The Official ChromaPure thread - Page 73

It is faster indeed.

Maybe for 'autocal' but I was meaning as a guide for manual adjustment.

At the moment, not all CP users will have the benefit of a complementary external device so 'autocalibrate' is not an option to them.

Yes I did mean these near zero dEs become academic. On the other hand. I have to say I did a litlle hand tweaking on some colors and got the luminance error much nearer to zero and that resulted in my highest color dE being 1.0 on magenta and this is by far the set has looked so far.

Keep in mind this is a new rear projector and I have done only dark room calibrations to this point. I just can't get over how good it looks. This was not a cheap display by any means. This new "Clear Screen" Mits Rear Projector just looks fabulous now. My 82" was already immersive but the color was never this saturated and the set was no where near as sharp as this 92".

Tom's software has just allowed me to get so much more out of the set than I would have. I could recalibrate every few days as the new lamp burned in. Obviously, the 82" just wasn't capable even with the DUO of approaching 0 dEs on primaries and gamma was always a twitchy affair.

The 92" display clearly calibrates very much better because it is better implemented but Tom's software just makes getting the color dead on, knock down easy. Auto does not nail grayscale like it does color, but I suspect it would not take much manual effort to get autocalibrate to dial in grayscale errors near academic as well.

I once said the autocalibrate with the DUO was like magic. I think I overstated the case at the time because I found later I ultimately did not continue to use it on the previous display.. the DUO and the Dynamic Iris were not that good a match. The Lumagen with full fields would have been a better match for the DI

But with this new set auto on the gets very close to magic. On Mits sets at least some preliminary thought needs to go into the whole thing about color clipping. The Mits' internal CMS still clips color at high brighness levels, but the DUO has its own quirks (one being oddly beneficial in this case). If you leave the Mits contrast high and adjust the contrast down slightly on the DUO with the DUO's Picture Control Menu you will drastically improved any high stimulus color clipping.

Bottom line I am very pleased with this new release. I hate the uninstall/reinstall process (yes I am lazy) and I would love to see user definable target gamma. This is great tool in the hands of a complete newbie and it make short work of calibrating for the more advanced enthusiast.

I think what I like most about auto-calibrate is that it is an amazing feature and that it keeps getting better and better. It won't make a silk purse out of sow's ear display even with a lot of hand tweaking, but if the set is capable of fully being dialed in, auto makes it a cakewalk.

Now if only it would set up the tripod, boot the notebook, and dimm the lights.....

Hi Gary,

I know you had a lot of pain with your last Mitsi (re many posts on old DVDO forum) so I am so glad you are now getting some joy.

Your comment about using Duo's contrast to stop clipping has worked perfectly for me as well.
My TV controls simply would not reduce enough to prevent it but as you say the Duo's control works very well (great clipping pattern with S & M disc).

On same subject I spoke to Ken at the now Simplay and he told me I should do any calibration with the Duo contrast at zero then adjust Duo contrast after calibration.
My mistake was not realising Duo 'Picture Controls' work on inputs only whereas CMS works on output only.

Hence I was calibrating at too low a 100 IRE 'Y' value which gave me far too low 10 and 20 IRE 'Y' values to measure effectively with the D3.

Agree of course, I didn't mean to have these for VP only but as a general option.

I don't know what they are smoking over at Simplay, but Ken's way doesn't solve my problem. I am not so focused on pretty patterns on Spears and Munsil color clipping. I just use the AVS color clipping test pattern. As long as color is not clipping below 235 I don't much care. You are basically talking almost everything above 235 is white anyway, and I am not clipping white whether I use the DUO Pic Contrast Control or not. I would much rather have the dynamic range. The 75 percent patterns get the color right at the heart of the matter and if a color runs out somewhere above 235 and white is not clipping or color shifting so be it. I mean it is nice if everything is that linear that you have more than enough max white and no white or color clipping all the way to a 109 percent.. but Mits 92" rear projectors are not infinite contrast machines.. there is a practical limit to how much light you are going to get out of a 180 watt lamp. Mits could certainly have worked harder at getting the same light and not clipping any color all the way out to the end.. it is just that in the real world it is probably mostly academic.

