Basic Guide to Color Calibration using a CMS (updated and enhanced) [Text View] - Page 7

sjschaff

02-14-10, 08:38 PM

Under RGBCYM the 3 sliders are interrelated. Hue is the least affected by the other 2, but saturation and brightness have quite an influence over each other. Any large changes in one control will almost certainly require an adjustment in the other.

After noting what you've said I gave it another whirl. Now I see what you mean. I had to dramatically lower saturation on all primary and secondary colors while increasing brightness substantially. The hue changes were trickier since I found that this slider seems to be counter intuitive. Rather than lowering hue to lower x I found it necessary to raise it, and vice versa.

I'll need to validate my results with the ColorMunki I'm about to receive, but at least I now have the process down.

I'm assuming that this method will work the same way with the JVC RS35 I'll be getting shortly. :D

spongebob

02-15-10, 09:08 AM

The LG 55LH90 appears to be very calibration friendly.
Here are some recent results using the Chroma 5 and internal patterns for gray and AVSHD for color. LD was on but full screen patterns were used.
Although only a 2D CMS is available, adjusting the color control for L and the CMS for color and hue gets it pretty close.

While I normally use ChromaPure, Calman was used here to get smaller .PDF file. The .PDF files produced by ChromaPure average 2 MB and exceed the AVS file upload limit. Is there a way to get these smaller ?

Are the LG's more calibration friendly (CMS) than the Samsung 6/8 series? (Plasmas)

thx

bob

catmother

02-15-10, 11:25 AM

Are the LG's more calibration friendly (CMS) than the Samsung 6/8 series? (Plasmas)

thx

bob

No idea, do not have one, but the LH90 series offer the following:

Picture wizard which provides internally generated patterns for contrast, brightness, color, tint and sharpness
Choice of 2 point and 10 point grayscale adjustment using externally or internally generated patterns
Internally generated grayscale patterns from 10 to 100% stimulus
Three gamma choices (none of them very good)
Color gamut standard and wide
Black level low, high
White balance cool, medium and warm

Perhaps someone who owns one of the Samsung 6000/8000 series would post here for comparison.

gurudad

02-15-10, 12:51 PM

Hi guys haven’t posted here in some time but I just bought the LG lh90. i was wondering, after reading the forum, is the a calibration that is generally preferred by the owners of the LG55 " lh90

catmother

02-15-10, 02:03 PM

LG 55LH90 calibrated with ChromaPure, Chroma 5 and the ChromaPure cal pattern disk using the Panasonic S97.
Testing with AVSHD on a Panny BD60 BR player shows the same results, so it is not due to the player.

Grayscale, Color Gamut etc look great and so does the PQ watching BR, HD DVD and OTA broadcast material. But what is up with Gamma ?
Have run the Gamma module and made some adjustments with the Contrast and Brightness control but these interact a great deal and affect the low and high end to different extents. Adjusting one end generally screws up the other end. Settled on a compromise.

Why the rise above 40% stimulus.

Have made several runs with different BL settings and while these affect Gamma to a minor extent, the general rise is still there.
The LG has Gamma presets for low, medium and high. Of these medium is the best choice.

Attaching a few files here.
The .zip is the .xls and the .txt must be renamed with a calx extension. AVS does not like .calx

In the calibration report no changes were made between pre and post calibration. Any differences are due to meter and/or TV drift.

spongebob

02-15-10, 03:28 PM

No idea, do not have one, but the LH90 series offer the following:

Picture wizard which provides internally generated patterns for contrast, brightness, color, tint and sharpness
Choice of 2 point and 10 point grayscale adjustment using externally or internally generated patterns
Internally generated grayscale patterns from 10 to 100% stimulus
Three gamma choices (none of them very good)
Color gamut standard and wide
Black level low, high
White balance cool, medium and warm

Perhaps someone who owns one of the Samsung 6000/8000 series would post here for comparison.

Thanks

How about CMS? 3 Point?

thx

bob

rowland.johnson

02-15-10, 05:06 PM

catmother

02-15-10, 05:14 PM

Thanks

How about CMS? 3 Point?

thx

bob

No, 2D CMS only. As posted before the only adjustment for Lightness is the Color control but it seems to get the job done.

spongebob

02-15-10, 05:19 PM

No, 2D CMS only. As posted before the only adjustment for Lightness is the Color control but it seems to get the job done.

Thanks

I think the best Sammy's are only 2D also.

bob

TomHuffman

02-15-10, 05:31 PM

I think the best Sammy's are only 2D also.The 650 series and above Samsung plasmas and LCDs have a 3D CMS.

thomasl

02-15-10, 05:48 PM

Thanks

I think the best Sammy's are only 2D also.

Unless Samsung has changed their CMS implementation on their latest sets, the Samsung CMS is a 6 axis CMS that lets you adjust saturation, hue and brightness of each primary/secondary color. I just checked the manual for PN50B850 and the CMS looks to be the same as the one on my A550 LCD.

What the Samsungs are missing and would be nice to have is finer grained grayscale controls (like the 10pt control the LGs apparently have). I've noticed that the Samsungs I own all tend to have spikes of either red or blue are certain points.

cheers,

--tom

rfbrown94

02-15-10, 06:00 PM

Are the LG's more calibration friendly (CMS) than the Samsung 6/8 series? (Plasmas)

thx

bob

There are no tv's on the market with a CMS as good as samsungs. The only problem I have with the samsung is that the gamma isn't linear and there is no way to make it linear. Regardless, it's still OK.

spongebob

02-15-10, 06:46 PM

There are no tv's on the market with a CMS as good as samsungs. The only problem I have with the samsung is that the gamma isn't linear and there is no way to make it linear. Regardless, it's still OK.

Wow, that's impressive. Are there any cal reports here from a 650 or 860 Sammy Plasma?

thx

bob

mhdiab

02-16-10, 01:57 PM

First of all thanks Tom for a fantastic write-up. I have read it once but will read it many more times to get a better understanding.

So I will most likely buy the DVDO DUO with the new CMS upgrade. I realize it is not 100% there, but I am going to live with 95% considering some other options that it comes with.

