This is false.
Originally Posted by irkuck
^DLP has practically no persistence (microseconds range).
Most DLP is usually currently fairly high persistence, so there is more motion blur on most DLP than on plasma/CRT displays, unless it uses interpolation or black frame insertion. Most DLP projectors currently fail the TestUFO Panning Map Test
at 960 pixels/second, you cannot read the street name labels because of the high persistence. Just try it. Try it and weep.
Transition/GtG = pixel switching/movement time
Persistence = pixel static/visibility time.Persistence is NOT the same as transition / GtG.
Start studying (use a stutter-free browser
for these animations to be accurate):
1. Animation of why persistence creates motion blur: www.testufo.com/eyetracking
2. Animation of how length of pixel static time coorelates to motion blurring: www.testufo.com/blackframes
3. Animation of various persistence lengths. Shorter pixel static time = less motion blur: www.testufo.com/blackframes#count=3
Note: These are software based animations using the full refresh cycle (8.3ms persistence steps at 120Hz, or 16.7ms persistence steps at 60Hz) in order to control persistence, so there will be lots of flicker, unlike high-frequency hardware-driven PWM, which can operate at frequencies far beyond the human flicker-detectability threshold.
Yes, DLP pixels can switch in microseconds, but DLP pixels do not "stay still" for just a microsecond.
That is NOT persistence. Persistence is NOT switching time. Persistence is STATIC time.The only way to create 1 microsecond persistence is essentially the following:
1. Flash a pixel for just one microsecond, once per refresh, any refresh. (this would be a very dark picture due to the million-to-one duty cycle).
2. Fill all timeslots. (1 million frame per second at 1 million Hz).
DLPs cannot do that, and virtually no technology currently achieves true 1 microsecond persistence; not even CRTs.
At 100% bright white, on most DLP projectors, a DLP pixel is typically continuously on for the whole refresh cycle (16.7 milliseconds at 60Hz). So you get 16.7ms persistence at 60fps@60Hz when not using interpolation, and when not using black frame insertion. As you track your eyes on moving objects on a screen, your eyes are in a different position at the beginning of a refresh than at the end of a refresh
. That causes the frame to be blurred across retinas, creating motion blur, as demonstrated at www.testufo.com/eyetracking
.... Mathematically, 1ms of persistence translates to 1 pixel of motion blurring during tracking your eyes on 1000 pixel/second full framerate=Hz moving objects. (This has been reliably true for clean square-wave persistence, such as LightBoost, BENQ Blur Reduction, and GSYNCs' ULMB feature, according to photodiode oscilloscope tests -- these are new low-persistence gaming monitors that has less motion blur than DLP/plasma, and actually has slightly less persistence than the Sony FW900 CRT -- finally, an LCD with less motion blur than certain CRT
). Softer persistence curves such as phosphor decayon CRT/plasma fuzzies this math up a bit, but the principle is still the same -- short-persistence CRTs have less motion blur/ghosting than long-persistence CRTs and thus remain true.
DLP pixels use PWM, which switches them on/off rapidly. On average, persistence of a DLP goes from the first visibility of the first PWM flash of a specific refresh, to the last visibility of the final PWM flash of a specific refresh. Typically, this is nearly the full refresh cycle -- e.g. 16.7ms of motion blurring at 60Hz (1/60sec = 16.7ms). Motion interpolation adds extra refresh cycles, and reduces persistence, so this is one common technique. Pixels of a specific refresh are visible for shorter times == shorter persistence. Black frame insertion is another technique of reducing persistence. Longer black period and shorter pixel visibility times == shorter persistence. There are numerous DLP projectors that does one or the other or both (interpolation or black frames, or both).However, DLP projectors *can* be made lower persistence
There are some good DLP's with fairly low persistence (<4ms). For example, DLP with large-ratio black frame insertion. It does reduce temporal time for the DLP to generate color. Some PC gamers found out
that on some DLPs such as Optoma GT720 by using true 120Hz (not stereoscopic 3D) combined with enabling black frame insertion, reduces persistence. On this model (GT720), enabling 3D mode even for non-stereoscopic 2D gaming, is another method of black frame insertion since 3D modes on DLPs adds a black period between refreshes as a guard delay for shutter glasses open/close -- this is effectively black frame insertion and can be reused for motion blur reduction too, since it reduces persistence. This creates only 4ms persistence and it actually barely passes the TestUFO Panning Map Test (readable map labels at 960 pixels/second). Most home theater DLP projectors, even $5000 models, are currently unable to run a Game Mode / Computer Mode with a low-persistence, but this is gradually changing.
As long as there is an acceptable tradeoff between persistence and color depth (lower persistence = less temporal time to generate colors), DLP are very capable of low-persistence. There is a 500Hz true-refresh-rate scientific DLP projector available at places like VPixx Technologies Inc. And I see nothing stopping virtual reality vendors using DLP chips as a method of creating low-persistence. However, it's extremely difficult for DLP to achieve 1 millisecond persistence while maintaing the full 8-bit color depth. (microsecond? Dream on. That's switch time, not persistence)
(Owner of BlurBusters -- considered an authoritative resource among PC gamers wanting "better than 60Hz" displays.)Edited by Mark Rejhon - 1/22/14 at 11:44am