Originally Posted by Mashie Saldana
So hi refresh rate is nothing new, there were/are LCD's running at 600Hz.
Faked refresh rate, not true refresh rate.
Big Hz was often marketing:
- Plasma subfield frequency (which were like temporal dithers) of a lower Hz, e.g. 10 subfields per true refresh cycle image.
- Temporal dither frequency; e.g. using 10 temporal dithered refresh cycles to generate one frame / one refresh cycle.
- Scanning backlight frequency or strobe backlight length (real Hz was often still 60 or 120 in these times)
- Motion clarity ratio based on 1/600sec strobe flash length even though refresh rate was only 60Hz or 120Hz (it's true that a strobed 1/600sec flash will have roughly the same motion blur of a true [email protected], but you will still have things like the phantom-array effect / stroboscopic effect like www.testufo.com/mousearrow
- Repeat-refresh frequency (e.g. refreshing the display repeatedly to improve color quality or speed up pixel response)
Motion blur is frame visibility time, whether the frame is visible for full duration of refresh cycle or is padded by black frames. A 1ms flash at 120Hz, would have the same amount of display motion blur as 1ms frames filling the whole second ([email protected]
) -- that's why it's so hard to eliminate motion blur strobelessly.
In those times, 600Hz has never been done by 600 unique non-faked frames in 600 refresh cycles. Even interpolation (Motionflow) didn't go at that high framerate.
Today, we finally can experiment with it now with the ultra-Hz displays that scientists/researchers now finally have -- true 600 unique frames at a true 600 Hz was not possible until relatively recently.
Originally Posted by Mashie Saldana
What I'm talking about here is capturing video at 120Hz or more which is wasted
Commonly perpetuated myth, understandably founded on old assumptions. But, buddy, it's time to lay that myth to rest as science has already proven otherwise. And camera shutter speeds can easily be faster than display persistence.
As founder of Blur Busters and inventor of TestUFO, I have ultra-Hz displays here.
While the camera shutter is indeed a limiting factor, remember, many sports camera can do a 1/1000sec shutter.
It depends on the camera shutter. 1/1000sec shutter video looks much clearer at 480fps at 480Hz on my experimental 480Hz display
. I was the world first to write mainstream tests about 480Hz.
We are already doing UltraHFR experiments
and I have a big reduser.net thread about Ultra HFR
To have blurless sample-and-hold, and have 1/1000sec camera shutter without phantom array effects and without motion blur, you need to fill the whole second with 1ms frames, aka [email protected]
Basically flickerless CRT clarity without the need for phosphor/impulsing/black frames/black periods/etc. To match 1ms motion blur and also completely avoid stroboscopic effects. Basically low-persistence sample-and-hold. Real life does not strobe, does not flicker, and strobing/BFI is a humankind band-aid until we've achieved ultra-Hz.
Also, source persistence and destination persistence is additive. A 1/60sec camera shutter and a 1/60sec sample-hold display == creates 2/60sec combined motion blur. But now with camera shutter of 1/1000sec with UltraHFR video running at 240fps, even a 240Hz sample-hold display is a motion blur bottleneck (1ms source persistence, 4ms destination persistence).
When HDTVs first arrived, some camera operators had to slow down camera motion and increase shutter, to increase motion blur, to hide display limitations better -- you no longer had CRT motion clarity pans of yesteryear of CRTs. So to prevent excess motion blur, camera operation changed. Only on certain material (e.g. sunny day downhill skiing or such) that you see the 1/1000sec shutter used, and then you see the stroboscopic effects of low-Hz. I see it sometimes on video material, and so you've got a tradeoff between camera shutter blur or stroboscopic effect. Can't fix both simultaneously without ultra-Hz.
Motion blur from persistence behaves the same way on camera end and display end.
1ms display persistence = 1 pixel of motion blur per 1000 pixels/sec motion on a sample-hold-display
1ms shutter persistence = 1 mm of motion blur per 1000 mm/sec motion on a photograph
(or angular motion, if you prefer to measure that way)
60fps @ 60Hz = 16.7ms motion blur on sample-hold displays
120fps @ 120Hz = 8.3ms motion blur on sample-hold displays
1000fps @ 1000Hz = 1ms motion blur on sample-hold displays
To get the same improvement of 60fps->120fps (8.3ms improvement), one observes the need to jump extremely dramatically up the diminishing-points-of-returns curve 120fps->1000fps (7.3ms improvement). There are cheap ways to achieve ultra-Hz that display manufacturers are not yet milking, and also on old fashioned assumptions.
Not everyone needs ultra-Hz but the benefits for UltraHFR video are confirmed plainly human visible. Have you ever seen UltraHFR like I have, like [email protected]
, [email protected]
and [email protected]
Yes, we can fix motion blur by using strobe/flicker/phosphor/CRT flicker/light modulation. But real life doesn't flicker, and the only way to fix stroboscopic effects AND motion blur effects in video simultaneously, and get 100% strobeless (and phantom-array-less) CRT motion clarity with UltraHFR video, is ultra-fps at ultra-Hz.
It's already scientifically proven that to have cake (lack of stroboscopics/wagonwheel effects) AND eat it too (lack of motion blur), requires an extremely dramatic climb.
NVIDIA scientists such as Morgan Mcguire agree with me, and many people at NVIDIA are huge fans of me now, considering me as a refresh rate visionary, since very few people get to experiment with ultra-Hz displays and make observations that nobody else has ever seen.
Having the fortunate experience of working with display manufacturers, I've written an excellent Hertz myth-busting article, confirmed by several scientists, see Blur Busters Law And The Amazing Journey To Future 1000Hz Displays
(Complete with peer reviewed links, annotated animations & photographic proofs)
I also have a peer reviewed paper, co-authored with NIST.gov, NOKIA, Keltek, on a display motion-blur measurement technique
, so I know what I am writing about.