I have none of the equipment needed and want to verify some things from the first post

1) Get a Colorimeter - seems like a no-brainer. I have an issue with which one to get. Any real difference between the two below? Others to consider for the price -- they seem like the right one based on the first post

http://www.amazon.com/gp/product/B000JLO31M/ref=s9_simi_gw_p23_t2?pf_rd_m=ATVPDKIKX0DER&pf_rd_s=center-2&pf_rd_r=116BRGB2F64BBTME71PT&pf_rd_t=101&pf_rd_p=470938631&pf_rd_i=507846

http://www.amazon.com/gp/product/B000CR78CE/ref=s9_simi_gw_p23_t1?pf_rd_m=ATVPDKIKX0DER&pf_rd_s=center-2&pf_rd_r=116BRGB2F64BBTME71PT&pf_rd_t=101&pf_rd_p=470938631&pf_rd_i=507846

2) Software -- I was simply going to get the HCFR that seems to work with those Colorimeters - will I miss out compared to a package deal with Calman included?

3) Test Patters - burning tonight :)

4) Light Meter optional -- ok so I need it to set the white level correct? So is it really optional how "bad" if I don't get one. It seems to be another $175 and while that alone isn't "that much" it adds upp.

Right now I am looking at enough that I just need to confirm if I should live without the light meter.

Thanks!

Knd

02-16-10, 02:12 PM

catmother

02-16-10, 02:27 PM

mhdiab

02-16-10, 03:14 PM

Get the cheaper colorimeter. The only difference is the software that X-Rite provides (its the same meter), and you're not going to use their software anyhow. HCFR works fine with this meter. I have been using both for the last 9 months.

You don't need a light meter, as you set white level (contrast) via the test disk also. So with the free disc from AVS, the free software from HCFR, and this meter, you can get started.

The following guide is very informative:
http://www.curtpalme.com/forum/viewtopic.php?t=10457&start=0&postdays=0&postorder=asc&highlight=

Thank you!

EDIT: Not even 10% into that link and WOW that is a true step by step guide. With that the first post here and many other posts available I am buying and going for sure. Thanks again

catmother

02-17-10, 11:00 AM

My strategy is to leave the Gamma setting at Medium and then to first balance the primaries at 10 - 100 IRE. Then, using the Y at 100 IRE calculate the Y required at each 10 - 90 IRE to achieve the target gamma (2.22 in my case). I then adjust the primaries at each 10 - 90 IRE for balance and the calculated Y. Usually, after 3 or 4 iterations the gamma looks like the attached.

Excellent advise.
In retrospect I should have realized this. It seems obvious now that adjusting the luminance at each of the 10 points would bring Gamma into compliance.
Alternating between the Gamma and White Balance modules in ChromaPure to adjust Gamma and check for RGB balance (laborious to be sure) did the job.

I actually started with the existing 10 point values and used the Gamma module to bring the Gamma to 2.2 at each point. Then used the White Balance module to set the gray scale, checking Gamma at each point and iterating.
Fortunately ChromaPure shows Y in the WB module but the spreadsheet available on AVS was used to check the required Y at each point.
After this was done, the Chromaticity and CMS values were checked and adjusted as required.
Although a bit more tweaking is required the result at this point is very good.

Post 1519

TomHuffman

02-17-10, 02:09 PM

Alternating between the Gamma and White Balance modules in ChromaPure to adjust Gamma and check for RGB balance (laborious to be sure) did the job.

I actually started with the existing 10 point values and used the Gamma module to bring the Gamma to 2.2 at each point. Then used the White Balance module to set the gray scale, checking Gamma at each point and iterating.
Fortunately ChromaPure shows Y in the WB module but the spreadsheet available on AVS was used to check the required Y at each point.
After this was done, the Chromaticity and CMS values were checked and adjusted as required.
Although a bit more tweaking is required the result at this point is very good.I would have done this in reverse order. Get the RGB balance correct at each point using the White Balance Module. Then use the Gamma Module to adjust the light output at each point. Since adjusting the light output up or down involves equal adjustments to RGB, this should have a minimal affect on the white balance. I haven't worked on an LG, but this is how it works on the Radiance, which has similar 10-point controls.

catmother

02-17-10, 05:41 PM

I would have done this in reverse order. Get the RGB balance correct at each point using the White Balance Module. Then use the Gamma Module to adjust the light output at each point. Since adjusting the light output up or down involves equal adjustments to RGB, this should have a minimal affect on the white balance. I haven't worked on an LG, but this is how it works on the Radiance, which has similar 10-point controls.

Yes, that is exactly the way it was done since the RGB (grayscale) was already balanced at the start.
The Gamma module was used to measure Gamma in continuous mode starting at 100% stim and and successively increasing R G an B by the same number of clicks until Gamma reached the correct value, 2.2 in this case. This was repeated at each stim value down to 10.

I suspected that equal increments of R G and B would unbalance the the associated gray point and that was indeed the case hence the need to iterate between White Balance and Gamma. On the LG the relationship between R G and B is not linear and also G has more effect in luminance followed by B and R. Strange that. I would have thought that R would alter luminance more that B.
In the end it worked out with a nice Gamma and a good grayscale

You have also pointed out in an earlier post that the Color control should be increased on or two clicks to reduce the Color error. But with a 2D CMS I had to use color to balance lightness in the primaries and secondaries.
The question is which is more important here for PQ, color error or lightness error.

BTW, the fellow who initially calibrated this 55LH90 last October trained at Sencore. He told me that Sencore training is done exclusively on LG TV's for flat panel work.
Perhaps because the extensive calibration controls exercises the skill of the trainees.

Most LG in the LH series LH40, LH 50, LH80, LH90 have the same user accessible calibration controls as do most of their plasma sets.

rowland.johnson

02-17-10, 08:01 PM

Yes, that is exactly the way it was done since the RGB (grayscale) was already balanced at the start.
The Gamma module was used to measure Gamma in continuous mode starting at 100% stim and and successively increasing R G an B by the same number of clicks until Gamma reached the correct value, 2.2 in this case. This was repeated at each stim value down to 10.

