Actually, I can tell when strobe crosstalk photography is reasonably accurate or not. It's VERY tricky to correctly photograph, and you should trust me when I give instructions to people on how to photograph display motion blur -- sometimes the photographs are taken wrong, and sometimes the photographs are good enough.
ericl wrote:The pictures are misleading AF because your eyes don't see what cameras pick up. (I think we all know that)
While that is true, it's worth noting that I'm the author of a peer reviewed improved method of photographing display motion blur, see
pursuit camera paper and several reviewers now use this technique. It's much more WYSIWYG now, as you can see of this photograph of
http://www.testufo.com/ghosting
Originally photographed by Jorim under my instruction, this is (now) my Acer XB252Q in Overdrive Normal, Extreme, and OFF during 240 Hertz at 960 pixels per second.
- Ladder track is WYSIWYG
- Amount of ghosting is WYSIWG
- Amount of blur is WYSIWYG
Obviously, this will often miss some subtle temporal artifacts (e.g. DLP dithering, LCD Inversion) but in fact, those has been even successfully photographed using variants of the pursuit camera technique.
The horizontal ladder track is the temporal test pattern used to make sure it was photographed accurately. If it looks virtually exactly the same on the photograph as you did with your human eyes, then you've got the most accurate possible motion blur photographs. The important thing is that you want long exposure + accurate sync to get as blurless as possible that a long-exposure panning camera photograph (~1/30sec) allows (roughly the approximate integration period of human vision) -- and then voila -- your photograph has the same amount of motion blur as it gave your human eyes. This is the most accurate method of photographing display motion blur and related artifacts such as ghosting/coronas/crosstalk/etc
As you can see.... this is the (mostly) WYSIWYG method of photographing display motion blur. A moving camera capturing multiple refresh cycles is a relatively accurate stand-in for a moving eyeball. A panning or rotating camera is the analog for a rotating eyeball tracking moving objects and is thusly the correctly method of capturing display motion blur (for those who don't understand eyetracking is the majority cause of display motion blur, please see
TestUFO Eye Tracking Motion Blur.
My innovative free invention (that replicated a $30,000 motion blur photographing technique, into blogger's hands) -- that other websites now use (RTINGS, TFTCentral, TomsHardware, SweClockers, etc) -- makes display motion blur photography much more accurate. If you want to test using a pursuit camera on a display,
read about other websites that use my invention, and then also
follow these pursuit camera instructions and also related
forum thread.
However, there is one specific situation where static is nearly as accurate as pursuiting, because of the ultrashort persistence time having no chance to motionblur: Strobe crosstalk photography is sufficiently accurate for static photographs as long as you photograph only one refresh cycle worth of strobe backlight flash.