Agent_Buckshot wrote: ↑25 Jun 2025, 17:30
As for why you'd want the framerate to match the refresh rate on impulsed/strobing displays I was able to find this chart explaining that it's to reduce "double-image" artifacts that while observably similar to crosstalk artifacts when using blur reduction tech on LCD's aren't necessarily the same (for sake of this discussion we'll focus on persistence as reducing crosstalk when using blur reduction tech on LCD's is a separate discussion altogether)
strobed-display-image-duplicates.png
This actually a photo of a CRT.
But it's exactly the same for short-persistence strobed backlights.
Doesn't matter what impulsed display (plasma, CRT, strobed LCD, BFI'd OLED etc) but if you view the coverpage of
www.testufo.com you will see similar effects. You may get more/less motion blur per strobe.
Correct. That's why you want framerate=Hz to keep framerate=stroberate to avoid duplicate images.
For LCD, double images are caused by frames being visible more than once per flash, for multiple independent reasons:
1. Non-strobe-crosstalk reason: Frame being flashed multiple times
2. Strobe crosstalk reason: Frame visible because LCD GtG was longer than one refresh cycle.
3. Both of the above simultaneously (even worse; usually a triple image effect since crosstalk is a lagbehind effect on the most recent flash).
If you have a 240Hz display (60Hz displays are not as educational for this specific motion test; extra refresh rate headroom makes this test maximally educational) -- play with TestUFO Variable Persistence BFI to study the blur physics:
EDUCATIONAL VIEW ON LCD/OLED (not CRT)
1. TestUFO Variable-Persistence Black Frame Insertion:
testufo.com/blackframes#count=4
2. TestUFO Black Frame Insertion Double Images
testufo.com/blackframes#multistrobe=2
That being said:
TestUFO PowerPoint education animation demos:
Learning Level When Viewed on 240Hz+ LCD/OLED: Very high
Learning Level When Viewed on 60Hz LCD/OLED: Moderate
Learning Level When Viewed on CRT: Confusing (because this demo is designed to convert sample and hold into impulsed)
As you can see, software-based black frame insertion works successfully as a software-based method of reducing display motion blur.
The more Hz you have (480Hz), the more you can emulate a short-persistence display via software blackframe insertion. For example 60fps at 480Hz with 1 visible frame, 7 black frames = 87.5% less motion blur. My CRT electron beam simulator at testufo.com/crt does something similar.
Agent_Buckshot wrote: ↑25 Jun 2025, 17:30
Digging a little deeper online into what determines the actual persistence you end up with on a given CRT or Plasma, I found this updated persistence chart from Blur Busters that lists examples of what would provide you with a given level of persistence based on the display you're along with additional relevant info such as the refresh rate and whether or not you're using blur reduction tech
It's more simple/elementary than that. CRT persistence is so brief (<1ms) and non-squarewave (fast rise, slow fall - like a sawtooth on an oscilloscope), so it's hard to put a number on it. But CRT fade 90% is usually the cutoffpoint, so when you measure that, you see almost no motion blur unless you have extremely fast motion speeds (e.g. 3000 pixels/sec at 1920x1080 on a CRT, and few CRTs showed 1920x1080).
Don't forget that higher resolutions amplify visibility of motion blur (I call this the
Vicious Cycle Effect, better resolution means you need more Hz or shorter pulsewidths to more hide motion blur). One screenwidth/sec motion at 320x240 at 1ms MPRT will be below detectability, since 1/1000th of 320-pixels-wide is less than half a pixel.
One screenwidth/sec motion at 3840x2160 at 1ms still produces almost 4 pixels of motion blur (almost 4000 pixels/sec, divided by 1000ms in a second, equals 4 pixels of motion blur, ala Blur Busters Law)
Agent_Buckshot wrote: ↑25 Jun 2025, 17:30
One could assume based on this chart that for the most part any given CRT will give you ≈1ms of persistence, but based on the older "Impulsed Displays" persistence chart one could also assume that there are other factors one should consider when looking for a CRT or Plasma, as one could end up with a CRT that doesn't necessarily have ≈1ms of persistence or a Plasma that has ≈4ms of persistence.
It's hard to give an exact number because of phosphor fade. It's a triangle wave (sawtooth) on an oscilloscope rather than a squarewave (like a LCD strobe backlight turning on/off, or an OLED BFI turning on/off), but 1ms is actually a conservative number; it just has very faint after effects after 1ms, like a faint green/yellow ghosting effect for white objects on black background, that not everyone notices.
There's no "cliff" in the light fadeoff of a phosphor fade, so it's hard to give an exact milliseconds endpoint.