I suspected that equal increments of R G and B would unbalance the the associated gray point and that was indeed the case hence the need to iterate between White Balance and Gamma. On the LG the relationship between R G and B is not linear and also G has more effect in luminance followed by B and R. Strange that. I would have thought that R would alter luminance more that B.
In the end it worked out with a nice Gamma and a good grayscale

My technique is different but functionally equivalent. Once I know the requisite Y for each IRE I use the continuous measurement capability to achieve x = .3127, y = .329, and the requisite Y at each IRE. I.E. I don't do a balance pass of the IREs, followed by a gamma pass of the IREs. I do balance and gamma in the same pass. I suspect I might converge in fewer passes because it's apparent how the balance and luminance adjustments affect each other at each IRE.

TomHuffman

02-17-10, 08:29 PM

I suspected that equal increments of R G and B would unbalance the the associated gray point and that was indeed the case hence the need to iterate between White Balance and Gamma. On the LG the relationship between R G and B is not linear and also G has more effect in luminance followed by B and R. Strange that. I would have thought that R would alter luminance more that B.
In the end it worked out with a nice Gamma and a good grayscale

You have also pointed out in an earlier post that the Color control should be increased on or two clicks to reduce the Color error. But with a 2D CMS I had to use color to balance lightness in the primaries and secondaries.
The question is which is more important here for PQ, color error or lightness error.That's interesting. On the Radiance the effect is linear. The fact that G had the greatest effect is not surprising. Most of the light output of the panel is in the G channel.

When I recommended raising the Color control, I was talking about improving the primaries and secondaries. I noticed in your calibration report that the lightness values of some of the colors were a little low. The Color control should have a negligible effect on the white balance.

thomasl

02-18-10, 05:27 AM

That's interesting. On the Radiance the effect is linear. The fact that G had the greatest effect is not surprising. Most of the light output of the panel is in the G channel.

Yes, on the Samsungs when I raise/lower the RGB Offsets or Gains, the affect is linear - with only a change in the luminance response and not in the color balance. Depending on the particular model, moving these has a similar affect as moving contrast/brightness typically just a finer grained version of those controls. For example, if the brightness control seems like it would be optimal to be set between two numbers then you could use the RGB offsets to tweak it.

cheers,

--tom

catmother

02-18-10, 10:10 AM

My technique is different but functionally equivalent. Once I know the requisite Y for each IRE I use the continuous measurement capability to achieve x = .3127, y = .329, and the requisite Y at each IRE. I.E. I don't do a balance pass of the IREs, followed by a gamma pass of the IREsThe chief reason for doing this is that the process was started with a previously calibrated grayscale with dE =< 1.

I do balance and gamma in the same pass. I suspect I might converge in fewer passes because it's apparent how the balance and luminance adjustments affect each other at each IRE.After seeing the nonlinearity I switched to your method.

But thanks, it was your previous post that tipped me off to a way to correct Gamma.
Did you experience the nonlinear effect, equal increments (i.e changing R G and B by equal amounts) alters the white balance ?
A later poster here states that the Samsung is linear like the Radience.

rowland.johnson

02-18-10, 10:25 AM

Did you experience the nonlinear effect, equal increments (i.e changing R G and B by equal amounts) alters the white balance ?

Oh yeah. Green was the most, followed by Red, then Blue. I'm assuming that it's because of the luminance in each of the channels. Green 0.7152, Red 0.2126, and Blue 0.0722. One thing I did notice is that sometimes it seems that adjusting the Red 1 tick in one direction causes the measurement to go the opposite direction. But then adjusting the next tick in the same direction causes the measurement to go in the right direction. Seems like the Red response may be not quite monotonic.

It gets tedious at the lower IREs because it takes so long for my i1 D2 to gather enough light to produce a reading. At 10 IRE I make an adjustment and wait 1 minute to see the effect. 30 seconds might do it but I wanted to be sure. I'll be getting a Chroma 5 in the next day or two. Maybe that'll help some.

catmother

02-18-10, 10:30 AM

The rising Gamma also resulted after calibration last November by Bill Hergonson of Coast Calibration
The attached shows the pre and post Gamma on the LH90. Even the OOTB Gamma shows a rise.
At first I suspected the BR player used to display the patterns, but Bill used a Sencore generator and it also happens with the LG internally generated patterns, so it is intrinsic to the TV. Improperly programmed video processor perhaps.

His other results were excellent though.

bodosom

02-18-10, 11:00 AM

I'll be getting a Chroma 5 in the next day or two. Maybe that'll help some.

It should help. Not as much as a ColorMunki but still an improvement.

mhdiab

02-18-10, 11:11 AM

Knd

02-18-10, 11:50 AM

So I wanted to confirm something from the Curt Palme site.

I am going to calibrate a Projector (RS2) and was going to buy the EyeOne Display 2/LT - well that was until I read the statement below. Does anyone use the EyeOne for projectors? What is your thought about this? The Spyder 3 is significantly more in their package but I obviously don't want to waste the money on a product that won't work very well for projectors. I don't care about the tripod part as I have one

The EyeOne Display 2/LT meter is the least expensive, reasonably accurate sensor recommended for home theater enthusiasts for direct view display calibration (not projector calibration - see the CalMAN Enhanced Spyder3 below). We do not recommend the Display 2/LT for front projector calibration as it does not include a tripod mount and the diffuser is not great for illuminant measurement (pointed directly at the lens on front projectors). The EyeOne Pro, the Chroma 5 and the CalMAN Enhanced Spyder 3 are better choices for front projector calibration (see below).

I use the LT from a tripod with no problems whatsoever. I originally just used masking tape to hold it in place and it worked fine, but my tripod was a little short, so I bought a $5 piece of aluminum and bent a right angle into it and drilled a hole it in, bolted it to the tripod removable bracket and now have about a 24 inch extension to the tripod. I just use 2" wide masking tape to hold it to the aluminum extension I made. Don't use the diffuser, just remove it and point the LT at the screen. I position mine so that one of the tripod legs is against the screen wall. That way, every calibration the meter is the same distance from the screen (about 20"). I then move the meter up and down, left and right and tilted forward, tilted back, until I get the highest Y reading (I use HCFR and set it to take continuous readings). I then lock everything down on the tripod and start calibrating.

I've never used a spyder probe, but am completely happy with the LT.

mhdiab

02-18-10, 12:08 PM

Thanks for the feedback. Yeah the diffuser comment was kind of confusing to me. First it said don't use the LT with projector due to bad diffuser. Then separately I saw over there - remove the diffuser.........

appreciate the feedback - anyone else feel free to chime in. Looks like the Eyeone is all I need and a lot "easier" to spend just over $150 vs $380.......

