AVS › AVS Forum › 3D Central › 3D Displays › How much crosstalk (ghosting) has your 3D TV? Test it!
New Posts  All Forums:Forum Nav:

How much crosstalk (ghosting) has your 3D TV? Test it! - Page 2

post #31 of 155
11/24/10 at 11:35am
Quote:
Originally Posted by upsilandre View Post

VT20 switch to 96hz mode when you use 24hz source (bluray 3D for exemple) and "24p film" enable
for my test i use .mpo 3D images (personal 3D crosstalk pattern) and my PS3 with the 3D viewer "Playmemories". i can switch "Playmemories" between 24hz mode (96hz page flipping on VT20) and 60hz mode (120hz page flipping) for compare and the difference is high

Thanks for sharing that.
I created a .MPO file of the image from the video that Nanite kindly provided which can be downloaded here:
http://dl.dropbox.com/u/7350380/testpattern.mpo
The side by side jpeg can be downloaded here:
http://3dfrank.com/testpattern.jpg
Reply
Reply

AVS Top Picks

post #32 of 155
12/2/10 at 9:29am
Thread Starter 
I got the Xpand X103 universal glasses from Amazon and they are much better (not perfect) than the original ones! I have about the same result in the crosstalk test with these but the bleeding is much dimmer. More proof that this test doesnt tell the truth about the overall 3D performance.

In practice there is less crosstalk. It feels that 95% of the irritating crosstalk is gone. Even the samples from Valentines Day works now. Crosstalk is still there but its so much dimmer that it doesnt destroy the 3D effect.

The glasses are also bit brighter and has less color tint in the 'open' state. They are also as comfortable as the original Samsung ones.

Only minor con is that the glasses has some glare on the corners, they reflect light from behind you.
Reply
Reply
post #33 of 155
12/13/10 at 9:50am
Panasonic P42GT25

Left eye open/right closed -> top half screen smaller visible number -> 1
Left eye open/right closed -> bottom half screen smaller visible number -> 10
Right eye open/left closed -> top half screen smaller visible number -> 10
Right eye open/left closed -> bottom half screen smaller visible number -> 1

Is the 638 series Mitsubishi as good as the 838?
Reply
Reply
post #34 of 155
12/13/10 at 10:18am
As noted earlier, my Mitsubishi 82738 had a perfect score on this test.
However, when viewing this side by side test pattern the crosstalk is quite evident.
http://3dfrank.com/indianheadtpsbs.jpg
Reply
Reply
post #35 of 155
12/13/10 at 12:25pm
Thread Starter 
Thanks for the .mpo file Frank. Now I could test the 'frame packed' performance with the PS3 'Play Memories' app. Much better results. Samsung plasmas doesnt perform that well with side-by-side material. Dunno if its the same with Samsung LCDs and other manufacturers sets.

Results from the .mpo file played with PS3 'Play Memories' 3D pic gallery app:
3D TV model: Samsung PS50C7705 50" Plasma (same as US PN50C8000)
3D glasses model: Samsung original battery powered & XPand X103 Unversal glasses.
The results are the same with both glasses but Xpand gives better real life results as the crosstalk is dimmer.

Left eye open/right closed -> top half screen smaller visible number -> 1
Left eye open/right closed -> bottom half screen smaller visible number -> 10
Right eye open/left closed -> top half screen smaller visible number -> 10
Right eye open/left closed -> bottom half screen smaller visible number -> 1

Results are the same with both 720p/60Hz and 1080p/24Hz.
Much better compared to side-by-side material 2-3 vs 10. There seems to be room for improvement in timings with Samsung plasmas in sbs mode.
Reply
Reply
post #36 of 155
12/13/10 at 5:48pm
Quote:
Originally Posted by Frank View Post

As noted earlier, my Mitsubishi 82738 had a perfect score on this test.
However, when viewing this side by side test pattern the crosstalk is quite evident.
http://3dfrank.com/indianheadtpsbs.jpg

