Official JVC RS20 / HD750 Calibration and CMS thread (NEW FIRMWARE V1.1) - Page 18

I hope that is how it works for me. My i1pro is due on Monday. I am just getting started down this road.

I calibrated another RS20 today and I have some new data. From the two units I have seen, it seems that the gamuts are very consistent, but the 6500K Color Temp that THX uses varies quite a bit.

First, the THX mode

THX	x	y	Y	CIE94	L	S	H
W	0.315	0.337	1.000	4.6
R	0.642	0.330	0.215	0.4	0.4%	0.6%	0.0%
G	0.313	0.589	0.715	2.0	0.0%	-4.6%	-4.4%
B	0.147	0.064	0.071	2.0	-0.9%	-2.6%	-2.3%
Y	0.423	0.504	0.899	1.4	-1.2%	0.2%	-2.4%
C	0.234	0.340	0.781	2.8	-0.3%	-6.9%	-10.5%
M	0.331	0.163	0.280	1.6	-0.7%	-4.0%	5.6%

Next, the User 1 mode pre-calibration.

User1	x	y	Y	CIE94	L	S	H
W	0.314	0.338	1.000	4.6
R	0.672	0.326	0.220	5.9	1.5%	12.6%	-1.0%
G	0.286	0.699	0.707	9.5	-0.4%	22.6%	2.3%
B	0.140	0.052	0.064	2.2	-5.8%	5.9%	-1.9%
Y	0.442	0.548	0.928	9.9	0.0%	17.2%	1.4%
C	0.206	0.343	0.773	6.0	-0.7%	24.0%	-10.0%
M	0.330	0.149	0.286	1.8	0.1%	7.1%	4.5%

Next, the THX mode at 50% saturation.

50%	R	G	B
x	0.479	0.317	0.244
y	0.334	0.458	0.213
Y	0.212	0.715	0.072
CIE94	1.0	2.4	2.2
L	0.0%	0.0%	0.0%
S	-0.4%	-6.3%	-16.8%
H	2.1%	-7.5%	1.4%

As you can see, red is essentially perfect. Green is off a little, but that is as much due to a correctable hue problem. Blue has the biggest problem, but the dE is relatively small. It seems that this PJ is fairly linear throughout the color space.

Finally, offsets between my Chroma 5 colorimeter and Orb spectroradiometer.

	x	y
W	0.004	0.006
R	0.000	0.001
G	-0.002	0.004
B	0.004	-0.001
Y	-0.002	0.003
C	0.003	0.008
M	0.006	0.001

As you can see, the offsets are very small, with white and cyan having the biggest problem. This is consistent with what I saw on the other RS20 as well. This PJ is a fairly easy unit to calibrate with a good colorimeter, which is not surprising as it uses a standard UHP bulb.

As I reported earlier, a post-calibration User 1 measures almost perfectly in all respects.

First, sorry if I missed this and it has already been discussed.

I sold my RS20, so can't do this myself, but I did calibrate it before I sold it and I had to lower the blue gain quite a bit. I don't remember how much extra blue I would have had with all the gains at 0 though. Since some other numbers I've seen indicate that other people have had to lower the gain for blue a fair amount, I'm wondering if anybody is considering doing this with a color filter instead. A filter that filters some green and more blue would seem ideal, but I'm not sure how much would be called for. I have an FL-Day filter, but no RS20 to try it on.

Anyway, one reason I think this could be extra helpful is what dcouzin posted about here:

http://www.avsforum.com/avs-vb/showthread.php?t=1145337

Basically, our vision in the darkest stuff tends to be skewed much more toward seeing blue and not red. So, while getting rid of some blue in the absolute black while lowering the blue in reference white optically could help the on/off contrast ratio measurement some, it could actually help the perception even more than the measurement using the photopic curve (basically what applies to brighter scenes) would indicate.

BTW: I used Calman with my EyeOne and one of the DPG-1000 Digital Pattern Generators from them. It was so nice to be able to kick off a set of measurements and just let the tools auto change the test patterns and do the measurements instead of having to manually change the test pattern and kick off the reading every time. I was able to do some other things around the house in the breaks while it was measuring a sequence for me.

--Darin

While I was measuring the primaries on my RS20 I was having some trouble measuring the blue primary. I ended up looking at the spectral curve in Calman and one thing it showed me is that the amount of red the tool was indicating for the blue primary (I think 100% level, but not positive) where I was aiming for REC.709 blue was about the same amount of red that it would indicate in a 30%stim gray image. The amount of signal the meter was reading in wavelengths in the green range was lower than it indicated for green in a 30% stim gray image. I should have saved images of those 2 spectral graphs so I could post them, but I didn't.

But for anybody using a spectroradiometer I think it could be interesting to look at what the spectral graph indicates is being read for levels of green and red while trying to adjust the blue primary, compared to what it reads at some different gray levels.

In my case I had put the projector about halfway to the screen with a smaller than normal image to try to get brighter images to hopefully help the accuracy of the measurements. I think my white level for this one was a little under 30 ft-lamberts.

--Darin

I had both the Planar 8150 and RS20 before this and still have the 8150.