Tripod I have one that should do it fine and as you say attaching it shouldn't be a major issue

TomHuffman

02-18-10, 12:10 PM

If you have a way to fix it to a tripod, then the D2 is perfectly fine for front projectors. For best results take measurements directly off the screen and do not use the diffuser. Even if you do use the diffuser for lamp readings, you can take off-screen readings also and use those as a correction for any inaccuracies introduced by the diffuser.

I am going to calibrate a Projector (RS2) and was going to buy the EyeOne Display 2/LT - well that was until I read the statement below. Does anyone use the EyeOne for projectors? What is your thought about this?

The EyeOne Display 2/LT meter is the least expensive, reasonably accurate sensor recommended for home theater enthusiasts for direct view display calibration (not projector calibration - see the CalMAN Enhanced Spyder3 below). We do not recommend the Display 2/LT for front projector calibration as it does not include a tripod mount and the diffuser is not great for illuminant measurement (pointed directly at the lens on front projectors). The EyeOne Pro, the Chroma 5 and the CalMAN Enhanced Spyder 3 are better choices for front projector calibration (see below).

catmother

02-18-10, 12:53 PM

Oh yeah. Green was the most, followed by Red, then Blue. I'm assuming that it's because of the luminance in each of the channels. Green 0.7152, Red 0.2126, and Blue 0.0722. One thing I did notice is that sometimes it seems that adjusting the Red 1 tick in one direction causes the measurement to go the opposite direction. But then adjusting the next tick in the same direction causes the measurement to go in the right direction. Seems like the Red response may be not quite monotonic.Yes indeed, have noticed that as well. Annoying but not fatal.

It gets tedious at the lower IREs because it takes so long for my i1 D2 to gather enough light to produce a reading. At 10 IRE I make an adjustment and wait 1 minute to see the effect. 30 seconds might do it but I wanted to be sure. I'll be getting a Chroma 5 in the next day or two. Maybe that'll help some.

The Chroma 5 will be faster and I use multiple readings, usually 2, to smooth out the variations from one reading to the next. But be advised that HCFR does not support the C5. CalMan and ChromaPure do.
I use ChromaPure with the C5 and really like that program.

This post suggests the Colormunki:
http://www.avsforum.com/avs-vb/showpost.php?p=18160602&postcount=1530
but that radiometer is only supported by CalMan at this time. AFAIK it is not NIST certified, the C5 is. Suggest you read the ColorMunki thread on the Cal forum before deciding.
http://www.avsforum.com/avs-vb/showthread.php?t=1215746

mhdiab

02-18-10, 01:34 PM

rowland.johnson

02-19-10, 12:20 AM

The Chroma 5 will be faster and I use multiple readings, usually 2, to smooth out the variations from one reading to the next. But be advised that HCFR does not support the C5. CalMan and ChromaPure do.
I use ChromaPure with the C5 and really like that program.

This post suggests the Colormunki:
http://www.avsforum.com/avs-vb/showpost.php?p=18160602&postcount=1530
but that radiometer is only supported by CalMan at this time. AFAIK it is not NIST certified, the C5 is. Suggest you read the ColorMunki thread on the Cal forum before deciding.
http://www.avsforum.com/avs-vb/showthread.php?t=1215746

Received the ChromaPure/Chroma 5 today.

Knd

02-19-10, 08:19 AM

I had a request to show my homemade tripod mount, so attached are the photos (I couldn't see how to attach photos to the PM).

Knd

02-19-10, 08:21 AM

And 2 more.

I also own an i1Pro and this works fine with that, although I run the masking tape all the way around the body of the i1P.

catmother

02-19-10, 10:58 AM

Received the ChromaPure/Chroma 5 today.

Congratulations. Look forward to hearing hearing about your experience with the new toys.

mhdiab

02-19-10, 01:11 PM

knd - thanks a bunch

PlasmaPZ80U

02-20-10, 11:06 AM

The link on the ChromaPure Site under the Lightness and Brightness article doesn't work.

http://www.chromapure.com/colorscience-lightness.asp

"What gamma value is closest to a perfect inverse of the lightness function? This source reports it as 2.32. Others claim that it is 2.35."

catmother

02-20-10, 01:07 PM

bodosom

02-20-10, 01:13 PM

The link on the ChromaPure Site under the Lightness and Brightness article doesn't work.

Looks like a cut and paste error.

Or a CMS (content management system) [sorry couldn't resist] typo.

PlasmaPZ80U

02-20-10, 03:00 PM

The link was in the sentence I quoted, not the article link I posted above.

(http://www.chromapure.com/http%7C//www.brucelindbloom.com/index.html)

TomHuffman

02-21-10, 01:20 AM

Man, you guys are fast. I just put that up yesterday. Link fixed.

bodosom

02-21-10, 11:45 AM

Link fixed.

By the way, doesn't that really mean the best gamma is at least two values?

TomHuffman

02-21-10, 11:57 AM

By the way, doesn't that really mean the best gamma is at least two values?I suppose it offers a range of reasonable values, which I think is a reasonable range, but the 2.32 value is very close to what Poynton cites (2.35), who is the "other" source I was referring to.

thomasl

02-21-10, 12:46 PM

Did you experience the nonlinear effect, equal increments (i.e changing R G and B by equal amounts) alters the white balance ?
A later poster here states that the Samsung is linear like the Radience.

Another possible reason for non-linear affects would be if your color balance is not linear when making changes to luminance. When altering the luminance at one of the control points, you'll be shifting the luminance response a bit thus shifting the color balance a bit as well. For example, if you have a spike of blue (e.g. 105%) at 65% stimulus and you lower the luminance around 70% using the ~70% 11pt control then it may be that you shift the blue spike closer to 70% stimulus thus it looks as though the color balance at 70% has changed. If the color balance across the entire stimulus range is linear then this effect wouldn't happen. On one of my Samsung displays, I have a distinct blue spike between 50%-60% but where I see it (50,55,60) depends on where I put the contrast control. So, I could see this as being one reason why one might see changes in the color balance as you alter the luminance at certain points even after initially optimizing the color balance.

cheers,

--tom

Meret

02-21-10, 01:12 PM

I have a question for those who have a lot more experience and technical knowledge at this than I do, although if I expend too more hours at this 'hobby', I should be able to qualify for a graduate degree pretty soon :)

My question is, which has more affect in reducing posterization/color banding, reducing saturation (all primaries are oversaturated) or the primary's brightness (all were too bright). After getting close on my grayscale (red is the problem child which the built-in defaults allow to run way low so I let it run a bit low but much better than the defaults), I started seeing skin tone highlights blow out on many HD TV stations (haven't seen it occur on BR disks).

