Page 1 of 1

comparison of monitors on third-party sites

PostPosted: 14 May 2019, 07:25
by 1000WATT
An ordinary user studying the review on monitors most often uses simple logic of visual comparison and does not go into details.
In the review on the monitor pg258q https://www.tftcentral.co.uk/reviews/as ... pg258q.htm there is an image https://www.tftcentral.co.uk/images/asu ... it_144.jpg.
photos pg258q and pg278q taken at a speed of 960 pixels per sec.
Yes, from the technical side, that's right, but most of what moves makes it distance per second.(fov games, movies, etc.)
1920/960=2 2560/960=2,6
valid for FPS games, this picture(https://www.tftcentral.co.uk/images/asu ... it_144.jpg) is only if pg278q is used 1080p in mode black bars around the image.

Re: comparison of monitors on third-party sites

PostPosted: 16 May 2019, 12:18
by Chief Blur Buster
There are infinite possible motion speeds.
As a result, to compare monitors better, there is a need for a standardized motion speed.

Why The 960 Pixels/Sec Motion Speed Was Standardized
Blur Busters TestUFO standardized on 960 pixels/second because it's the closest number to 1000 pixels/second that is divisible by common refresh rates (60Hz, 120Hz, 240Hz), and translates extremely well to Blur Busters Law (1ms = 1 pixel of motion blur per 1000 pixels/sec). Also, it's fast enough to represent many common game motions, but not too fast for a manually-propelled pursuit camera (reviewers use our invention).

It's Easy To Convert motion blur to other motion speeds on same monitor
It's easy to convert motion blur to faster and slower speeds; twice the motion speed will have exactly twice the motion blur (for non-strobed modes). So a faster turn at 3000 pixels/sec will have approximately 3 times the motion blur seen in the photos. A blurrier-than-average monitor will retain the same blurrier-than-average times 3. And a clearer-than-average monitor will retain the same clearer-than-average times 3.

Same-monitor motion blur is predictable at other motion speeds...
....once there is one accurate motion-blur photo for one motion speed

Persistence-based display motion blur very predictably linear for a non-strobed mode. It never varies if on the same monitor -- it's very simple mathematics of motion blur; For eye-tracking-based situations, it is always perfectly exactly twice the motion blur at twice the motion speed -- with exactly the same corona/ghosting artifacts (just stretched 2x longer blur trail) assuming there's no source-based motion blur in the original motion material.

Different monitors do have different blurs. But when comparing motion speeds of the same mode on the same monitor, it's easily extrapolated. Refresh cycles don't even know how fast motion is, they're just simply flipbooking through static images at a rapid rate (60 images per second, 120 images per second, 144 images per second, 240 images per second). So the blur is the tracking itself. It's simply a function of eye tracking speed; faster eye tracking motion (or pursuit camera tracking motion as a clone of eye tracking motion) translates into more motion blur for the exact same pixel of exact same GtG/MPRT -- the pixel response never changes speed for different motion speeds, so the only variable is external; the tracking speed, and it's typically a mathematically perfect extrapolation.

If you change motion speeds at http://www.testufo.com you will see this happen.
- Halve motion speed, halve motion blur (of what was seen at 960 pix/sec on same monitor)
- Double motion speed, double motion blur. (of what was seen at 960 pix/sec on same monitor)

But to compare between 2 monitors, the world needed a standardized motion speed -- and Blur Busters has chosen 960 pixels/sec for a very good reason. Screen width motion in 2 seconds. Other websites have followed suit for this simplicity. A few may consider doing 1920 pixels/sec especially on 4K displays.

Testing multiple motion speeds is okay, as long as the standardized motion speed is tested
There's noting preventing sites from choosing other motion speeds, but 960 pixels/second has been the standardized motion speed for TestUFO tests and all the websites that use TestUFO or TestUFO-derived (i.e. RTINGS) motion tests. It's a really great goldilocks motion speed that allows reliable comparisions.

Web testing using multiple motion speed can potentially confuse people
But, if everybody used different motion speeds, it would be much harder to compare between different websites, and multiple motion speeds per monitors will confuse users (especially if you accidentally post screenshots out-of-context: a low-quality monitor can do 480 pixels/sec better than a high-quality monitor can do 1920 pixels/sec). If websites use different motion speeds, I strongly recommend them to embed the motion speed graphically at the bottom edge of the motion-blur photo -- to prevent context abuse.

This is exactly what you say, the ordinary user ..."often uses simple logic of visual comparison and does not go into details".

So that's why 960 pixels/sec has now become the consistent standard for testing motion on gaming monitors.

Re: comparison of monitors on third-party sites

PostPosted: 16 May 2019, 12:37
by Chief Blur Buster
Now about measuring motion blur of video or film material....

About motion blur built into movies/video
Don't forget that source persistence (camera persistence) and destination persistence (display persistence) is additive. That means a movie or video with 1/30sec camera shutter, combined with a display of 1/60sec persistence, translates to motion blur of (1/30sec + 1/60sec) = 3/60sec of combined motion blur.

Reliable measurement of display motion blur requires material with no source-based motion blur
Websites don't typically test for this; because the source material blur is currently unknowable. However, you can standardize on a test video or video test pattern, and that is how some television websites do it. But if you want to isolate display motion blur (without camera shutter motion blur variable), you need source material with zero motion blur -- like the TestUFO motion tests. That makes measurement of display motion blur reliable. Just like turning off "GPU Motion Blur Effects", since you don't want source-based motion blur to interfere with motion blur measurements.

Source blur (video) and destination blur (display) can be separately measured and later be added together successfully reliably
That said, once you measure display motion blur, it's actually quite predictably additive to whatever source persistence there is (original camera shutter). So you just add to whatever successfully measured motion blur there is. Once you measure blur of one video, it can be reliably added to whatever blur measurements of any display. Conversely, once you measure blur of one display, it can be reliably added to whatever blur measurements of any display. It's a two way street there as the blurs are independent but are exactly mathematically additive.