AVS Forum banner
Status
Not open for further replies.

Motion blurring on PDP and LCD displays

926 views 19 replies 9 participants last post by  xrox 
#1 ·
Have been doing some reading on digital (pixel based) displays and motion-blurring. I thought that pixel response time was the only parameter related to motion-video sharpness. But interestingly it is not.


All pixel based displays have the same problem. They are hold-type displays. This means that each pixel is emmiting during the entire duration of the frame (1/60th of a second) in order to maximize luminence.


CRT’s on the other hand are impulse displays where each point (phosphor dot) is only excited once during the trace of each frame.


So if you display two frames with fast motion on a hold-type display you will integrate the two frames visually and perceive motion-blurring or edge-blurring. So even the fasted LCD response time will still not come close to a crt as far as motion-blurring is concerned.


There are ways around it being developed.


Also, I’m reading up on Dynamic False Contouring which Plasma displays are subject to due to the way they produced grayscale.


Anyways, just posted for interest sake.......
 
See less See more
#2 ·
There is a lot of difference in the way plasmas illuminate and the way LCDs illuminate, though.


Also, a "field" is 1/60 of a second and a "frame" is 1/30 of a second.


You can use mini bursts to illuminate the plasma of very small duration -- and theoertically several times per frame. On an LCD, by contrast, the light is always on... The only question is how 'blocked" it is by the liquid crystal. In theory, the transistor could "pulse" on and off, but it probably wouldn't matter given the response time of the crystal.


Plasma, in other words, is more capable here and better. But as LCD response times fall to the 7-12ms range -- which will be happening over the next 2 years -- this advantage should be greatly mitigated. Unless further advances in plasma -- like faster phosphors for example -- become more common.


Mark
 
#3 ·
Rogo,


Very true, they do illuminate very differently (even though LCDs do use plasma).


But what I was trying to point out was that no matter what response time your pixels have (even nanoseconds). You will still have motion blurring due to the hold-type way of displaying.


Yes, plasma uses bursts of luminence over the duration of each frame but this is how they produces grayscale. The pixel is still illuminated during the entire duration of each frame in order to maximize overall luminence.


Also, when I say 1/60th of a second per frame I am talking about progressive scan displays which do not use interlaced fields. But plasmas do use subfields to produce grayscale.


------------------------------------------------------------------------------


Now, that being said, they have come up with patch solutions:


1) inserting black frames between every frame (increases flicker though)

2) Use subfield motion-compensation


Exactly how or what plasma displays use to overcome the hold-type limitations is tough to find out. But as I said before I just posted this for interests sake.
 
#5 ·
Dsmith,


It's interesting insn't it...............


The backlights in LCDs work on the same principle as each pixel in a plasma. Gas is ionized by electrical potential and the ionized gas emits ultra-violet light which excite phosphor which in turn emits visible light.


Plasma is just ionized gas.


In the case of LCDs the gas is mercury-vapor and the phosphor is white-phosphor which emits bright white light when excited by ultra-violet light.
 
#7 ·
Sorry, forgot to add that in the case of Plasmas the gas used is usually a mixture of helium-xenon or neon-xenon along with seperate red, green, and blue phosphors.


Several different types of phosphors have been used in Plasma displays as research has improved both the relaxation time of the phoshor and the lifetime.
 
#9 ·
xrox,


Thanks for the interesting info.


Without attempting to contradict such technical info, in practice I notice little motion blurring on my Panasonic plasma. (Especially in analog NTSC channels).


Once things get "really digital," e.g. digital source etc, I find motion artifacts can start becoming more bothersome.


Strangely enough, when plasmas first started showing up in the stores I was put off by the motion artifacts. Now, I find myself more annoyed by the motion and picture artifacts in many digital tube CRTs. They use what look like cheap 'n dirty processing to for line doubling/pull down etc. Even when I'm watching DVDs on a Loewe CRT, I see a stuttering to motion that is more intrusive than any I encounter on my plasma.

.
 
#10 ·
R. Harkness,


I too can not see any motion blurring on the Panasonic plasma. I haven't looked at to many others. The Pio and the Sony I looked at did show smearing during fast motion(albiet very slight). But I guess current models use some sort of solution for this problem. Either they use sub-field motion-compensation or black frames, or some complex integration software, i dunno? I really don't quite understand it all.........


