AVS › AVS Forum › Display Devices › Digital Hi-End Projectors - $3,000+ USD MSRP › Official JVC RS20 / HD750 Calibration and CMS thread
New Posts  All Forums:Forum Nav:

Official JVC RS20 / HD750 Calibration and CMS thread - Page 8

post #211 of 1782
12/23/08 at 11:05am
I'll repost, as no one saw my post.

I haven't worked with the RS20 color management yet, but why not try for perfect Y first on all colors, then work on bringing in the points to there designated places on the CIE chart?

You will also want 100% color windows for more accurate results.

Tom I have a question for you...

Example: a red 100% window on the AVS HD Calibration Disk in the RGB domain is

Green = 000
Blue = 000
Red = 235

Correct?

and a red 109% window (if one where to be avail) in the RGB domain would be

Green = 000
Blue = 000
Red = 254


is this correct?
Reply
Reply

AVS Top Picks

post #212 of 1782
12/23/08 at 11:25am
Quote:
Originally Posted by TomHuffman View Post

No, the 4.2% figure refers to the % deviation from correct lightness (L, lightness, S, saturation, H, hue).

Just look at the LSH numbers. They show exactly where the errors lie. I'm using CIE94 for dE analysis, rather than CIELUV, which although fine for grayscale, I find not as accurate for pri/sec color grading.

Thanks for your patience with me Tom. I got you percentages wrong (I thought they showed deviation from x,y,Y instead of LSH).

I'll give it another try, especially on blue which should get better now that I understand where it's wrong... Green and Yellow are almost spot on as far as x,y are concerned if I look at (and understand) your percentages. It's Y which is wrong, and I've reach the max of brightness, so I won't be able to get these any better unless I change strategy (I will explore Jason's route to see if I can get a better result with a color setting between -15 and -22).

Cyan is off everywhere, but I can't get it any closer even on the x,y space.

Back to the drawing board...
Reply
Reply
post #213 of 1782
12/23/08 at 11:29am
Quote:
Originally Posted by SOWK View Post

I'll repost, as no one saw my post.

I haven't worked with the RS20 color management yet, but why not try for perfect Y first on all colors, then work on bringing in the points to there designated places on the CIE chart?

It's because you haven't worked with it yet... Unfortunately if you bring color down too much, you don't have enough brightness to compensate, and parameters which should be independant are not. For example, brightness affects saturation, which makes things very difficult... If you get Y perfect for a color, and then use saturation to get it right on x,y, brightness changes as well (although less thant the other way round)... Try it, it's fun during long winter nights.
Reply
Reply
post #214 of 1782
12/23/08 at 12:10pm
Quote:
Originally Posted by SOWK View Post

Example: a red 100% window on the AVS HD Calibration Disk in the RGB domain is

Green = 000
Blue = 000
Red = 235

Correct?

and a red 109% window (if one where to be avail) in the RGB domain would be

Green = 000
Blue = 000
Red = 254


is this correct?

Seems right.
Reply
Reply
post #215 of 1782
12/24/08 at 5:35am
It could be very interisting to lauch test of all saturations of primaries and secondaries with color HCFR and publish the gamut results.
Personnaly, i had perfect positions of primary and secondary colors but saturations was unsaturated. For example 100% green was good but 75% green was at 60% instead of 75%.
Reply
Reply
post #216 of 1782
12/24/08 at 7:02am
I've never played w/ gamma much before - and I've been going back and forth between 2.3 and 2.4 (Jason dialed in 2.4 for my unit). Switching back and forth - there seems (to my eyes) to be just the very slightest difference between them (2.4 gives an ever so slightly darker image during low APL scenes). Maybe I'll test the Star Wars trash compacter scene when I get home as this is supposedly one where you can tell the difference.
Reply
Reply
post #217 of 1782
12/24/08 at 7:10am
Quote:
Originally Posted by John Ballentine View Post

I've never played w/ gamma much before - and I've been going back and forth between 2.3 and 2.4 (Jason dialed in 2.4 for my unit). Switching back and forth - there seems (to my eyes) to be just the very slightest difference between them (2.4 gives an ever so slightly darker image during low APL scenes). Maybe I'll test the Star Wars trash compacter scene when I get home as this is supposedly one where you can tell the difference.