I think there must be sets that you getvitually perfect color dEs at 25%, 50%, 75% and 100 % with the same configuration but there are probably darn few sets that will do that and also take the near perfection on color all the way out to 109 %.

Maybe someday the video processors and software will give us 20 step color configuration for both color xy and gamma on each color and autocalibrate for it.

This really makes sense to me.

My LCD refuses to give me full blue saturation at 100% patterns never mind 109% so being realistic probably suits most of us when it comes to expecting perfection from displays.

It is so dependent on the display.
After calibrating my LG 55LH90 with 100% sat patterns from the Duo the CIE diagram shows primary and secondary spot on except green which is slightly over saturated

This should not cause concern since 100% green is quite rare in real program material.

I take slightly over vs. under any day.... Heard a lot of good things about 55lh90. How's it black level compared to the new panny ?

The 55LH90 is the only LED lit set I own so can not compare it to the Panny.
IMO the black level is excellent and screen uniformity is perfect for color and monochrome full screen patterns.
Haloing can be seen on screen credits but is so slight one has to look for it.

Viewing demos of the LH5600 and LH6500 at Fry's I found the passive 3D impressive on demo material like Imax "Beneath the Sea" but I can not persuade myself to trade for edge lit LCD with problematic local dimming if even available.

Viewing folks wearing black or near black coats on well calibrated programs like Fox News one can see the shadow detail on the folds in the sleeves.

After two years and the most recent calibration with CP Pro, D3 Pro and iScan Duo I will keep the LH90 till it fails.

spent some quality time with the new release on saturday...

as reference, i'm using a d3 with a lumagen mini 3d... measuring off the lens... running chromapure on an ancient xp laptop... pj is a mits hc9000d, screen is a da-lite hphc...

pros:

- faster. and i thought it was fast before. it's done before you know it now. looks like tom tightened up the algorithms a bit, it doesn't seem to "chase the wrong direction" like it would sometimes do in the past.

- the second greyscale run after the gamma correction works well. this was a nice little addition. me likey.

cons: (and i'm not sure it's a con, the problem could be between the keyboard and the chair)

after the autocal (and subsequent tweaking, which was minimal), i still had plenty of time on my hands, so i decided to play arounf and experiment with different color temp presets on the pj (rather than using the "user" preset)... since i wasn't using the "user" option, i needed to set 100% white in the radiance...

used the "white" in the radiance cms to do this... no problem there... however, when running the gamut autocal, the autocal changed the white point (significantly), and completely hosed the greyscale as a result...

this is repeatable... i tried it 3 times to make sure it wasn't something i was forgetting (like doing a save, or whatever), and it does the same thing every time... white gets messed with (the red component gets set back to zero, and the green component changes to a different number than i set, blue stays 0 as it was originally set)....

anyone out there want to give this a shot and see if they get the same results as i do, or possibly shed some light on what the user is doing wrong?

Don't know if its relative (I have a Duo not a Mini), but I find that if I 'autocalibrate' all three (gamma, grayscale and gamma)), I get great results (to manually tweak).

However, if I then try to 'autocalibrate' gamma only this significantly affects the 'tweaked' grayscale.
Similarly using only grayscale affects gamma.

However a color only 'auto calibrate' seems ok for me.

^^^

thanks for the input...

that's consistent with "normal results" in the past for me...

the issue i bring up is only when using the radiance to do the white point adjustment and then doing a subsequent gamut autocal...

I would guess that the Mini CMS is being reset prior to running gamut autocal - I'm not sure if the CMS reset can be done without resetting white. Perhaps there needs to be an option to adjust WB after the reset but before Autocal runs.

