View Single Post
Old 02-04-11, 09:00 AM   #1
Registered User
Join Date: Jun 2009
Posts: 47,113
Post More About the Strange White Banding Within Recent 3D LCD Monitors

If you remember the review of the 23‚?≥ Planar SA2311W 3D monitor I did not too long ago, you should also remember the interesting effect I've observed in the photo above (the banding on the white color). After testing a more recent model, namely the 27-inch Acer HN274H 3D monitor I have also observed the same strange effect with my extreme stereo 3D crosstalk test photos‚?¶

The photo above is of the monitor showing the 3D test photo and not trough the lens of the shutter glasses, so don't use it to judge for level of ghosting as it is not intended for that, but it is to investigate the strange gradation of white I've noticed. This test 3D photo uses black on white for the two eyes as this makes the strange color gradation of the white to be most apparent and it turns out that this issue is somewhat related to the OverDrive function as well as the current level of brightness (the contrast setting) of the monitor. If I disable the OverDrive trough the system menu the problem with the gradation of the white disappears, but in turn I get a lot of crosstalk which is to be expected. If I push the contrast setting (the brightness) of the display higher or lower there is also difference in the banding of the white with the white turning out to be completely even after a specific value.

Of course the black on white stereo 3D crosstalk test photos are an extreme case that should not normally happen in real world use scenarios or it will be very hard to be seen with normal stereoscopic 3D content displayed on the screen. Nevertheless I wanted to know exactly what is causing this strange behavior, how exactly it is related to the use of OverDrive and why it was not present in the previous 3D-capable LCD monitors that I've tested already. So I've been in contact with Planar to get the answers to this questions and they were very helpful in explaining the exact reasons, also confirming my observations regarding the OD and contrast settings. I'm going to quote the exact response I've got from Planar below:
I'm going to try to explain why banding can be seen with certain stereo test patterns on Planar's 3D Vision monitor. First, the goal of stereoscopic visualization is to show the left image only to the left eye and the right image only to the right eye. We want to prevent residual of the left image showing up when the right image is displayed, and visa-versa. Because lines addressed near the top of the display have more time to settle than lines at the bottom of the screen, the display electronics need to drive pixels more aggressively to switch from left to right as you get closer to the bottom of the screen. The display electronics has some limitations, where it can't change the 'aggressiveness' on a line by line basis. There are finite steps that the current drive electronics can control. This causes the banding, which can be seen on test images with constant gray values in one eye and high contrast in the other eye. However, as you stated in the article, this artifact is difficult to see in normal stereoscopic images because you would very rarely have large areas of a common gray level with high contrast between the left and right eye. If such an image were exhibited it would be uncomfortable to view (on any 3D monitor) because of the large discrepancy between the two eyes. This banding issue is below the threshold of visibility in real world images; just when guys like you or me test the extremes.

You asked why banding can be seen in test images on the new panel and not on the older ones. The older panels used one overdrive value for the entire display. This resulted in low stereo crosstalk in the middle of the display, but noticeable ghosting at the top and bottom of the screen. The new SA2311W monitor has low stereo crosstalk across the entire screen.

As for the reason that changing the contrast setting can eliminate banding, consider that with a low contrast setting you would drive a normal image from 0 to 85% (for instance) of the range available. This leaves some headroom for LCD overdrive which is used to improve 3D quality by reducing stereo crosstalk. In this case, you would see banding when a black image is shown in one eye and a high gray level (white) image is shown in the other. When you increase the contrast setting, you eliminate the headroom for overdrive. All the white values may be driven near 100% so that banding disappears, but then you observe ghosting. In general, a lower contrast setting results in better 3D quality for normal images.

The bottom line is that the appearance of this strange effect is a result from the efforts being applied in order to reduce the level of crosstalk/ghosting we are getting on the more recent 3D monitors and in order to have less ghosting at the top and bottom of the screen as compared to the earlier models. As it was already mentioned, this strange behavior is observed only with extreme test conditions and should not be easily noticeable in real world usage scenarios. And the white banding in these test photos will most likely be something that we are going to be seeing a lot from now on in the new 3D monitors that are about to come later this year, but you should not be worried about that fact. The important thing is that we are getting less and less crosstalk/ghosting with newer 3D-capable LCDs, right‚?¶ ;)

Related posts you might be interested in:

News is offline   Reply With Quote