I agree with you, if you're just watching movies, then the refresh rate doesn't matter nearly as much, but how they process 3:2 or handle 24p input, etc. Happened to notice that the original poster said "occasional movie", so that might not be a priority. And, for me, I want to use the display for interactive and fast-motion applications -- for example, sports, computers or video games -- so refresh matters a great deal. That said, the HDGuru article has some pieces of now-outdated info about the benefits of scanning backlights.It's also true 120Hz has rapidly diminishing returns, so you need to do the big jump to 960Hz if you want startlingly clear benefit beyond 120Hz
(Assuming material that meets (1) full framerate at input; (2) no built in motion blur in the frames; and (3) sufficient fast motion) -- "120Hz" has mathematically 50% less motion blur than 60Hz. Doubling it again, has 75% less motion blur. Keep doubling to 480, it's now 87.5% less motion blur. Then double one more times to 960Hz, and you've got mathematically 93.75% less motion blur than 60Hz (though CMR 960 and Sony Motionflow XR 960 comes nowhere this close, due to various limiting factors such as persistence, scanning backlight diffusion spreading between on-segments versus off-segments, motion interpolation algorithm limitations, etc). The point here being, the jump 120Hz->960Hz (43.75% addition motion blur eliminiation) has almost as big jump as the jump betwen 60Hz->120Hz (50% motion blur elimination). So if you go beyond 120Hz, don't bother going to 240 or 480 -- definitely go all the way to 960, for the massive upgrade, just to do an adequate jump in the graph of diminishing returns. As a result, mathematically, the 120->960 upgrade is nearly the same huge upgrade as the 60->120 upgrade
... The "960" can be emulated either via interpolated frames (true 960fps), or via 1/960sec strobes (black frame insertion or scanning backlight), or a combination thereof (e.g. 240fps interpolation, that are strobed with 1/960sec backlight flashes) -- that's the technique both Samsung and Sony does with their respective "960" rated displays, even though they often fall short of the actual measurments (much like contrast ratio exaggerations). CRT displays have a 1000 or 2000 "Hz equivalence" (not actual Hz, but motion resolution equivalence of a LCD running at X fps @ X Hz) due to their phosphor decay of 0.5-1.0 milliseconds; and it's only recently that technologies have been found to allow LCD to get into that ballpark equivalence.
In the past, most old scanning backlights don't do a good job of reducing motion blur - often only by an imperceptible ~25%. Since that HDGuru article was created, better scanning backlights have been made -- e.g. in Samsung CMR 960 and Sony Motionflow XR 960 displays. They benefit situations when you're running material that excel in motion clarity on CRT displays, which are (1) running full-framerate material (60fps), and (2) material that has no built-in motion blur (fast camera shutter, non-excessive compression, etc), and (3) the motion is very fast (e.g. alpine skiing races, red bull air races, NASCAR, fast football pans, etc). They don't benefit situations such as movies very much (except for things like projector style effect from strobes, etc).
I've got a FAQ about scanning backlights
- presently, it's more geared towards the project I'm doing, but it's quite important to gain an understanding how the "ratings" like "Clear Motion Rate 960" came to being. In particular, read "Q: What is Samsung CMR 960 or Sony Motionflow XR 960?" ... They are sometimes viewed as marketing exaggerations by some reviewers. Measurements often show that they do not reflect real-world benchmarks (e.g. similiar exaggeration problem exists in contrast ratio claims versus actual measurement). However, there’s a honest actual scientific basis behind these numbers (it was very interesting when I figured out why they came up with the numbers), but measured numbers fall short of actual rating -- much like contrast ratio exaggerations. However, the science is sound and there is much improvement that can be done.
In the long term, a scanning backlight that equals CRT quality (without motion interpolation), needs approximately 150 watts per square foot (at current LED efficiencies), for the short single phosphor-style strobes required every frame, so scanning backlights optimized to run sequentially (CRT emulation scan pattern, at least at the millisecond granularity -- segment at a time, not scanline at a time) for motion blur reduction, are extremely challenging to create cheaply. Phosphor illuminates amazingly brightly on a CRT for a tiny time period, so backlights that simulate phosphor illumination, for as short as time period, to achieve the flicker stroboscopic sharpening of motion (ala CRT), need to be as bright as phosphor during a brief flash. 150 watts per square foot of LED backlight is incredibly expensive to manufacture -- imagine putting 800 watts of LED in a 47" HDTV (even if short strobes means it still averages out to 40 watts). In addition, the introduction of modern 120Hz 3D LCD's have enabled pixel persistence to be less than a frame of refresh (since they must refresh incredibly fast between frames while both shutters are closed during the first few milliseconds of a refresh), allowing an opportunity to completely bypass pixel persistence as the motion blur barrier (just keep it in total darkness), and strobe the backlight when the pixel persistence is practically (99%+) gone (Zero Motion Blur LCD
!). This prerequisite has opened the door to ultrahigh-peformance scanning backlights to be possible (once wattage is cheap enough). So my experiments are being started on a small 3D-compatible LCD panel glass (modified 23" 3D 120Hz computer monitor, Asus VG236H or Samsung S23A700D) through my Arduino-powered scanning backlight project (project is on my BlurBusters Blog
). You see many CRT users still using direct-view's in the Direct-View Forum because of excellent motion resolution. Some people hate the flicker, but my system is geared towards computer and at 120Hz native PC signal (which doesn't flicker to most people) -- for near-zero-added-input-lag motion blur elimination -- something that can't be done with interpolation (adds massive input lag). Obviously a niche "videogamer videophile" application, but interesting science behind it all. (See Science & References
Ultrahigh-performance scanning backlights scientifically optimized for motion blur elimination (90-95% motion blur elimination for the "perfect CRT-style sharp motion on LCD") are not yet available in any HDTV or computer monitors yet today, but I am hoping manufacturers will begin introducing this (at least as a configurable "High Peformance Game Mode" setting). Albiet a niche for high-motion-resolution enthusiasts, it is applicable to both desktop monitors and full-screen HDTV's. Then, this indeed, will be a "feature worth buying" for computer users and video gamers looking for a high end LCD display that has no visible motion blur, without the input lag of motion interpolation
. Until then, other attributes that you already mentioned -- your list is an excellent list of criteria.Edited by Mark Rejhon - 11/8/12 at 12:17am