^^^

yea, that was what i was thinking too... but what's odd is that the red component resets to 0, but i get a (seemingly) random value for green...

what's intriguing is that i used this methodology before, as i wasn't smart enough to figure out how to adjust the white balance on my previous pj... i did not run into a similar problem then, which is what led me to believe that it's a possible bug with the new release...

just blue skying from a software developer's standpoint, it almost looks like a register has leftover data in it that wasn't cleared/set correctly... dunno, that's merely a guess...

and as noted, the problem may be beteween the keyboard and the chair, and i'm making some kind of simple error that i'm too close to see...

Not sure I understand this. Is the problem that auto-cal is not resetting all values to zero when starting or is it something else? Could you describe the problem again briefly?

hi tom, thanks for chiming in...

basically, it's this...

- start with a cms on the lumagen set to defaults...
- use the "white" color in the lumagen cms to correct 100% white balance...
- run gamut autocal...
- values that i set for the "white" color in the lumagen get changed... in my case, the red component reset to 0, and the green component to a seemingly random number in the 100's, blue remained 0...

hope that's clearer...

First, the white value will get changed. It should be reset to zero. In cases like this, run the grayscale and gamut calibration together.
Second, RGB for white CMS should all have been reset to zero. Let me look into this.

^^^

ok, cool... thanks... when i get home tonight, i'll play with it in the suggested manner...

good enough on the second part...

Apart from a potential time saving, what advantage is there to running 'autocalibrate' by selecting gamma, gamut or grayscale seperately.?

lets say you were happy with gamma, happy with gamut but not happy with grayscale. You could hand tweak grayscale to get it perfect then autocalibrate the gamut again. One can think of may scenarios. With my previous DlP I would let it autocalibrate all three then hand run each and it worked well because while grayscale and gamma did not autocalibrate as one would like because of the windowed patterns on the DUO... my set had a non-defeatable dynamic iris.. it would get these numbers in the ballpark and do pretty well with gamut. So doing the whole thing starting from the autocalibrate results was easier especially on gamut.

One could for example could just rerurn grayscale and gamut again one gamma is good. It does not alway produce the same result.

For example I have used autocalibrate with the internal defauts. Then hand tweaked the internal settings a bit and ran gamut auto again, which produced better results.

Dear Tom,

I get a bit confused on the process of calibrating my new JVC RS-55 projector using my Lumagen w/ Chromapure autocal and the current X-rite pro 3 meter. Some AVS members suggest...

a. I go into the JVC service menu and manually set the white balance using Chromapure as a guide.

b. Manually set the gamut using JVC's menu's and the Chromapure as a guide.

c. Then run Chromapure autocal for the CMS and/or possibly all 3 calibrations again (gray scale, gamma, & CMS).

Is this the right process to get the best results for the JVC? I'm very new at doing this, so any specific guidance you could offer would be appreciated. Thanks so much.

Only that a display may not need calibrating in all three areas.

To start, you shouldn't have to manually calibrate anything. Just select reasonably accurate presets as your starting point before running auto-cal.

After auto-cal is finished you may want to make some small manual adjustments.

Hi tom

Any progress on specific support for the sharp elite panel besides just using the plasma profile?

I enabled ISF mode on the set and did an auto cal and manual dust up using the elite's controls - while I got excellent numerical results for greyscale and gamma, the color was notably off, and acm revealed some quite horrid tracking whether I started with 100 or 75 stim - red and magenta hues were particularly off among others, and I could adjust crudely those hues and lightness controls by eye better than I was getting trusting the numbers.

I don't know if the elite has major color decoding issues or what, but something was wacky. My next thought is to use the ISF mode but make no adjustments there and try the mini 3d. Using the mini 3d and I the standard user modes resulted in poor calibration, but I'm wondering if the ISF mode enables different processing (certainly the slider control values are different in ISF mode than the user accessible modes for hue, sat, and value amonth other settings)

So while it's better than using standard modes to calibrate, the ISF mode with Chromapure and the DIII is still showing anomalies - just don't know where the fault lies at the moment...

One of the main reasons it was suggested to set 100% WB in the JVC service menu is because last year adjusting the 7000 degree preset for 100% WB at 6500K yielded more light output than doing it any other way. This year I get about the same light output with either the 6500K or 7000K preset adjusted to 6500K (in the JVC SM).