Gamma is more of a personal preference based on the source...free free to change that as needed.
Reply
Reply
post #218 of 1782
12/24/08 at 9:30am
I just spent a few hours calibrating and learning the controls.

Thanks to Manni and Jeff for saving me a lot of work.

Greyscale and gamma calibration took about 15 minutes, after which I had a very flat greyscale and gamma dialed in nicely at 2.4. dEs of 1.X all around. Very nice.

I managed to get pretty good results with the gamut controls. There is definitely a limitation in the controls and I really suggest that we petition JVC for an upgrade here.

Still, as my results suggest, things turned out pretty good all around as compared to REC 709 with dE measured with the Accupel calculator.

I set the master color control at -7.

Here are my CMS settings:

R: H,5; S, -24; B, 21
Y: H, 2; S, -5; B, 30
G: H, -8; S, -5; B 30
C: H, 0; S, -18; B 30
B: H, 20; S, 11; B. -2
M: H, 0; S, -15; B, 20

These yielded the following measurements:

R: x, .641, y .334, Y 7.053, dE 2.7
G: x, .302, y .604, Y 22.789, dE 2.3
B: x, .155, y .073, Y 2.760, dE 3.7
Y: x, .423, y .509, Y 29.558, dE 2.0
C: x, .217, y .334, Y 22.779, dE 6.5
M: x, .322, y .157, Y 9.594, dE 2.5
W: x, .313, y .331, Y 33.282

There is some room to improve these a little, but not too much.

Cyan is the sore thumb sticking out. There is not enough headroom in the brightness control to raise it to where it should be. I quickly gave up. I found blue to be very stubborn. Most of the colors took just a few minutes to get where they needed to be. Blue can be improved but for some reason the controls do not respond as they should with this color, even though there is plenty of headroom left on them. It could be me but I spent as much time on blue as all the other colors combined. I don't know why.

I am pretty satisfied with these results numerically. I'll report back about how they look subjectively.
Reply
Reply
post #219 of 1782
12/24/08 at 9:37am
Those numbers look great lawguy. Good job.
Reply
Reply
post #220 of 1782
12/24/08 at 9:41am
Quote:
Originally Posted by mrlittlejeans View Post

Those numbers look great lawguy. Good job.


Thanks. They are tantalizingly close to where I would like them but I really have no complaints.

I really just wish that we had all the headroom built in to begin with. This color control work around seems to work all right but Jason's experience suggests that it won't work for everyone.
Reply
Reply
post #221 of 1782
12/24/08 at 9:45am
Quote:
Originally Posted by Lawguy View Post

There is definitely a limitation in the controls and I really suggest that we petition JVC for an upgrade here.

I seem to remember exactly that observation and recommendation being made somewhere else last week

Im sure a lawyer such as yourself would have no trouble organizing a posse and composing a suitably acerbic covering letter.

All the best for Christmas/Festivus
Reply
Reply
post #222 of 1782
12/24/08 at 10:08am
Quote:
Originally Posted by ChristopheCherel View Post

It could be very interisting to lauch test of all saturations of primaries and secondaries with color HCFR and publish the gamut results.
Personnaly, i had perfect positions of primary and secondary colors but saturations was unsaturated. For example 100% green was good but 75% green was at 60% instead of 75%.

Interesting. I had the chance to look at the HD750 at an exibition in Hannover/Germany a few weeks ago. Ekki showed me the newest HD750 pre-prod model.

The THX mode looked ok-ish with somewhat of a tinge to it. From comments here, it seems like this remains an issue with production units.

