Originally Posted by ADU
And also 2170P-4/SCOL & SHUE (which is what I used to tweak DVI on my TV instead of using the 2170P-3 controls, to save some time).
Indeed there are several places where Color, Hue, and other menu-slider offsets can be tweaked. It's important to be aware of them all:
(1) 2170P-4 #4-6/SPIO-SHUO appear to be global offsets for Picture, Color, and Hue, with values of 0-15. The service manual specifies defaults of 4-9-7, but my set came with 7-7-7 (the midpoints). If my Hue or Color sliders were off on all inputs, meaning for me that 31 is not the calibrated setting for *any* input, I would start here. I have left them alone. They show only one column in the service-data chart.
(2) 2170P-4 #1-3, SPIC-SHUE appear to be the same as (1) but are specific to 5 broad categories: DRC-processed signals, component inputs (V5/6), HDMI input, MS/ATSC (memory stick and digital tuner), or something called "passthrough." So these codes have 5 columns each in the data chart. They are pretty wide categories.
(3) 2170P-3 #13-15, UBOF-UHOF, for Brightness (black level), Color, and Hue. These are specific to each combination of video input *and* input-scan rate, 17 columns in the chart. Here is where you can balance these three parameters among your signal sources. I balance Brightness here, but I keep the Color and Hue values identical for DVD, MS, and all broadcast sources -- so a calibration from a DVD or MS source is valid for broadcast inputs. My VCRs are then tweaked to make them look like the calibrated sources.
None of these are directly part of color-matrix (decoding) calibration. The two parameters that affect its accuracy, however, must be set properly first: Color (amount) and Hue (tint, color phase). Brightness and Picture have no direct effect on color-decoding calibration.
It should be obvious that there are more than one way to "tune" the Color, Hue, Brightness, and even Picture parameters among the large number of unput/video combinations. A good strategy would be to accomplish calibration for the broadest categories of offsets first (#1, above), then do the finicky input-dependent diddling last (#3).