So after giving up for a bit and reading more on the various approaches and compromises to balancing the color gamut, I focused on the saturations and primary's overall brightness. Basically, most colors were over-saturated at 25/50/75/100% but way under-saturated at 25/50/75 if I brought the 100% primaries in line (saturation wise). So I again compromised :) and decreased each primary's brightness using the color control and making a trade-off in saturation to get close in brightness (red is about 10% low, green/blue are 4-8% high) with most of the 25/50/75% saturations pretty close to target saturation leaving 100% somewhat over-saturated.

This seemed to eliminate (or make it almost undetectable) the skin tone posterization on those less-than-perfect TV stations. While I know grayscale can impact this area, since I reduced the red signal to avoid any blowout or clipping/tinting, that suggests this really is mostly a color gamut issue. I'm a type of person that just 'needs' to know what what fixed it (everything I've read suggests it's a 8 bit color conversation error for the larger gamut), but I would like to know, which of the two issues is the most important to target, primary too bright or primary too saturated (since I have to use both color controls to deal with the brightness issue).

Thanks in advance for any insight you can provide so I can sleep better at night :

P.S. My overall color reproduction is substantially better at this point.

catmother

02-22-10, 12:30 PM

[quote=Meret;18178100]I have a question for those who have a lot more experience and technical knowledge at this than I do, although if I expend too more hours at this 'hobby', I should be able to qualify for a graduate degree pretty soon :)
Snip....
IMHO the first task is to check for design deficiencies in the TV, using a benchmark like S&M:
http://www.avsforum.com/avs-vb/showthread.php?t=1206484&highlight=spears

The S&M showed among other things that the LG BD390 player was seriously clipping whites above 235 reference and that my LG 55LH90 had problems with cadence.
The white clipping would affect calibration so I now use the ChromaPure calibration disc in an old Panasonic S97 or AVSHD in a Panasonic BD60. The Getgray pattern disc is also very good.

The more one knows about the TV the better one is prepared to tackle calibration.

turbe

02-22-10, 02:09 PM

I had a request to show my homemade tripod mount, so attached are the photos (I couldn't see how to attach photos to the PM).

Here's one that is offered (I worked with another person on design).

It works well with the standard Chroma 5 / I1 Pro Mount and handle heavier meters like the K-10 and has vertical stability.. (adjustable horizontally - K-10 can go 20" from the center of the tripod head to the end of the boot.).

http://www.avsforum.com/avs-vb/attachment.php?attachmentid=167964&d=1266869361

1stsamdude

02-25-10, 11:37 PM

Just wanted to say that I am glad I ran into this Forum. I read all of the very first few pages and will continue to read them in order untill all have been read to better understand calibration. I am new to this and this is a great read for me. Special thanks to Tom for his write-up. It was very informative. I have bought some tools and software and am now venturing forward with my quest.
I will be calibrating my Sammy PN58B860 Plasma when I feel I am ready to tackel the job.

Thanks to all:)

spongebob

02-28-10, 03:29 PM

Guys

I have a dpt-94 meter and HCFR. please look at my before and after CIE charts.

I know the Pio CMS sucks but I'm concerned that I had to move Green to -11 to get it to the after point.

Complete after settings: 0,-1,-11,-1,+2, -2

thx

bob

bodosom

02-28-10, 03:47 PM

please look at my before and after CIE charts.

The dE is more important than the charts.

spongebob

02-28-10, 03:51 PM

The dE is more important than the charts.

Do you mean from the gray scale run?

bob

spongebob

03-08-10, 12:29 AM

Tom

How do you see de values on the CIE chart and how do you interpret them?

I understand how they work on the rgb scale.

thx

bob

ElwayLite

03-08-10, 08:17 AM

Hi all, this is a great thread and the first post has helped me learn a great deal. I bought a Enhanced C5 a few weeks ago and have been going over the learning process. The FAQ for setting color and tint with a meter is priceless.

I love the control, to be able to experiment with your tv and find what works best over time. I have a Regza 55" 670U LED and would like some honest input.

*) I have tried both Gamma 2.25 and 2.33, using the Gamma slider on the tv, and feel 2.33 loses a little bit of detail (appears to crush a bit). It also is not close to the 2.2 gamma at 10% Tom recommends.

*) I've calibrated with both LD on and off (I use full field patterns on the AVS disc) because of a slight blooming issue in a dark room. When LD is on, and you watch a movie with dark scenes in a dark room (6500k backlight only), blooming can be very obvious and the you can see the zones turning on and off. With LD off, you do not see this, but you get the standard bright screen with no signal, which is quite annoying to me as an ex plasma owner.

Now, if I watch the movie in the early AM, with a small amount of ambient light coming in (no real reflections), I can no longer see blooming and have deep rich blacks. IMHO, these tv's, as well as other LCD's, are not made for the complete dark viewing like newer plasmas. I'm fine with this, because I have a 60w light off to the side that does not interfere with the tv, which is how I watch nightime tv.

*) I have a new DMP-BD85 coming, and will recalibrate when it gets here, to ensure there are no differences from the BDP-320. My question is, with an LCD like this with some ambient light, should I set the gamma slider where I get 2.2 at 10% and then deal with the result, or is say a 2.25 gamma, with a higher 2.28 (what I have now) at 10% low enough. That is with Gamma -2, so I can raise it to get the lower 10% close to 2.2, but the way this tv usually works, is that a 2.2 gamma at 10% will yield a lower total gamma than 2.2, so I will not be able to hit 2.2 at 10%, it'll be slightly higher.