What i found quite curious, though, was Ekki's own CMS calibration attempt. This was around the time where he wrote that the CMS is 'kinda tricky' to use... most people in the demonstration where like "oh, ah"

I did not like the picture at all. Primaries and secondaries might have been dialed in properly, but that is just that. I don't know how big any color luminance errors where at the time, but i told him:

"This looks really bad. Its like there is a gamma-esque problem to the saturation of the picture. Your mids are really grayish and lack color. What someone should do is a proper saturation-scale tracking kinda thing, similar to gray-scale tracking."

He was someone puzzled by my critique at first, but admitted that this might be a good idea... Just a lot of work

Anyways. Interesting thread, guys. Keep it going!

Best regards
Bjoern

P.S. Back to my cave where i am hiding for some years
Reply
Reply
post #223 of 1782
12/24/08 at 10:33am
Quote:
Originally Posted by coldmachine View Post

All the best for Christmas/Festivus

To you also, assuming, as always that work does not interfere. I am looking forward to a week or more of sitting on my butt.

You called the CMS correctly. The controls generally do what they should and are easy to use. They just need a little more range added in to control my OCD.

I also would like to try a different approach: raising the color control and seeing if the CMS controls work better from that initial position.

Have you spent more time with the RS20?

Off to church!
Reply
Reply
post #224 of 1782
12/24/08 at 10:34am
Quote:
Originally Posted by Lawguy View Post

I just spent a few hours calibrating and learning the controls.

Thanks to Manni and Jeff for saving me a lot of work.

Greyscale and gamma calibration took about 15 minutes, after which I had a very flat greyscale and gamma dialed in nicely at 2.4. dEs of 1.X all around. Very nice.

I managed to get pretty good results with the gamut controls. There is definitely a limitation in the controls and I really suggest that we petition JVC for an upgrade here.

Still, as my results suggest, things turned out pretty good all around as compared to REC 709 with dE measured with the Accupel calculator.

I set the master color control at -7.

Here are my CMS settings:

R: H,5; S, -24; B, 21
Y: H, 2; S, -5; B, 30
G: H, -8; S, -5; B 30
C: H, 0; S, -18; B 30
B: H, 20; S, 11; B. -2
M: H, 0; S, -15; B, 20

These yielded the following measurements:

R: x, .641, y .334, Y 7.053, dE 2.7
G: x, .302, y .604, Y 22.789, dE 2.3
B: x, .155, y .073, Y 2.760, dE 3.7
Y: x, .423, y .509, Y 29.558, dE 2.0
C: x, .217, y .334, Y 22.779, dE 6.5
M: x, .322, y .157, Y 9.594, dE 2.5
W: x, .313, y .331, Y 33.282

There is some room to improve these a little, but not too much.

Cyan is the sore thumb sticking out. There is not enough headroom in the brightness control to raise it to where it should be. I quickly gave up. I found blue to be very stubborn. Most of the colors took just a few minutes to get where they needed to be. Blue can be improved but for some reason the controls do not respond as they should with this color, even though there is plenty of headroom left on them. It could be me but I spent as much time on blue as all the other colors combined. I don't know why.

I am pretty satisfied with these results numerically. I'll report back about how they look subjectively.

Well done Lawguy, these numbers look very good (although let's wait for Tom's verdict, he may disagree!)

Could you please tell us what your other settings are (iris, contrast, brightness, rgb low/high end) just to get a sense of where you stand. I've noticed that the contrast setting had a huge effect on the gamut (change it even slightly and measure again, you'll see it's completely off). If you stayed with a 0 setting for contrast, I may give it a try as your numbers are much better than mine.

It's not just you regarding blue, I struggled with exactly the same problem, trying to control it, and haven't managed to get it spot on on my last attempt.

I anyone could answer my question above regarding calibrating the component input of my 750 for DVD playing from my BD player, it would be much appreciated.

Happy festivities to everyone!
Reply
Reply
post #225 of 1782
12/24/08 at 10:37am
Quote:
Originally Posted by Manni01 View Post

Well done Lawguy, these numbers look very good (although let's wait for Tom's verdict, he may disagree!)