If you move your head around do you see an area that's blocked also move? I think you're seeing the limits of the tech in today's LCD shutter glasses. From Andrew Woods's crosstalk paper:

Ordinarily the only crosstalk present with timesequential
3D on DLP is due to the LCS glasses.
The important contributors to crosstalk for time-sequential
3D on DLP are:
The performance of the liquid crystal cells in the shutter
glasses (see Section V-A), and
The timing of the shuttering of the glasses with respect to
the display of images on the screen (see Fig. 14).
Reply
Reply
post #37 of 155
12/14/10 at 9:04am
I will try the test tonight on my Samsung UN46C7000, but one interesting thing I noticed was that the crosstalk of white text on a black background is evident on my display whereas black text on a white background there is zero apparent crosstalk even when both are on the same screen. Not sure what this means as in both cases you will have rapid switching of pixels from black and white to produce the 3D image.
Reply
Reply
post #38 of 155
12/16/10 at 7:15pm
POPPABK, cross-talk varies from scene to scene, but is especially noticeable when the picture element that's closest to the viewer in the 3d field is also the brightest... so if the white text in your example is intended to be "closer" than the black element, you will more likely experience cross-talk. It's less pronounced when the darker element is at the forefront.

As per expert opinion, Cross-talk should be a non-issue on PDPs (plasma displays), but occurs mainly when the glasses' LCD based shutters fail to block out enough light when the shutters are in the "off" position. To eliminate any cross-talk caused by bright elements in high contrast situations, the shutters must be able to block out 100% of the light, which is impossible with Liquid Crystals... this is why manufacturers pre-tint the optics of their glasses to block out about 25% of light from the start.

Also, the LCDs in these glasses do not transition from "on" to "off" fast enough to keep up with the naturally faster dissipating plasma phosphors... this can cause some of the cross-talk effects more commonly seen in LCD-LED 3D diplays with content running over 24fps.

Why manufacturers chose not to supply better adapted glasses to their lines of plasma displays is quite beyond me, but as an intermediate level HDTV calibrator & avid Videophile, I can tell you that the problem lies in the glasses (& sometimes video processing), not the Plasma panels themselves.

Lets hope PDP manufacturers realize this & produce better adapted eye wear for our picture perfect Plasmas.
Reply
Reply
post #39 of 155
12/21/10 at 10:11am
Great Tester!

LG 55LX6500 (mp4 Standard). I'll give the mpo a try as well.

Left eye open/right closed -> top half screen smaller visible number -> 1
Left eye open/right closed -> bottom half screen smaller visible number -> 6
Right eye open/left closed -> top half screen smaller visible number -> 6
Right eye open/left closed -> bottom half screen smaller visible number -> 1

(mp4 Cinema/Game)

Left eye open/right closed -> top half screen smaller visible number -> 2
Left eye open/right closed -> bottom half screen smaller visible number -> 10
Right eye open/left closed -> top half screen smaller visible number -> 11
Right eye open/left closed -> bottom half screen smaller visible number -> 3


Glad to see we are not alone.
Reply
Reply
post #40 of 155
12/21/10 at 12:34pm
LG 47LX6500 (MP4) - Viewed from in TV media player.

Left eye open/right closed -> top half screen smaller visible number -> 1
Left eye open/right closed -> bottom half screen smaller visible number -> 4
Right eye open/left closed -> top half screen smaller visible number -> 1
Right eye open/left closed -> bottom half screen smaller visible number -> 1

About as bad of crosstalk as it gets I'm afraid.
Reply
Reply
post #41 of 155
12/21/10 at 9:39pm
Has anyone run this test on one of the Samsung 3d Plasmas after the update the supposidly helps with the cross talk?
Reply
Reply
post #42 of 155
12/22/10 at 7:47pm
I'm really jealous of you guys seeing "NOTHING" where you're supposed to see nothing...