If I am making any sense ;). I've set brightness with 3 diff discs (always 9-10), and it always takes major brightness adjustments from there to effect the 10% number towards 2.2, as in one of Tom's recommendations. So, it's my thinking that I have to set brightness right, then get my gamma where I want with the slider.

Long story short, it's my guestimation that I need to set brightness, shoot for an overall 2.2 gamma with the slider, which will prob yield 2.23 or 2.24 at 10%.

Thanks for any straightening of my confusion ;)

TomHuffman

03-08-10, 08:58 AM

How do you see de values on the CIE chart and how do you interpret them?

I understand how they work on the rgb scale.A CIE chart doesn't show dE values. All it shows is a rough visual indication of hue and saturation errors.

TomHuffman

03-08-10, 09:04 AM

My question is, with an LCD like this with some ambient light, should I set the gamma slider where I get 2.2 at 10% and then deal with the result, or is say a 2.25 gamma, with a higher 2.28 (what I have now) at 10% low enough. That is with Gamma -2, so I can raise it to get the lower 10% close to 2.2, but the way this tv usually works, is that a 2.2 gamma at 10% will yield a lower total gamma than 2.2, so I will not be able to hit 2.2 at 10%, it'll be slightly higher.

If I am making any sense ;). I've set brightness with 3 diff discs (always 9-10), and it always takes major brightness adjustments from there to effect the 10% number towards 2.2, as in one of Tom's recommendations. So, it's my thinking that I have to set brightness right, then get my gamma where I want with the slider.

Long story short, it's my guestimation that I need to set brightness, shoot for an overall 2.2 gamma with the slider, which will prob yield 2.23 or 2.24 at 10%.I want to comment on this because it has come up before. As I tried to make clear in the original post, the 2.2 advice for setting brightness should only be used as a way making a more precise adjustment to the one you have already made visually using a pluge pattern. In fact, the AVSHD has a kind of pluge pattern for setting brightness that is so precise that no further aid is necessary.

The only reason I ever offered this advice was because I noticed that using the traditional 2% above black/2% below black pluge pattern left considerable wiggle room. The 2.2 gamma at 10% stim instruction was intended only as way of determining where WITHIN THAT RANGE you should set brightness. You cannot use this method by itself because many displays will not have a natural 2.2 gamma.

spongebob

03-08-10, 09:07 AM

A CIE chart doesn't show dE values. All it shows is a rough visual indication of hue and saturation errors.

That's what I thought. Was was referred in post #1555?

thx

bob

ElwayLite

03-08-10, 09:09 AM

Thanks Tom, and I do like the AVS disc pattern for Contrast and Brightness.

So, if I feel there is some crush/detail loss at 2.33, then I should set brightness properly and reduce the gamma slider(-3 to -2 or -1) to get closer to a 2.2 curve?

spongebob

03-08-10, 09:09 AM

I want to comment on this because it has come up before. As I tried to make clear in the original post, the 2.2 advice for setting brightness should only be used as a way making a more precise adjustment to the one you have already made visually using a pluge pattern. In fact, the AVSHD has a kind of pluge pattern for setting brightness that is so precise that no further aid is necessary.

.

Tom

Where on the disc is that?

thx

bob

ElwayLite

03-08-10, 09:10 AM

Check the AVS709 thread and download and burn it.

TomHuffman

03-08-10, 09:10 AM

That's what I thought. Was was referred in post #1555?The dE value of a color is a more accurate and comprehensive way of determining its accuracy than its position on a CIE chart.

TomHuffman

03-08-10, 09:11 AM

ElwayLite

03-08-10, 09:15 AM

The dE value of a color is a more accurate and comprehensive way of determining its accuracy than its position on a CIE chart.

When adjusting Tint/Cyan, using your method, I use the dE level too. On the 21% Red vs White, would the dE value be best for this too, or just trying to hit the 21% number?

bodosom

03-08-10, 09:34 AM

*) I have tried both Gamma 2.25 and 2.33, using the Gamma slider on the tv, and feel 2.33 loses a little bit of detail (appears to crush a bit). It also is not close to the 2.2 gamma at 10% Tom recommends.

Although he takes some pains to clarify this in the first post as well as here (http://www.chromapure.com/colorscience-gamma.asp) and elsewhere I think it's time to modify item 4.

ElwayLite

03-08-10, 09:39 AM

Its looking like I should reset user settings, adjust bright and contrast, tighten up grayscale, then check the gamma slider starting at 0, and see which setting gives me the flattest line between 2.2 and 2.3. Then go back and check the other settings.

spongebob

03-08-10, 09:59 AM

Check the AVS709 thread and download and burn it.

I have it. Wondered what pattern Tom was referring to?

thx

bob

ElwayLite

03-08-10, 10:21 AM

I assumed it was the standard brightness pattern where you want bars 17 and higher showing and 16 and below off.

TomHuffman

03-08-10, 10:22 AM

When adjusting Tint/Cyan, using your method, I use the dE level too. On the 21% Red vs White, would the dE value be best for this too, or just trying to hit the 21% number?dE is a comprehensive method for grading color difference. 21%-of-white is simply a straightforward way to set the brightness of red correctly. Since the Color control has the biggest effect on color brightness (within a reasonable range of adjustment), it's a good way to adjust the Color control.

spongebob

03-08-10, 10:26 AM

The dE value of a color is a more accurate and comprehensive way of determining its accuracy than its position on a CIE chart.

Tom

Can you look at this reading and tell me why Magenta's de is 28, even though it looks closer than yellow that has a de of 8.9?

This is from a Pio PRO-101FD.

thx

bob

ElwayLite

03-08-10, 10:26 AM

de is a comprehensive method for grading color difference. 21%-of-white is simply a straightforward way to set the brightness of red correctly. Since the color control has the biggest effect on color brightness (within a reasonable range of adjustment), it's a good way to adjust the color control.

10-4.

spongebob

03-08-10, 10:37 AM

A CIE chart doesn't show dE values. All it shows is a rough visual indication of hue and saturation errors.

Using, HCFR, when I hover my mouse over a reading on the CIE chart, It gives me a de reading. Am I missing something?

thx

bob

spongebob

03-08-10, 11:05 AM

I assumed it was the standard brightness pattern where you want bars 17 and higher showing and 16 and below off.