Could you please tell us what your other settings are (iris, contrast, brightness, rgb low/high end) just to get a sense of where you stand. I've noticed that the contrast setting had a huge effect on the gamut (change it even slightly and measure again, you'll see it's completely off). If you stayed with a 0 setting for contrast, I may give it a try as your numbers are much better than mine.

It's not just you regarding blue, I struggled with exactly the same problem, trying to control it, and haven't managed to get it spot on on my last attempt.

I anyone could answer my question above regarding calibrating the component input of my 750 for DVD playing from my BD player, it would be much appreciated.

Happy festivities to everyone!

Not time for a full reply but the iris is completely closed. Contrast and brightness are at or near zero for both.
Reply
Reply
post #226 of 1782
12/24/08 at 10:39am
I found an issue with my Lumagen which required a factory reset so I to start everything from scratch. I decided to do something different which was to dial in the saturation levels visually to taste and then calibrate the luminance levels for all the colors (75%) as well as some minor hue tweaks. It's something of a half way between the native gamut of the projector and 709. I think you can calibrate the life out of a projector.
Reply
Reply
post #227 of 1782
12/24/08 at 11:00am
My calibration at 25 hours.
values are on the comments inside the file.

 

calib. 25h..zip 5.3212890625k . file
Reply
Reply
post #228 of 1782
12/24/08 at 11:41am
Thread Starter 
Quote:
Originally Posted by Bjoern Roy View Post

"This looks really bad. Its like there is a gamma-esque problem to the saturation of the picture. Your mids are really grayish and lack color. What someone should do is a proper saturation-scale tracking kinda thing, similar to gray-scale tracking."

This is a really good point. I don't think anyone has yet measured how well colors track in LSH within the defined colorspace. It would be interesting to verify the tracking inside the gamut once the primaries have been corrected.

This touches on another issue - one of the problems with uncorrected, oversaturated primaries is that it introduces color errors within the Rec. 709 gamut, not just at the boundaries where it's most oversaturated. Using a non-linear stretch would help to minimize the error within the gamut with oversaturated primaries. So measuring the LSH tracking would help to see if such a technique is being employed and if so how well it works.

There are lots of measurements needed to do this properly though and so it would be better done using an automated setup with a computer controlled color generator and sensor.
Reply
Reply
post #229 of 1782
12/24/08 at 3:17pm
Quote:
Originally Posted by Lawguy View Post

I set the master color control at -7.

Here are my CMS settings:

R: H,5; S, -24; B, 21
Y: H, 2; S, -5; B, 30
G: H, -8; S, -5; B 30
C: H, 0; S, -18; B 30
B: H, 20; S, 11; B. -2
M: H, 0; S, -15; B, 20

These yielded the following measurements:

R: x, .641, y .334, Y 7.053, dE 2.7
G: x, .302, y .604, Y 22.789, dE 2.3
B: x, .155, y .073, Y 2.760, dE 3.7
Y: x, .423, y .509, Y 29.558, dE 2.0
C: x, .217, y .334, Y 22.779, dE 6.5
M: x, .322, y .157, Y 9.594, dE 2.5
W: x, .313, y .331, Y 33.282

There is some room to improve these a little, but not too much.

Cyan is the sore thumb sticking out. There is not enough headroom in the brightness control to raise it to where it should be. I quickly gave up. I found blue to be very stubborn. Most of the colors took just a few minutes to get where they needed to be. Blue can be improved but for some reason the controls do not respond as they should with this color, even though there is plenty of headroom left on them. It could be me but I spent as much time on blue as all the other colors combined. I don't know why.

As you say, blue and especially cyan are the problem children.

CIE94--L-----S-----H
R 1.3 -0.1% -0.9% 1.0%
G 1.6 -1.7% 0.6% -0.7%
B 3.5 7.1% -8.4% -1.2%
Y 1.7 -1.7% 1.9% -0.8%
C 5.2 -5.4% 9.6% -4.4%
M 0.5 0.5% -2.0% 0.7%
Reply
Reply
post #230 of 1782
12/24/08 at 3:23pm
Quote:
Originally Posted by Lawguy View Post

Not time for a full reply but the iris is completely closed. Contrast and brightness are at or near zero for both.