FWIW, I have an LG 47LX6500 with similar results to dburckh.
Reply
Reply
post #43 of 155
1/2/11 at 1:52pm
I'd be curious to hear what kind of crosstalk the new passive TV from Vizio has.
Reply
Reply
post #44 of 155
1/3/11 at 4:27am
Quote:
Originally Posted by BlackShark View Post

This test looks interesting but could be improved : it only shows absolute black/white contrast. It would be very interesting to see what happens when you use various grey backgrounds all the way up to pure white. It would show the inverse crosstalk that many LCD displays have (and improve the score of other displays a lot, since bright numbers on black background is the worst case scenario for plasma and polarised technology)

Here is one of my displays :
3D display : Zalman Trimon ZM220 computer monitor (2008)
3D glasses : Zalman 3D glasses shipped with the display

This display is a custom Zalman circular polarised 3D interlaced display, compatible with RealD but 45° off angle from RealD standard. The picture is still watchable but ghosting is 50% stronger and changes coulour to purple if used with RealD glasses
Also due to the interlacing of the µpol filter technology, this display has a sweet spot, horizontal view angle is pretty wide but the vertical on is very very narrow, you have to be at the right height otherwise you get total crosstalk. The values reported here are when watched inside the sweet spot

The display is very shiny, crosstalk tends to get hidden behind reflections.
Brightness 50%, Contrast 50%

DARK ROOM (no reflections)
Left eye open/right closed -> top half screen smaller visible number -> 1
Left eye open/right closed -> bottom half screen smaller visible number -> 10
Right eye open/left closed -> top half screen smaller visible number -> 10
Right eye open/left closed -> bottom half screen smaller visible number -> 1

BRIGHT ROOM (lots of reflections)
Left eye open/right closed -> top half screen smaller visible number -> 2
Left eye open/right closed -> bottom half screen smaller visible number -> 20
Right eye open/left closed -> top half screen smaller visible number -> 20
Right eye open/left closed -> bottom half screen smaller visible number -> 2

Greetings BlackShark.

It would be interesting to see the performance of your polarised dual-projector system.

Thanks and best regards.
Reply
Reply
post #45 of 155
1/7/11 at 7:17am
I really enjoyed running this test on my 60" LG PX950. Unfortunately, it faired about the same as most of the results posted... I am impressed that the DLP showed nothing at all when it was supposed to... To think, I bought the Plasma because I read 3D performance was better than LCD. I still got a great deal, and one test doesn't mean real world 3D watching is awful, but we all want our equipment to work.

What's funny to me, is that if I put my desktop background up on the TV, and turn on SBS 3D, and close one eye, I see exactly half the desktop background, if I close the other eye, I see the other half. My background isn't symmetrical, and I looked for pieces of one side of the background in the other, but never saw them... Is it just a problem with pure black backgrounds vs non-black on the other side?
Reply
Reply
post #46 of 155
1/15/11 at 4:45am
Hello again i ran the Test on the dual projector setup and here are the results :

My passive polarised 3D dual projector setup :
Dual Epson EH-TW3500 (euro version of the 8100)
Silverfabric linear polarised filters POLAR-L-50
Harkness Spectral 240 "silverscreen"
Stereomax linear polarised glasses

Left eye open/right closed -> top half screen smaller visible number -> 1
Left eye open/right closed -> bottom half screen smaller visible number -> 15 (normal) 20 (tuned for lowest crosstalk but unbalanced colour)
Right eye open/left closed -> top half screen smaller visible number -> 15 (normal) 20 (tuned for lowest crosstalk but unbalanced colour)
Right eye open/left closed -> bottom half screen smaller visible number -> 1

(edited, interverted two lines)
Reply
Reply
post #47 of 155
1/15/11 at 8:34am
Quote:
Originally Posted by BlackShark View Post

Hello again i ran the Test on the dual projector setup and here are the results :

My passive polarised 3D dual projector setup :
Dual Epson EH-TW3500 (euro version of the 8100)
Silverfabric linear polarised filters POLAR-L-50
Harkness Spectral 240 "silverscreen"
Stereomax linear polarised glasses

Left eye open/right closed -> top half screen smaller visible number -> 1
Left eye open/right closed -> bottom half screen smaller visible number -> 15 (normal) 20 (tuned for lowest crosstalk but unbalanced colour)
Right eye open/left closed -> top half screen smaller visible number -> 1
Right eye open/left closed -> bottom half screen smaller visible number -> 15 (normal) 20 (tuned for lowest crosstalk but unbalanced colour)

Greetings BlackShark. Thanks for posting the numbers.

