-game limiters will overshoot the cap.

Is RTSS one of the better frame rate limiters? Does it not add an overlay? Why not just use NVidia Inspector Frame Rate Limiter that comes from NVidia drivers? Is RTSS somehow better?

NVidia Frame Rate Limiter had options I do not understand - Noop, Flip by Flip, etc. Which one do I want for a G-Sync display? I have in-game V-Sync disabled, NVidia Inspector V-Sync Forced On, and G-Sync forced on for FullScreen and Borderless mode. My monitor goes as high as 165Hz, so I figured 160fps is about right. Which setting should I stick to it in NVidia Inspector then? Default? I obviously want Frame Rate Limiter enabled to get rid of max FPS/Hz input lag.

Or should I use RTSS?

Optimal G-SYNC settings for reference:
https://www.blurbusters.com/gsync/gsync ... ttings/14/

In-game vs. RTSS vs. Nvidia limiters input lag tests:
https://www.blurbusters.com/gsync/gsync ... ttings/11/

jorimt wrote:Optimal G-SYNC settings for reference:
https://www.blurbusters.com/gsync/gsync ... ttings/14/

In-game vs. RTSS vs. Nvidia limiters input lag tests:
https://www.blurbusters.com/gsync/gsync ... ttings/11/

Huge Thanks! Could you post a shot of where I'd input those values for Scanline Sync? I only see a single on/off option that does not specify whether it is Scanline Sync 0 or 1.

I figured it out, I think. The settings are manually adjusted in Global profile (opened with Notepad). I have them like this now for NVidia Forced V-Sync + G-Sync on 165Hz display:
[Framerate]
Limit=160
LimitDenominator=1
LimitTime=0
SyncScanline0=1
SyncScanline1=0
SyncTimeout=0
SyncPeriods=0

The lines in bold are the ones OP suggests are the best for my configuration, right?

Nope, not working... SyncScanline0 must be 0 for Grim Dawn to even launch. It sometimes launches when SyncScaline0=1, but FPS fluctuate from 80 to 160 and the game stutters hardcore...

You don't need to edit the file. You can enter a scanline directly in the RTSS UI. Unless you're using an older RTSS version. The latest version allows you to enter a scanline value directly in the UI, and you click the "scanline sync" button to switch between different sync modes (normal, x*2 and x/2).

If a game is severely stuttering with s-sync, that means there's too much GPU load. s-sync works best when GPU load is less than about 60% or so, maybe 70%.

Note: scanline sync is only useful when you don't have g-sync. It's basically "poor people's g-sync." If you do have g-sync, you don't need scanline sync (and it will probably not even work correctly with g-sync.)

MonarchX wrote:I have them like this now for NVidia Forced V-Sync + G-Sync on 165Hz display

Why are you using Scanline sync with G-SYNC?

Scanline sync is, as far as I'm aware, user controllable tearline steering, which means it isn't technically a form of V-SYNC at all, but a V-SYNC OFF method that mimics the look of traditional V-SYNC, but without the lag.

It's not intended to be used in conjunction with VRR, but to be used standalone if, for instance, your display doesn't have VRR capabilities.

Scanline sync should NOT be combined with GSYNC/FreeSync.

Scanline sync is a "VSYNC ON" lookalike with lower lag than "VSYNC ON"

Scanline Sync should only be used when you need VSYNC ON for any reason (e.g. improved framepacing with ULMB)