As for LCDs, they must first overcome the pixel response issue before the hold-type problem becomes noticeable. And according to the papers I've read it is very much noticeable without a good solution in place.


I found it interesting though


Cheers,

xrox
 
#11 ·
Quote:
Originally posted by xrox



As for LCDs, they must first overcome the pixel response issue before the hold-type problem becomes noticeable. And according to the papers I've read it is very much noticeable without a good solution in place.

there are so many new LCD's being advertised with much faster response times that they are being compared to plasmas- not sure if I believe all the brightness/repsonse time specs the LCD makers claim


from what I have seen there is still no comparison: even with the newest/fastest LCD's- they still do not come close to the response time of a plasma


perhaps this will this change in 1 or 2 years but I am skeptical


Mark
 
#12 ·
Plasma is not just "ionized gas." Ionized gas is gas that has become ionized. Plasma is actual the "fourth state of matter" and describe a level of excitation beyond more ionization. It is much harder to achieve than what goes in a fluorescent light bulb, even though it is similar in how its achieved.
 
#13 ·
Markrubin,


I agree. I have been quoted 3 different response times for the Sharp 30" LCD panel. I've heard 20ms, 16ms, and 12ms. I've also been told that anything under 20ms is not noticeable.


I think this is bull! I can definitely see smearing and blurring during motion playback.
 
#14 ·
Rogo,


I don't believe so.


the "fourth state of matter is............." Ionized Gas.


trust me on this.....


And yes, flourescent bulbs, neon signs, HID lamps and any other ionized gas is a PLASMA.


In the case of Plasma TV, the ionized gas is a mixture of Helium-neon-xenon



Cheers,

xrox
 
#15 ·
Xrox, I'll take your word for it as I stopped taking physics in high school.


Still, there is this for your perusal, whatever the heck it means.


:confused:


" Now the "Electric Universe" is a different viewpoint. Notice, for example, that its definition of "plasma" is not the conventional one of "ionized gas". That latter definition jumps to the conclusion that you can understand something about plasma by falling back on what you know about ideal gasses and thermal ionization. The ideal gas law is an important insight in the conventional view, but it becomes a blindfold in the electric view, preventing you from seeing what's before your eyes. Rather, "plasma" is an emergent (i.e., higher level or statistical level) orderliness of complex electrical forces: such properties as filamentation, long-range attraction and short-range repulsion, braiding, characteristic velocities, formation and decay of plasmoids, and identity of properties at different scales."


Mark
 
#16 ·
I hated physics in high school but somehow I ended up working in the field and loving it. Go figure???


I really like that quote you posted. It tries to describe plasma as a specific state during ionization (kinda like when all the planets align). It goes against common definitions but I won't go as far as saying it's bull. It is more likely that it is just beyond me (way beyond me LOL)


Anyway,

Just for your interest I will try and explain how a plasma ball works. It is kinda cool. If you ionize gas at very low pressure you get a glow discharge (such as the case in flourescent lights and neon lights. If the gas is at very high pressure it becomes very resistive so you require a huge electrical potential to arc through the gas such as in a lightning bolt and HID lamps on those ubiquitous BMW's.


Now, a plasma ball is unique because it is at an in-between pressure (a very specific one) that "just barely" allows ionization. Since the gas is so resistant to ionization the electrons struggle to jump through the gas and this appears as waving fingers of plasma.


coolness (I'm sad, I know)


Anyway, I apologize for getting way, way, way off topic


Cheers,

xrox
 
#17 ·
xrox,


a few things:

1. LCD blurring is composed of three totally separate blur issues:

a. pixel response time

b. pixel decay time.

c. processing blur (read below).


2. In essence, each pixel on the plasma is actually three separate entities bundled together - R, G, and B subpixels. These flicker independantly of eachother and for various plasmas even have different electrical properties and are even sized differently (blue usually receiving the larger space).


These subpixels flicker all the time. In fact, there is no state in which they do not flicker. They flicker to produce individual gray scales.


They instantly flicker from one luminance value to another as the frame changes.


HOWEVER, most plasmas do not change pixel values at the input rate (this is completely different than CRTs, which use the refresh rates as the rate the pixels are ignited).


MOST plasmas have their own internal rate at which the screen is refreshed. This has been demonstrated most recently by the Pioneer being internally marked at between 70-72Hz.


The internal scaler compensates for this by producing "tween frames", which are basically transitionary frames that reduce the preceived distance that pixel vector have to travel between two separate source frames.