How would you rate the score of 15 subjectively, and would you be able to give a rough estimation in terms of the percentage of polarised light rejected?

DLP TVs and projectors seem to be able to achieve perfect crosstalk elimination. Is there any inherent advantage in this technology?

Best regards.

Quote:
Originally Posted by Frank View Post

Mitsubishi 82838 DLP
I just downloaded the file and dragged it into stereoscopic player and viewed it in a 3D window and got these results.

Left eye open/right closed -> top half screen smaller visible number -> 1
Left eye open/right closed -> bottom half screen smaller visible number -> NOTHING
Right eye open/left closed -> top half screen smaller visible number -> NOTHING
Right eye open/left closed -> bottom half screen smaller visible number -> 1

Viewing it full screen yielded the same results.
Reply
Reply
post #48 of 155
1/16/11 at 5:24am
Thread Starter 
15 is quite good result. Other factor is how dark are the bleeding numbers (off transparency). This test doesnt measure it and its also a big factor that affects the 3D experience.
Reply
Reply
post #49 of 155
1/16/11 at 5:49am
Crosstalk on passive polarised systems is always identical, it's a linear bleed relative to the contrast between the two eyes.
15 is very dark, barely visible, and it progressively gets brighter and brighter to reach it's maximum at the end.

It's impossible for me to be able to tell how much light is bleeding/blocked in %, the human eye has a very non-linear response to the amount of light.
For instance , when I put a filter in front of one projector, I know technically that I instantly loose about 50~55% of the light but visually it looks like I only lost 20% or so.

Similarly, I know from the manufacturer specs that crosstalk is somewhere around 0.9~1.5% because the polarisation extinction ratio of the screen is 130:1 peak, and averaging throughout the field of view somewhere around 100:1, however visually speaking it goes from 0% (no ghosting visible) straight to something around 10 % (on extreme high contrast / crosstalk test scene).

In order to provide an actual measurement, I'd need an instrument to precisely measure the light, which I do not have.
Reply
Reply
post #50 of 155
1/20/11 at 2:54pm
Just tested an LG 55LX9500 using XPAND 103s and a custom 3D ghosting pattern fed from a VideoForge. The ghosting on the LG was 10 on a 0-12 scale. By comparison we tested a JVC RS40 pj and it was less than 1 on the same scale using the same glasses.
Reply
Reply
post #51 of 155
1/22/11 at 10:16pm
Thanks everyone for posting in this thread as I am very interested in your findings. I am looking for a 3D display right now and I have been reviewing these posts for help choosing the best one. Could someone please help me understand what the numbers mean in terms of performance? I understand what the best performance is 1,nothing,nothing,1, but does higher or lower numbers mean that it is better? What is the best 3d plasma out there right now? Thanks so much everyone.
Reply
Reply
post #52 of 155
1/23/11 at 3:08am
Thread Starter 
Quote:
Originally Posted by WiFi-Spy View Post

using XPAND 103s and a custom 3D ghosting pattern fed from a VideoForge. The ghosting on the LG was 10 on a 0-12 scale.

That custom test dont tell nothing to us (higher number good or bad?). Could you do the test using test pattern from the first post? Are the XPands performing better than the original ones?

Quote:
Originally Posted by gr3y3y3s View Post

but does higher or lower numbers mean that it is better?