I have trouble with that pattern :)

bob

ElwayLite

03-08-10, 11:13 AM

I have trouble with that pattern :)

bob

??

spongebob

03-08-10, 11:33 AM

??

hard to distinguish the difference in the flashing bars. I think there are better pluge patterns.

bob

TomHuffman

03-08-10, 11:52 AM

Using, HCFR, when I hover my mouse over a reading on the CIE chart, It gives me a de reading. Am I missing something?No, that's just something HCFR provides. It is not part of the CIE chart itself.

bodosom

03-08-10, 12:19 PM

hard to distinguish the difference in the flashing bars.

That's because you have an Elite monitor (or it's too bright for proper calibration) -- although they might reduce the brightness of the labels in the next version which is also a problem for Kuro monitors in a light controlled environment.

spongebob

03-08-10, 03:04 PM

If you guys have HCFR, can you check mile file above (#1573) and see why the magenta Delta E is so high compared to where it is on the chart and compared to the other colors?

thx

bob

bodosom

03-08-10, 03:18 PM

If you guys have HCFR

There's a thread for C-HCFR reports/advice (http://www.avsforum.com/avs-vb/showthread.php?t=786966).

spongebob

03-08-10, 03:30 PM

There's a thread for C-HCFR reports/advice (http://www.avsforum.com/avs-vb/showthread.php?t=786966).

Done, but I thought this was CMS related and that most people here used or were familiar with HCFR.

thx

bob

TomHuffman

03-08-10, 03:34 PM

Can you look at this reading and tell me why Magenta's de is 28, even though it looks closer than yellow that has a de of 8.9?

This is from a Pio PRO-101FD.I don't know what you are looking at.

spongebob

03-08-10, 03:45 PM

I don't know what you are looking at.

Using HCFR, I am looking at the Delta E readings for the pri/sec. colors. You can check it in the measurements window under primaries and secondaries or you can hover the mouse over the reading on the CIE chart. Magenta DE is 28 and looks to be fairly close on the chart where it should be. Other colors are much closer (Delta E) Sorry if I'm not being clear enough.

bob

glaufman

03-08-10, 03:56 PM

What Tom's saying is that when we open your file, the dE for the Magenta primary is 7.4, not 28, whether you look at it in the measurements window or hover of the Magenta in the CIE chart.

PlasmaPZ80U

03-08-10, 03:59 PM

Done, but I thought this was CMS related and that most people here used or were familiar with HCFR.

thx

bob

I've replied to your post in the ColorHCFR thread. It appears magenta is 7.1% too dim relative to the target luminance.

spongebob

03-08-10, 04:40 PM

What Tom's saying is that when we open your file, the dE for the Magenta primary is 7.4, not 28, whether you look at it in the measurements window or hover of the Magenta in the CIE chart.

I must be having a twilight zone day, then :)

It says DE 28 when I open it. I'll post again after i just confirmed that.

I did get your response on the other thread, Plasma. Thanks

bob

glaufman

03-08-10, 04:55 PM

Check what you have set as your references.

spongebob

03-08-10, 05:35 PM

Check what you have set as your references.

REC 709.

thx

bob

glaufman

03-09-10, 08:03 AM

And the white point?
169475

spongebob

03-09-10, 08:23 AM

And the white point?
169475

Sorry, Not sure what you mean. I do check white when doing a run but I see the missing data. All my runs are missing the white delta e numbers but I have no idea why!

thx

bob

glaufman

03-09-10, 08:51 AM

I meant what white point are you using for a reference?

spongebob

03-09-10, 09:15 AM

I meant what white point are you using for a reference?

I try to remember to set it at D65 but it doesn't save in the preferences.

I could have forgot before this run. Would that make a difference?

thx

bob

glaufman

03-09-10, 04:42 PM

I think so. Open your file, play with the white point reference setting, and see how your DE changes.

spongebob

03-09-10, 05:19 PM

I think so. Open your file, play with the white point reference setting, and see how your DE changes.

No difference.

thx

bob

glaufman

03-09-10, 05:27 PM

Strange, it did on my end. What version HCFR are you using, and on what OS?

spongebob

03-09-10, 07:02 PM

Strange, it did on my end. What version HCFR are you using, and on what OS?

Version 2.0, Vista 32 bit.

What is strange is when I go to preferences and check the D65 box, it won't let me apply (grayed out). Because it's already selected?

bob

spongebob

03-09-10, 07:04 PM

No difference.

thx

bob

Sorry, changed to D75 or D55, big change on chart.

bob

glaufman

03-10-10, 08:30 AM

So what does the dE become with D65 as the reference?
You could try upgrading to 2.1.0... but for the record, I'm running 2.0.1 here at work...

spongebob

03-10-10, 08:44 AM

So what does the dE become with D65 as the reference?
You could try upgrading to 2.1.0... but for the record, I'm running 2.0.1 here at work...

My first charts are correct @ D65. I couldn't change to D65 cuz it was already set there LOL.

What about colorspace 1 or 2 on the Kuro? Would changing that give different RGB/CIE numbers?

thx

bob

glaufman

03-10-10, 09:29 AM

spongebob

03-10-10, 11:07 AM

I would probably give different numbers, but what the display is putting out isn't as much the problem as why you're getting different dEs out of HCFR than the rest of us...
do you have another file open you're using as a reference measure?

No other files open. have closed and opened HCFR many times, same numbers. Any other settings that could be it?

BTW, what de are you guys getting for magenta, again?

thx

bob

spongebob

03-10-10, 11:17 AM

I don't know what you are looking at.

I'm looking at this!

bob

TomHuffman

03-10-10, 12:57 PM

I'm looking at this!Both cyan and magenta are incorrectly reported. They should be 2.8 and 7.4, respectively using CIELUV against a Rec. 709 target. RGBY are correct.

I seem to recall that an early build of HCFR had this error. This is an HCFR-specific problem.

spongebob

03-10-10, 03:04 PM

Both cyan and magenta are incorrectly reported. They should be 2.8 and 7.4, respectively using CIELUV against a Rec. 709 target. RGBY are correct.

I seem to recall that an early build of HCFR had this error. This is an HCFR-specific problem.

I'm downloading 2.1 right now. Was running 2.0.