That said, the example given above by ccotenj is another good reason for adjusting the 100% WB in the projector and it is dead simple to do. I also believe this is the method suggested by Lumagen.

BTW, I found little difference in adjusting gamut in the JVC or the Mini. But in both cases I could not reduce the blue dE at 75% saturation using either system. The good news though is that the Standard color profile gamut dE is good enough as-is so I ended up without a gamut correction and its corresponding loss of lumens.

BTW, Last year saturation tracking became an issue when the JVC CMS was adjusted to correct more than a tiny error. That doesn't seem to be as bad this year but if you end up correcting gamut (with either system) I suggest you use the Advanced CMS module in Chromapure to verify saturation tracking isn't whacky.

The meter mode has nothing to do with this. This is something weird about this display, which I haven't spent enough time with to troubleshoot. It sounds like the set's CMS results in very non-linear results throughout the color space.

What do you mean that "standard user modes resulted in poor calibration"? Could you be as specific here as you were in describing calibration from the user menu? FWIW, I saw the same thing with a user calibration.

Did you try the Sharp's controls in the ISF menu or just the Mini?

Selecting all three seems to give me the best compromise between them.
If I then try to say adjust just gamma because my grayscale is close this seens to give me worse grayscale than before.

In other words I find that there is less manual tweaking work needed if I do all three together.

Please don't misunderstand, 'autocalibrate' is giving me a great first attempt which I am more than happy to manually tweak.
Its only if I try to 'tweak' the one that is least accurate by singly 'autocalibrating' that things go worse not better.

Would you expect this?

Hi Tom

Sorry I wasn't clear....

Initially, I had tried calibrating the set using several of the user modes (THX, Movie, etc.) using the standard CMS controls (not the ISF mode). This was after you have visited, and I did not get good results.

I also tried to use a Mini to calibrate these modes, but all of the initial RGB readings showed severely undersaturated unless I turned on Local Dimming, which brought the primary sats up some, but still way to low, or one of the Advanced local dimming modes, which shot the primaries way up. But all I've read is that those features should be turned off when calibrating, and in fact, when you enable the ISF mode, the set turns them off automatically.

The ISF mode is is different it seems, as I do not see the undersaturated primaries, and I thought all was going well (using the Greyscale and CMS in the Elite's ISF mode; gamma/greyscale tweaking was done in the Mini simultaneously, as there is no way to set gamma properly in the Elite alone)

They despite EXCELLENT numbers and charts, the colors looked off when viewing - watching the Rangers/Canandiens game showed that red was really magenta, and in fact, the ACM chart (which I did not save foolishly, but will run again and do so), showed a severe shift toward magenta, and the points themselves where way off. Same for all colors really; some 25/50 or 50/75 points were nearly bunched up on one another.

Skin tones were too saturated, yellows too pronounced... really odd.

I actually went back in and pulled Red hue and Magenta hue back more toward center from their more radical positions (to get Red "right" when calibrating, Hue was moved quite a bit to Magenta). This manual tweak helped somewhat, but the tracking is not right at all.

My next thought is to enable ISF mode, but leave the Elite Greyscale and CMS controlled untouched (expect White Point), and then run a Lumagen only cal and see what happens.

I do have the detailed calibration report, but if you were to read it, you would likely say fantastic!

I am on my third Elite because of DSE issues and my concerns over calibration - frankly, while they are about to send a fourth! set, I am really thinking there is something fundamentally wrong with the way this set decodes color. If you are still planning on supporting auto cal to the Elite directly, I'm curious what you will find

If you need to stop by again and spend more time to figure this panel out, you are welcome to do so...

Thanks

That certainly was the case last year when it was demonstated at the Value Electronics shootout. Even in the low quality youtube video it was plain to see. Blue was purple and orange was yellow.

A patch was supposedly being worked on but I've not heard whether it was released or if the problem has been addressed.