Higher number is better. This test does not tell the whole truth about the 3D crosstalk. This pattern doesnt test the overall off-transparency. For example there could be two TV models with same score but the other has numbers much darker (less transparency). I see this happening with same TV with two different glasses models also.

There are also other factors to 3D performance. Black level (dynamic range), color space etc.
I am not so happy with black level with my Samsung plasma in 3D mode.
You should wait for 2011 models.
Reply
Reply
post #53 of 155
1/23/11 at 11:48am
I have the 65638, and while I can angle my head to block out even the highest numbers, there's no way for me to block out every element due to the shutterglasses having an uneven coverage, apparently. The default Mitsubishi/Samsung IR glasses are pretty sensitive to head tilt, while the ViewSonic PGD-150's fare a little bit better. They vary in how large an area they cover from the center of the lenses where they're most effective. I can only sit back 9 feet max, but I can get the Viewsonic 150's to cover the entire screen almost entirely, though I need to have my head at the perfect, perfect angle. The Mitsu IR's are going to require me to sit back probably to 13-15ft, though I can't get that far back in my room.

So this confirms for me that current IR glasses can contribute to ghosting. It's only about 1% of the time, but it would be nice to have that perfect isolation in all cases. Anyone have universal IR glasses recommendations? I've also tried the XPAND X103's and they were about the same as the ViewSonics. I am leery about DLP Link now because I get these brief flickers occasionally (no sunlight, and just a dim halogen lamp behind me).
Reply
Reply
post #54 of 155
1/23/11 at 12:56pm
Samsung UN46C7000
I tried this but my results are totally inconsistent and highly dependent on settings and viewing location.

Backlight 20, viewed directly from front:
Left eye open/right closed -> top half screen smaller visible number -> 1
Left eye open/right closed -> bottom half screen smaller visible number -> 20
Right eye open/left closed -> top half screen smaller visible number -> 10
Right eye open/left closed -> bottom half screen smaller visible number -> 1
Backlight 8, viewed directly from front ~7 feet:
Left eye open/right closed -> top half screen smaller visible number -> 1
Left eye open/right closed -> bottom half screen smaller visible number -> 30
Right eye open/left closed -> top half screen smaller visible number -> 22
Right eye open/left closed -> bottom half screen smaller visible number -> 1
Backlight 8, viewed directly from front ~1 foot (glasses still operating):
Left eye open/right closed -> top half screen smaller visible number -> 1
Left eye open/right closed -> bottom half screen smaller visible number -> 1
Right eye open/left closed -> top half screen smaller visible number -> 1
Right eye open/left closed -> bottom half screen smaller visible number -> 1
Backlight 8, viewed 45 degree angle from front ~7 feet,:
Left eye open/right closed -> top half screen smaller visible number -> 1
Left eye open/right closed -> bottom half screen smaller visible number -> 2
Right eye open/left closed -> top half screen smaller visible number -> 2
Right eye open/left closed -> bottom half screen smaller visible number -> 1
Backlight 8, viewed directly from front ~7 feet, dynamic contrast off:
Left eye open/right closed -> top half screen smaller visible number -> 1
Left eye open/right closed -> bottom half screen smaller visible number -> 42
Right eye open/left closed -> top half screen smaller visible number -> 37
Right eye open/left closed -> bottom half screen smaller visible number -> 1
Backlight 20, viewed directly from front ~7 feet, dynamic contrast off:
Left eye open/right closed -> top half screen smaller visible number -> 1
Left eye open/right closed -> bottom half screen smaller visible number -> 36
Right eye open/left closed -> top half screen smaller visible number -> 29
Right eye open/left closed -> bottom half screen smaller visible number -> 1
Reply
Reply
post #55 of 155
1/23/11 at 5:04pm
Quote:
Originally Posted by poppabk View Post

Samsung UN46C7000
I tried this but my results are totally inconsistent and highly dependent on settings and viewing location.