Agent_Buckshot wrote: ↑25 Jun 2025, 17:30
As for why I keep bringing up ≈4ms in relation to Plasma TV's, while doing more research I was able to find this YouTube video titled "Explaining the Plasma TVs Motion Clarity based on the Blur Busters Law" from the YouTube channel "Plasma TV for Gaming" which does a pretty good job of explaining how the Plasma TV tech works and how it ends up delivering ≈4ms of persistence in a simple & relatable fashion.
Yes. That's why 480fps 480Hz OLED has less motion blur than a plasma display (even a Kuro). The problem is you need to keep framerates very high. Or use extremely strong motion blur reduction (often better than the weak 50%:50% BFI, since it doesn't reduce blur as much as large-ratio BFI).
Agent_Buckshot wrote: ↑25 Jun 2025, 17:30
Unfortunately when looking for similar resources that explain what goes into determining the persistence of a CRT I came up short.
That's because:
1. CRT persistence was mostly meaningless at low resolutions (perfectly zero for retro-resolutions like 320x240; only barely noticeable at 1920x1080 fast scrolls)
2. CRT was long obsolete before we had high-Hz high-resolution displays
So we don't have any scientific papers comparing the two. However, I do have some resources at blurbusters.com/area51 -- including
www.blurbusters.com/why960
Remember.... 1ms MPRT at 1 screenwidth per second at Super Mario 256x224 = only one-quarter-pixel motion blur = zero persistence (motion blur is a rounding error below human visible noise margin). That's why you never talk about persistence.
Now, MPRT is more important when we have 4K today, with all that high resolution fast scrolling.
When viewing
www.testufo.com/map on an OLED display
As you can see at
www.blurbusters.com/120vs480
When it comes to *specifically* OLED (480Hz OLED), 120fps vs 480fps is more visible for scrolling situation than 60fps vs 120fps on OLED. This is unimportant for impulsed displays because 60fps framerate=Hz at 1ms (or less) = zero persistence for most stuff you view on a CRT. But with sample and hold, you need ultrashort frametime (high framerate for flickerless blur busting), or ultrashort refreshtime (if adding software-simulating impulsed on sample hold).
Agent_Buckshot wrote: ↑25 Jun 2025, 17:30
If anyone wishes to help demystify this topic I would greatly appreciate it; I'm having a great time using my CRT monitors for gaming and while the motion clarity is observably much better, I'm itching to learn more about what goes into the actual "Impulse/Strobing" technology used by these older displays and by extension blur reduction tech such as backlight strobing and black frame insertion on newer LCD displays.
TL;DR: More Hz the merrier for BOTH methods of blur busting (strobeless blur busting and strobed blur busting)
1. Shorter frametime = less blur
2. Shorter refreshtime = ability to use strong motion blur reduction algorithms such as
www.testufo.com/crt
You can use hardware-based strobe backlight or BFI, or software-based motion blur reduction algorithms. Some software such as RetroArch now includes software-based motion blur reduction. Which is better? Usually hardware, BUT sometimes software-based blur busting outperforms hardware-based blur busting, because sometimes manufacturers waterdown OLED BFI to only 50%:50% ratio, while with software BFI with large ratios (e.g. 60 vs 480), resulting in (1 visible frame for every 7 black frames) = 12.5%:87.% ratio of blur = 87.5% less motion blur = 2ms MPRT via software-based BFI! (better than Kuro plasma!)
For better or for worse, that's why framegen-based blur busting looks great on OLED, since you don't have flicker. But if you don't mind flicker of CRT, then CRT is more organic (no fake frames). However, if you want to understand the multiple options of blur busting, there is a new article
https://blurbusters.com/flicker-vs-fram ... e-methods/ -- which covers the multiple methods of blur busting.
(IMPORTANT: Don't view
www.testufo.com/crt on a CRT tube; it doesn't look good -- view on at least on a 120Hz+ OLED or 240Hz LCD. This CRT simulator algorithm is DESIGNED for sample and hold, to convert sample and hold to a rough emulation of a CRT tube. It performs best on OLED at 240:60 ratio or 480:60 ratio). Zero blur in Super Mario Brothers on 480Hz OLED, since Super Mario Brothers 256x224 scrolling usually doesn't need less than 2ms MPRT, because of the low retro resolution... But RetroArch sometimes flickers erratically on this because it's a software simulation and the software simulation of a CRT tube electron beam isn't always stable;
However, it's something that will become more important as CRT tubes become harder and harder to find;
Keep your CRTs, but at borrow or visit somebody's cheap 240Hz LCD/OLED, to self-educate. TestUFO training demos are very educational when it piggybacks on brute Hz for impulsed display simulation.