Lawguy:
I'd also appreciate any other settings you are willing to supply. I plan on trying out several and seeing which works best on my particular one.

Please include lamp setting, and color temp and any offsets.
I'd also like to know if you have a hi power screen since I have one.

Thanks in advance and enjoy the new toy!
Reply
Reply
post #231 of 1782
12/24/08 at 3:31pm
Quote:
Originally Posted by Bjoern Roy View Post

What i found quite curious, though, was Ekki's own CMS calibration attempt. This was around the time where he wrote that the CMS is 'kinda tricky' to use... most people in the demonstration where like "oh, ah"

I did not like the picture at all. Primaries and secondaries might have been dialed in properly, but that is just that. I don't know how big any color luminance errors where at the time, but i told him:

"This looks really bad. Its like there is a gamma-esque problem to the saturation of the picture. Your mids are really grayish and lack color. What someone should do is a proper saturation-scale tracking kinda thing, similar to gray-scale tracking."

Given what we've seen here, I suspect they were quite large. This suggests what I suspected: that he simply attempted to get a good-looking CIE chart and then left it at that.

BTW, could someone try the following: measure the xy points (you can ignore luminance for this test) for RGBCYM using a 100% test pattern. Then repeat using a 75%, then 50%, and then 25% patterns. Then please report the 48 data points.

Something I found with the Radiance is that it corrects the primaries at different levels of stimulus quite well, but that it had more trouble with the secondaries. I'd be curious to see how the post-calibration RS20 tracks. For background, see

http://www.avsforum.com/avs-vb/showp...postcount=3101

and subsequent posts.
Reply
Reply
post #232 of 1782
12/24/08 at 4:27pm
I think some of you Master Calibrators and or Projector Companies should develop a software interface so projector owners could plug in a PC to the projector/hdmi/RS232 along with a colorimeter, in their own viewing environment and have it auto calibrated on line.

Sure, not as good as having a pro visit ones home, but a darn good income stream for the providers and a good service for the projector owners.

Priced reasonably could be a big hit!
Reply
Reply
post #233 of 1782
12/24/08 at 4:34pm
Thread Starter 
Quote:
Originally Posted by Highjinx View Post

I think some of you Master Calibrators and or Projector Companies should develop a software interface so projector owners could plug in a PC to the projector/hdmi/RS232 along with a colorimeter, in their own viewing environment and have it auto calibrated on line.

Sure, not as good as having a pro visit ones home, but a darn good income stream for the providers and a good service for the projector owners.

Priced reasonably could be a big hit!

Dilard by Milori used to do something like this for the old JVC G series. Most calibration is automated nowadays just not ISF style calibration. I had the pleasure of visiting William Phelps lab in the bay area and seeing how he calibrates and pretty much everything is automated using his own tools and equipment. Factory JVC calibration is also automated.
Reply
Reply
post #234 of 1782
12/24/08 at 4:39pm
Hi, I am having Michael TLV of Lion A/V Consultants Network come over on boxing day to calibrate and give me some instruction on setting up the new HD750. He said we could be at it for 4 hrs or so. I will let you know the results in a few days.
Reply
Reply
post #235 of 1782
12/24/08 at 5:17pm
Quote:
Originally Posted by Mark Petersen View Post

Dilard by Milori used to do something like this for the old JVC G series. Most calibration is automated nowadays just not ISF style calibration. I had the pleasure of visiting William Phelps lab in the bay area and seeing how he calibrates and pretty much everything is automated using his own tools and equipment. Factory JVC calibration is also automated.

Would be terrific to see such a service on line. Pay per use or annual subscription.

Select the colorimeter one has, the sever based software does the rest......bring it on!
Reply
Reply
post #236 of 1782
12/24/08 at 5:27pm
There is a projector manufacturer working on an auto calibration kit. Unfortunately I can't say much more about it.