Backlight 20, viewed directly from front:
Left eye open/right closed -> top half screen smaller visible number -> 1
Left eye open/right closed -> bottom half screen smaller visible number -> 20
Right eye open/left closed -> top half screen smaller visible number -> 10
Right eye open/left closed -> bottom half screen smaller visible number -> 1
Backlight 8, viewed directly from front ~7 feet:
Left eye open/right closed -> top half screen smaller visible number -> 1
Left eye open/right closed -> bottom half screen smaller visible number -> 30
Right eye open/left closed -> top half screen smaller visible number -> 22
Right eye open/left closed -> bottom half screen smaller visible number -> 1
Backlight 8, viewed directly from front ~1 foot (glasses still operating):
Left eye open/right closed -> top half screen smaller visible number -> 1
Left eye open/right closed -> bottom half screen smaller visible number -> 1
Right eye open/left closed -> top half screen smaller visible number -> 1
Right eye open/left closed -> bottom half screen smaller visible number -> 1
Backlight 8, viewed 45 degree angle from front ~7 feet,:
Left eye open/right closed -> top half screen smaller visible number -> 1
Left eye open/right closed -> bottom half screen smaller visible number -> 2
Right eye open/left closed -> top half screen smaller visible number -> 2
Right eye open/left closed -> bottom half screen smaller visible number -> 1
Backlight 8, viewed directly from front ~7 feet, dynamic contrast off:
Left eye open/right closed -> top half screen smaller visible number -> 1
Left eye open/right closed -> bottom half screen smaller visible number -> 42
Right eye open/left closed -> top half screen smaller visible number -> 37
Right eye open/left closed -> bottom half screen smaller visible number -> 1
Backlight 20, viewed directly from front ~7 feet, dynamic contrast off:
Left eye open/right closed -> top half screen smaller visible number -> 1
Left eye open/right closed -> bottom half screen smaller visible number -> 36
Right eye open/left closed -> top half screen smaller visible number -> 29
Right eye open/left closed -> bottom half screen smaller visible number -> 1

There is no way you are getting those results with that tv, sorry bro. Your results do not come close to the results posted by other owners of your tv
Reply
Reply
post #56 of 155
1/23/11 at 5:26pm
Quote:
Originally Posted by neo0285 View Post

There is no way you are getting those results with that tv, sorry bro. Your results do not come close to the results posted by other owners of your tv

Those are definitely the results. TV had been on all day in a warm room, which may have improved the results. I am gonna try while cold at a later date. What I don't understand is why the left and right eyes are different, I've got a feeling it is a scanning speed issue. I need an inverted test pattern to test it though.
Oh yeah, test was also done in daylight, it might be different in the dark.
Reply
Reply
post #57 of 155
1/23/11 at 8:50pm
I hate to say this, but the graphic you are using to measure the crosstalk is not going to give you a meaningfull figure, I think.

To truly measure the crosstalk, ie the extinction ratio, you must use a 100 % white signal from the opposing eye, which should be blanked out by the eye that shouldn't see it.

Only then, by measuring the percentage of the signal that gets thru to the other eye, from this 100 % signal, can you judge the extinction/crosstalk.

I think you'd be better off using something like I've attached below - the eye that is ALLOWED to see, will see 30 boxes, with grey scales from 1 % to 30 % white (against a 0 % Black background). The opposing eye gets 100 % white circles, which fit into the 0 % cutouts in the boxes.

However much crosstalk is appearing, will show inside the boxes, and you must match the greyscale against the surrounding box to get a match on the correct level of crosstalk.

The signal used MUST be 100%, and you then judge how much of that 100% signal got through to the eye that shouldn't have seen it.

In a perfect system, with no crosstalk, all the boxes would show a pitch black circle - the 100% white circles on the opposing eye would be blocked entirely. I only went to 30%, as anything that does worse that that should be laughed out of the room and is unsuitable for discussion .

