Originally Posted by Ignace
Look carefully in the greenish background and in the 2 pink blurry areas in the background, I saw more contouring using the CMS compared to THX. I must stress that we are speaking of little contouring, so you have to look carefully! But you can see that the colors are a bit more "jumping around" in smooth transitions
I checked carefully your image in Capture NX.
There is no banding and the JPEG compression is clean.
The bokey is nice but in the mentionned areas green and pink are blured and mixed. There is a lot of very fine color variations; the saturation gradient is smooth but visible.
So your test material is good.
Some days ago I reported that the saturation curve in HCHR is not linear. mainly in the 90-100% area where the slope in very abrupt.http://www.avsforum.com/avs-vb/showt...4#post16479404
As a result some pixels having very close colors in that portion of the curve are reduced to the same value and seen as larger area similar to banding.
This is similar to artefact in low quality compressed jpeg files.
In some areas switching the CMS does enlarge them.
Two different cms settings will produce different saturation curves, and therefore different positions and colors of 'banding'.
Switching from one to the other animates the pixels.
The statistical distribution of the blur is modified and is dependant of the non-linearity.
There is an other aspect of the problem above Saturation 75%: the saturation curves of the primaries are very distant from each other. The effect is the same.
my test results:
CMS off has almost no effect since it is linear
Gamut transformed by pixel shader is very good
THX is not too bad and smooth
CMS depends on the shape of the saturation curve, but is visible in any case
Originally Posted by Lawguy
Very interesting! You have clearly taken more measurements than anyone else.
I don't have reference to your actual readings right now, but overall I recall this looking very, very good. Something does appear to be going on, mainly between 97% and 100% saturations. What do you think it is? I don't have the tools to take as many readings as you do but I don't see any issue when comparing 100% and 75% saturations.
Sorry for the delay.
Do this test in HCFR: use the internal patterns, in the menu select 40 colors for the saturation test and look carefully at the screen; the two last transitions are very strong and will jump to your eyes.
I do not understand what is wrong in the CMS at that level.
In the 90% sat area the relative positions and distances of primaries saturations are of the same order as their relative brightness level.
That is a remark, not an explaination...
Regarding the saturations, the CMS is not totally fixed