--Darin

Thank you! I thought I was going crazy. I have been a long time member that does not post much (mostly lurk and read) but i am having the same problem. I cannot get a dE on blue under 6, but I can dial everything else in almost perfectly. I am seeing some of the others calibration results with a blue dE of under 1 and I just can't figure it out you are doing it.

I also wanted to add, I just purchased an i1 Pro with Calman software. I saw some people asking about this earlier. I will say, that as a beginner I found the Calman software difficult to use. I finally switched to HCFR, which it seems most use in this thread anyway, and found it much easier to use in conjunction with the greyscale and color calibration for dummies tutorial. It has been a great learning experience and I am learning so much more about my RS20- I highly recommend it.

Since I'm out of lurking mode, thanks to all of you that have posted, especially Manni and LDVD. I have gained a ton of knowledge from this thread and have enjoyed using some of your settings before I attempted my own calibration.

Hi Manni,

hasn't it no problem on this point (under 100 %) ?




I'm sorry I can't open your (other) files, how is the "Y" (lightness) ?

A professional certified ISF said that it was (In his opinion), the more important issue with the first firmware (the variation on "Y" when he used the C.M.S.) ?

Hi Manni,

I don't know why I can't open your files (I use the HCFR Software since many years ), but it's not important.

If I use your "pictures" (with the measure at 100 % and at 75 %) I obtain this :

It was not very important in absolute value, but in relative value ?

 

0k . file

I should add one thing. Derek (of SpectraCal) recommended not going over a specific amount for white with my EyeOne Beamer. I think it was around 100 ft-lamberts. I'm not sure how much this applies to the blue primary (like whether putting the projector even closer to the screen for that one, or the meter closer to the projector if measuring that way, would help).

--Darin

Hi,

I use this (you calibration picture at 75 %, if I don't make a mistake, it's what I understand ) :

Hi,

I'm confused, but the Y for the "white", it is not the Brightness ?

and it's not linked with the "Y" (and x and y) on primary and secondary "Y" ?

So, when, the "Y" white is at 16,29 at 75 % (against 28,8 at 100 %) the other color mustn't went on the relatives references ?

Sorry, maybe it's a mistake

Measure off the screen or from the projector but at the end of the day if the measurements off the screen are different, you have a problem.

I have no idea how much of an impact the screen has. I think there is some and when you add it together with unit to unit variation, it really makes a visible, if as Millerwill describes, small difference.

I was wondering if you guys also do some visual checks after calibration. I mean: do you perform a color decoder check afterwards?

My CMS results (calibrated with I1 Lt, later this week my I1 pro will arrive) are very, very different from most settings I see on this thread, especially hue on red, green and blue, but they measure spot on (below delta error 2, below xy error 0.002 and below Y error 2%). Both on 75% and 100%.
Also the visual color decoder check (the 3 flashing RGB bars, the flashing RGB color brightness check and the RGB clipping check on the AVS HD disk) show no visible errors at all.
When I enter the values I see in this thread, both my measured and my visible results are completely off.

These are my CMS settings:
R 36, -23, 16
Y 8, -46, 36
G 24, -37, 28
C 4, -43, 34
B 36, -17, 10
M 8, -37, 25

Any thoughts about these huge differences?

*Edit may 16: Don't try the numbers above, they are not good. I've got a i1 Pro now and it seems my i1 D2 was a long way off.

When one starts to discuss 75% saturation, this has to mean "Chroma level". For example if you display a 75% SMPTE bar pattern, the signal (stimulus) is 75%. If you measure the White "Y", there will be a component of RGB that totals to the White "Y". If the blue component Y = 1.0, then, at the proper chroma level, the "Y" for the Blue bar should be 1.0. If the Blue "Y" = 0.75, the "color" control would be set to low causing it to under saturate the colors. All of the other colors in the pattern would be too low in "Y" as well. If they were not all 75% of expected "Y" values for their respective color, color decoding would be off (assuming the 100% x,y,Y values for White, Primaries and Secondaries are correct)

As far as I'm aware of, there are no sources for 75% Stim, 75% saturation patterns. The only patterns I'm aware of use 100% stim, but vary the saturation (chroma level). These are found on the AVSHD709 downloadable disc. These are the patterns I use for my calibrations. They are in the "saturation windows" menu option. For each of the primaries and secondaries, there are 0, 25%, 50%, 75% and 100% saturation (chroma level), but all are referenced to 100% Y, not 75% Y and assume a display gamma of 2.22.

I love the fact that these patterns exist because they have proven VERY useful for me.

Dan

It does look to me that that lower saturations are a little off. 75% saturation of red, green and yellow for example seem undersaturated versus the target points. This could be due to your gamma though. What gamma are you measuring?

Who's chart is this by the way? I'm a little confused. I'm not sure if Thebes was showing someone elses calibration results or his own.

Dan

What we are referring to as 75% saturation is much simpler than this. It is just distance from the white point. For example if red = x0.640, y0.330 and white = x0.3127, y0.329, then the distance between them is x0.3273, y0.001.

So 50% red saturation is 0.3127 + (.5 * 0.3273) for x and 0.329 +(.5 *0.001) for y or x0.476, y0.330. It is just half way from red to the white point in the Rec. 709 gamut. The Y value doesn't matter for the purpose of this measurement.