Below are examples of the left and right images that I feel would give a much more accurate idea of the extinction ratio, and the crosstalk (they are reduced size, of course, since they are only attachments - but you get the idea:
LL
LL
Reply
Reply
post #58 of 155
1/24/11 at 11:38am
That is a very very interesting pattern rdjam which allows not only seeing ghosting but also measuring it.
The only thing I think it lacks is the luminance value of each square to make it easier to translate into a text result, could you add it ? (luminance 0->255 scale would probably be more accurate than a percentage)
could you also make a side-by-side version (to make it easier to launch in any 3D photo viewer)

However I disagree with your remark considering the usefulness of Nanite's ghosting test. The two tests have different objectives.

Your test consists in throwing the maximum brightness and seeing how much goes through.
Nanite's test consists in finding at what threshold crosstalk becomes a problem (at what point do you start to see it)
Your test could also give the result of Nanite's test if you pushed further with brighter squares

Finally I'd add that these tests only show one or two elements in the crosstalk characteristics : none of these tests show the big picture because in only measures maximums and minimums. Passive polarisation has a linear response so it's ok but LCD screens with shutter glasses have demonstrated issues with inverse crosstalk depending on the colour combinations (not a light bleed, but a light gap) due to the LCD overdrive overcompensating during the fast 120Hz switch between the left and right views.
Reply
Reply
post #59 of 155
1/24/11 at 3:54pm
Quote:
Originally Posted by BlackShark View Post

That is a very very interesting pattern rdjam which allows not only seeing ghosting but also measuring it.
The only thing I think it lacks is the luminance value of each square to make it easier to translate into a text result, could you add it ? (luminance 0->255 scale would probably be more accurate than a percentage)
could you also make a side-by-side version (to make it easier to launch in any 3D photo viewer)

However I disagree with your remark considering the usefulness of Nanite's ghosting test. The two tests have different objectives.

Your test consists in throwing the maximum brightness and seeing how much goes through.
Nanite's test consists in finding at what threshold crosstalk becomes a problem (at what point do you start to see it)
Your test could also give the result of Nanite's test if you pushed further with brighter squares

Finally I'd add that these tests only show one or two elements in the crosstalk characteristics : none of these tests show the big picture because in only measures maximums and minimums. Passive polarisation has a linear response so it's ok but LCD screens with shutter glasses have demonstrated issues with inverse crosstalk depending on the colour combinations (not a light bleed, but a light gap) due to the LCD overdrive overcompensating during the fast 120Hz switch between the left and right views.

I hear you. But low luminance crosstalk is hardly noticeable most of the time. But just because you can't notice it, doesn't mean it's not there.

If you are going to measure it, you must use full luminance interference, so you can see the percentage. If there is 20% crosstalk at full luminance, there is still 20% crosstalk at low luminance too, but you just might not see it.

Yeah, I definitely will add some labeling. It's a first effort to try to test various screen materials for polarisation retention in the Ultimate 3D Projection System thread. I'll get it pretty fairly soon

Yes, and I agree on the other patterns. Because of my focus on passive at the moment, I havent given much thought to other scenarios in the big picture.
Reply
Reply
post #60 of 155
1/25/11 at 2:13pm
Quote:
Originally Posted by rdjam View Post

If you are going to measure it, you must use full luminance interference, so you can see the percentage. If there is 20% crosstalk at full luminance, there is still 20% crosstalk at low luminance too, but you just might not see it.

Well the problem with crosstalk and extinction ratios is.... they are different things. The extinction ratio is just a single measurement (the one that is the most interesting to market : the best value), crosstalk covers the entire picture at any colour and any contrast level.
It's true passive polarised have a linear crosstalk (you get x% at any brightness), however even with passive polarised, it is technically possible to end up with materials with a non-linear response.
Actually all materials have a non-linear response, but you need extremely high light intensity to notice it, I do not know if any material with non linear response at usual light levels are actually available outside of research laboratories.
Reply
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: 3D Displays

AVS Top Picks

AVS › AVS Forum › 3D Central › 3D Displays › How much crosstalk (ghosting) has your 3D TV? Test it!