service menu adjustments to improve PQ
Since I got my HLR5067w 6 weeks ago, I've been bothered by
(1) the obscenely high contrast of the default settings;
(2) obvious "graininess"/"noise"/"texture" of dark areas in low contrast scenes;
(3) extreme green cast to the same dark areas of low contrast scenes.
by fiddling around with various setting in the service menus, I've virtually eliminated all these problems without turning off DeNIE. I blew $30 for the service manual, which was a total waste of money unless you're interested in making board level repairs or understanding the details of the video circuits (I was interested only in calibration tips--there weren't many).
(1) can be fixed by turning down the user menu contrast settings from the default of 90 to about 60--70 (for the "standard" preset). OR:
Except for "my color control", all user menu settings are reset to the factory defaults every time you enter the service menus (how incredibly stupid engineers can be!) so if you're happy with using the default settings on the user menu, you can adjust the "sub-contrast" & "sub-brightness" for good PQ while you're in the service menus. The settings labeled "sub-contrast" provide a baseline contrast (actually white level--it would more accurately be called the "glare control") which controls the affect of a given setting in the corresponding user menu's contrast--i.e. moving sub-contrast down in the DeNIE menu causes the effect of (the default setting of) contrast=90 on the user menu to actually be lower contrast. Likewise, the DeNIE setting of "sub-brightness" provides a baseline for the user menu's brightness (actually black level) settings.
BTW, DeNIE is set up as follows on my set: TV mode and HDMI share settings; all Svideo inputs share settings; Component 1 and 2 each have their own settings; and the VGA PC input has its own settings.
(2) is much smaller then 16x16 pixels and so is, I believe, really MPEG quantization noise, an artifact of the low signal to noise ratio of MPEG encoding in dark areas, especially if the compression is quite high or the image has been encoded/decoded more then once or with a less-the-great encoder. Since I have trouble spelling "quantization" and being a photographer by training, I usually call this "dark-area-graininess". It's not really a "macroblock problem" as is often stated on this forum. Anyway, I found service menu items that fixed this:
(a) changing the "Index Delay". Mine was at 42. Using the test pattern in the index delay service menu, I found that the gray scale test pattern broke up (started changing color from gray) with settings below 28 and above 45. As advised in the service manual, I set mine to 36, the middle of the range.
BTW, the index delay is used to synchronize the electronics with the mechanics of the spinning color wheel--e.g. to make sure that the red signal happens just as the middle of the red segments of the color wheel pass in front of the light beam and not while part of the green filter is in the beam. Hence, it affects all inputs and varies slightly with each color wheel.
(b) changing the Gamma. Mine was at 4 and I wound up setting it to 1. The values run from 1--9; the service manual says it selects a lookup table, and it looks to me like they're arranged in random order because they bounce all over the place; e.g. 2 was almost identical to 4, but 3 was Much lighter and contrastier then either and much like 6--9. I found 1 & 2 were similar to 4 & 5 with 1 & 5 being darker then 2 & 4 and 1 & 2 being slightly darker and producing less dark-area-graininess then the others. All gammas above 5 where useless. Gamma is a system-wide setting.
I also turned down the Lamp Boost setting (also system-wide). Mine was at 15 and I lowered it to 0, which made the overall picture very slightly less bright. My theory is that this affects the voltage applied to the lamp, so a lower value will prolong lamp life. It might also change the color temp slightly, and I did notice (I think) a very slightly bluer color at 0 (although logic says the color temp should go warmer, not cooler). I did notice that if I turned the lamp boost up past 19, the picture turned significantly greener.
(3) can be largely eliminated by lowering the DeNIE setting called "Grn-B-coeff". This, of course, applies to all inputs in a DeNIE settings group. Mine was at 128 (as were all the "x-x-coeff" settings), reducing it to 115 for the HDMI input (which I use for my DVD player) cleaned up DVDs without noticeably affecting the TV input that shares the same DeNIE settings group. Nor does it seem to have a noticeable effect on high contrast images. I also lowered Component 2 (which I use for my SA8300HD STB) grn-B-coeff to 123.
There's also a "Grn-C-coeff" setting; I'm not sure what effect the two different Grn-x-coeff have (green-brightness and green-contrast?? one of these may have the same effect as the user mode "My color control" settings). Anyway, the Grn-B-coeff seems to work better for getting rid of green in dark areas of low contrast scenes, which is what I was after. My theory is that some combination of coeff (and possibly other DeNIE settings) can be used to get the same effect as the user mode Tint control would have if it was not disabled for some inputs.
BTW, the CCA settings are intended to correct for minor variations in the color/density of the filters from color wheel to color wheel. A replacement color wheel should come with a factory calibration label that specifies what the correct settings for that wheel. I think it works like this: the "D" settings in the right-hand column are the desired/target numbers for the desired color gambit space; the electronics manipulate the electronic signal to scale it from the actual value for that filter to match the desired value. Without an expensive colorimeter or paying for a calibration, it's pretty much impossible to calibrate CCA, so trusting the factory setup seems wise. Anyway, I've found that changing the settings quite a bit has virtually no noticeable effect on the image. CCA settings are grouped by inputs as follows: TV & HDMI; all Svideo; all Component; VGA/PC).
I found that Activator gain (mine was at 90; a system wide setting) made relatively little difference and you have to look at the test pattern Very closely to see any changes at all. It seems to affect sharpness/focus/convergence. (of course, there's no convergence per se in a DLP, since the same mirrors are used for all colors, but Activator Gain may effect the rise time of the signal that flips the mirrors, thus changing the timing of the mirrors.