This produces a smoother motion, but can introduce other motion artifacts (usually not that bad, though, unless you're feeding in the wrong rate through DVI, which causes a severe tearing artifact).


3. Of course, this is problematic for rates that are so utterly different from 70-72Hz. The worst of this is in PAL, where the rate (50Hz for video sources) is so completely different then the 70-72Hz that the plasma runs natively at.


Feeding such sources causes a severe artifact with plasmas that do not support 2:2 pulldown (those can reduce the rate to 25 fps which allows sufficient tween frames). It's an artifact that is something between smearing and banding and is apparent primarily on on high motion pixels.


This artifact is not unique to plasmas, but occurs on many sample & hold devices, such as DLPs. We've conducted a test about a year ago, where we compared Sharp 9000U to Sharp 9000E. The only technical difference between the units is 2:2 pulldown. The 2:2 pulldown doubles the amount of tween frames. Now, forget the PQ improvement due to 2:2 pulldown, and talk about flicker. There was no flicker on either one, but there was motion judder (similar to flicker) on the U unit (non 2:2 pulldown) and there was the banding/smearing artifact too. The Sharp 9000E worked perfectly with no judder or banding/smearing on the exact same source.
 
#18 ·
Quote:
Originally posted by xrox
I have been quoted 3 different response times for the Sharp 30" LCD panel. I've heard 20ms, 16ms, and 12ms. I've also been told that anything under 20ms is not noticeable.


I think this is bull! I can definitely see smearing and blurring during motion playback.
Let me get this straight. Aren't we talking about 'milli-seconds' (1/1000th of a second) here ?? NTSC runs at 60Hz, and it seems to me that LCDs response time is A LOT faster than 60 Hz. (solely according to their spec...milli-seconds!)


If you see the motion blurring on LCDs, do you also notice it on DLPs ??

I maybe wrong but, the 'motion blurring' you mentioned, isn't this actually temporal dithering (motion dithering) artefact ?
 
#19 ·
Oferlaor,


You pretty much got it right with the integration time being too small. Just picture one pixel during the playback of motion video. OK, now during the duration of one frame the pixel fires 5 times (5 subfields). But the next frame has some spatial movement. So the pixel will fire 5 more times displaying this spatial movement.


But since there is not enough time between frames for our brains to process the info we see the two pixels (same pixel on different frames) blended due to temporal integration. In other words we se blurriness even though it is not there.


Quote:
Originally posted by oferlaor
xrox,


a few things:

1. LCD blurring is composed of three totally separate blur issues:

a. pixel response time

b. pixel decay time.

c. processing blur (read below).

Ok, when you say pixel decay time what are you referring to? Do you mean the fall time? LCDs pixels have both rise and fall times while plasma has only fall time due to phosphor decay.


As for processing blur you must mean "motion-blurring" which is the term industry uses to describe pixel-integration blurring (described above) of hold-type displays (also, see my first post). The actual blurring has nothing to due with processing though. Unless you are talking about de-interlacing which does introduces process motion artifacts.


You're on the right track when it comes to refresh rate. The insertion of black frames has traditionally been a patch for motion-blurring on hold-type displays but the problem is it introduces flicker and reduces luminence (NOT GOOD)


I've done some digging and found out that the Fujitsu 50" model uses a subfield-motion-compensation (not talking about de-interlacer here) algorithm instead of black frames. I could write what I've read if anyone is interested.


Not only does this improve motion-blurring but also reduces dynamic false contouring.


I have no idea what other models use though?
 
#20 ·
Nusim,


The motion blurring I was referring to is not a pixel response time issue but a temporal integration problem as you mention.


It’s all very confusing and I’m not helping much explain it.


Here goes again,


OK, you have two frames. There is motion between frames (ie an object has moved, so in frame 1 the object is not in the same place as frame 2)


Now, on a hold-type display (which means the pixels are emmiting light over the entire duration of every frame) there is not enough inactive time (black pixel) between frames for our brains to properly resolve the two pixels in question. We integrate them and thus see a blurred image.


Now on a CRT each pixel is excited only once during the entire trace of each frame so there is more time for our brains to resolve the two pixels in question. But (and I’m just reading this now) if you increase the refresh rate to >100Hz you start to see the same problem.


Is that any clearer? Sorry if I make no sense, I’m learning it as I read.
 
Status
Not open for further replies.
You have insufficient privileges to reply here.
Top