P.S. Don't PM me, I still won't tell you.
Reply
Reply
post #237 of 1782
12/24/08 at 7:38pm
Quote:
Originally Posted by Free View Post

There is a projector manufacturer working on an auto calibration kit. Unfortunately I can't say much more about it.

P.S. Don't PM me, I still won't tell you.

Free = DLP

Sim.....could it Be?

Kiddin' of course!
Reply
Reply
post #238 of 1782
12/24/08 at 9:14pm
Quote:
Originally Posted by Lawguy View Post

I just spent a few hours calibrating and learning the controls.

Thanks to Manni and Jeff for saving me a lot of work.

Greyscale and gamma calibration took about 15 minutes, after which I had a very flat greyscale and gamma dialed in nicely at 2.4. dEs of 1.X all around. Very nice.

I managed to get pretty good results with the gamut controls. There is definitely a limitation in the controls and I really suggest that we petition JVC for an upgrade here.

Still, as my results suggest, things turned out pretty good all around as compared to REC 709 with dE measured with the Accupel calculator.

I set the master color control at -7.

Here are my CMS settings:

R: H,5; S, -24; B, 21
Y: H, 2; S, -5; B, 30
G: H, -8; S, -5; B 30
C: H, 0; S, -18; B 30
B: H, 20; S, 11; B. -2
M: H, 0; S, -15; B, 20

These yielded the following measurements:

R: x, .641, y .334, Y 7.053, dE 2.7
G: x, .302, y .604, Y 22.789, dE 2.3
B: x, .155, y .073, Y 2.760, dE 3.7
Y: x, .423, y .509, Y 29.558, dE 2.0
C: x, .217, y .334, Y 22.779, dE 6.5
M: x, .322, y .157, Y 9.594, dE 2.5
W: x, .313, y .331, Y 33.282

There is some room to improve these a little, but not too much.

Cyan is the sore thumb sticking out. There is not enough headroom in the brightness control to raise it to where it should be. I quickly gave up. I found blue to be very stubborn. Most of the colors took just a few minutes to get where they needed to be. Blue can be improved but for some reason the controls do not respond as they should with this color, even though there is plenty of headroom left on them. It could be me but I spent as much time on blue as all the other colors combined. I don't know why.

I am pretty satisfied with these results numerically. I'll report back about how they look subjectively.

Nice work. Here is what your CMS chart looks like with those measurements for quick reference.
LL
Reply
Reply
post #239 of 1782
12/24/08 at 9:17pm
Quote:
Originally Posted by TomHuffman View Post

Given what we've seen here, I suspect they were quite large. This suggests what I suspected: that he simply attempted to get a good-looking CIE chart and then left it at that.

BTW, could someone try the following: measure the xy points (you can ignore luminance for this test) for RGBCYM using a 100% test pattern. Then repeat using a 75%, then 50%, and then 25% patterns. Then please report the 48 data points.

Something I found with the Radiance is that it corrects the primaries at different levels of stimulus quite well, but that it had more trouble with the secondaries. I'd be curious to see how the post-calibration RS20 tracks. For background, see

http://www.avsforum.com/avs-vb/showp...postcount=3101

and subsequent posts.

Tom - what's a good source for such test patterns? I normally use the AVD HD 709 disc but can't recall if such patterns exist on there.
Reply
Reply
post #240 of 1782
12/24/08 at 9:21pm
Is there a way to get the RS20 to keep some of the controls/settings displayed longer or indefinitely until closed? For instance when working with the gain and bias controls these controls close out if you don't hit anything for like 15 seconds. When I calibrated mine I think I spent more time going back through all the menu sequences each time to get back to the control to make a tweak than I did determining what the tweaks should be... Same thing for the lens controls.

IIRC with the RS20 I think the controls would stay up longer or perhaps indefinitely if the service menu was toggled. Anyone try this with the RS20?
Reply
Reply
New Posts  All Forums:Forum Nav:
  Return Home

AVS Top Picks

AVS › AVS Forum › Display Devices › Digital Hi-End Projectors - $3,000+ USD MSRP › Official JVC RS20 / HD750 Calibration and CMS thread