Thanks, Tom

bob

edit: 2.1 Fixed it !!!

spongebob

03-10-10, 07:42 PM

Tom,

What is the Yx value on the spreadsheet? I'm trying to learn about CIE94, etc and want to enter my HCFR readings into the chart. What is the HCFR equivalent?

thx

bob

ElwayLite

03-11-10, 09:15 PM

Hey Tom, don't want to derail your thread, but I was wondering what your take is on local dimming lcd's. Calibrate on or off? I understand you can use full field patterns on the AVS disc, and once you calibrate with it off or on, you have to leave it that way. I'm coming to the opinion that LD on looks real good, but kills detail.

Have you calibrated any LD sets?

TomHuffman

03-12-10, 01:53 PM

ElwayLite

03-12-10, 02:07 PM

Thanks. What I've found, is that overall, I prefer the LD on, but it definitely has its issues. IMHO, this keeps Plasma as the top tech. The Regza is real hard to dial in with LD off, and deltaE's under 3 are hard to come by. Much better with it on.

I goofed around last night and found I was able to get more light output with a lot less backlight and contrast, by turning off the scanning backlight and it also allowed for a much straighter gamma line at 2.32. Dunno why the scanning backlight would effect gamma(part of it is prob the lower backlight and contrast), but it does. I'm amazed at how diff features effect different things between the brands and models. I feel for professional calibrators.

catmother

03-14-10, 08:21 PM

Yes, with mixed results. Don't really have any advice to give.

You are too modest Tom.
Using ChromaPure and a Chroma 5 to calibrate the 55LH90 the results are very good. More importantly OTA and even CATV look great on my set.
http://www.avsforum.com/avs-vb/showpost.php?p=18230186&postcount=331

This calibration was done with LD on and field patterns using the ChromaPure pattern disc playing on a Panasonic S97 SD DVD player.

Scott_R_K

04-05-10, 03:44 PM

It doesn't look like this has been Posted anywhere yet , but for those of you interested in Poynton , here is a Link (pdf) of the first Article that SpectrCal will be hosting...

http://storage.pardot.com/2742/29527/Poyntons_Vector_Mar_2010.pdf

Enjoy :D

Scott.................................

bodosom

04-05-10, 04:08 PM

It doesn't look like this has been Posted anywhere yet

nt!, kicking it old-school! Next up stilb.

tqn

04-19-10, 01:50 AM

Hi Tom,
what do you think of the "75% saturation/100% brightness" method to set primary and secondary colors? Basically that's aiming to hit the 75% saturation point for colors instead of the usual 100% saturation point.

http://www.avsforum.com/avs-vb/showthread.php?t=1134710

It makes great sense to me if the display colors are grossly under-saturated at 25/50/75% saturation points.

For example, I calibrated my Samsung A650 plasma using the conventional method of 100% saturation/100% brightness colors (of course after setting brightness, contrast, gamma, grayscale, etc), but was never really happy with the results. I somehow came across the thread linked to above and learned about "75% saturation" and "saturation-luminance" (big thanks to Dan's excellent spreadsheet).

I was quite surprised to see my plasma's performance here. The 25/50/75% saturation points for most of the colors were well short of the mark. Here is the color gamut AFTER I tried my best to use 75% saturation method. The Samsung primaries and secondaries can't be saturated any further, so that's about as good as it gets.

http://i123.photobucket.com/albums/o298/tqnn/Colorgamut.jpg

Perhaps more disturbing are the saturation-luminance measurements I obtained. Here are the charts for RED, which are the most problematic (and noticeably incorrect) color. Below are 2 charts. The main color control is set to default in both. In the first chart, the RED brightness is default. In the second chart, the RED brightness (only) was lowered using the Sammie's 3D CMS.

http://i123.photobucket.com/albums/o298/tqnn/RedSaturationdefaults.jpg

http://i123.photobucket.com/albums/o298/tqnn/RedSaturationColordefaultRed20.jpg

The effect I see of adjusting red brightness is to get the 100% red saturation point close to the correct brightness. But look at 75% saturation and lower! They are too bright. Is there any way I can get this *entire* curve to lie reasonably flat and close to the reference line?

TomHuffman

04-19-10, 09:28 AM

what do you think of the "75% saturation/100% brightness" method to set primary and secondary colors? Basically that's aiming to hit the 75% saturation point for colors instead of the usual 100% saturation point.Not sure what I think. I guess that it seems reasonable to me to choose any CMS setting that does the best job possible of minimizing error for as many data points as you have. Taking saturation and luminance readings from multiple levels just gives more data points. Of course, what you are likely to find is that no one CMS setting will get the results right for all points, so it is a matter of compromise.

spongebob

04-19-10, 10:18 AM

Not sure what I think. I guess that it seems reasonable to me to choose any CMS setting that does the best job possible of minimizing error for as many data points as you have. Taking saturation and luminance readings from multiple levels just gives more data points. Of course, what you are likely to find is that no CMS settings get the results right for all points, so it is a matter of compromise.

But isn't 75% more "real world" than 100%?

bob

TomHuffman

04-19-10, 11:31 AM

But isn't 75% more "real world" than 100%?I can't think of any reason for why a particular shade of red would be more "real world" than another. In fact, the entire Rec. 709 gamut is not particularly "real world" insofar as it excludes many higher saturation colors entirely.

PlasmaPZ80U

04-19-10, 11:37 AM

But isn't 75% more "real world" than 100%?

bob

Do you mean 100% saturation/75% stimulus or 75% saturation/100% stimulus? I believe 100% saturation/75% stimulus is better than 100% saturation/100% stimulus since 75% stimulus produces about 50% light output, which is closer to most content than 100% light output.

rowland.johnson

04-19-10, 12:06 PM

I can't think of any reason for why a particular shade of red would be more "real world" than another. In fact, the entire Rec. 709 gamut is not particularly "real world" insofar as it excludes many higher saturation colors entirely.

Maybe what he meant is that in most content 75% saturation is seen more than 100% stauration. Anyway, with my LG 47LH90 I had to choose between getting either 100% saturation right and getting 75% saturation and lower right. Adjusting for a correct 75% saturation response produced the smallest error at the remaining staurations. And, assuming that there was more content at 75% and lower made it easier to accept the tradeoff

TomHuffman

04-19-10, 01:00 PM

That's how I understood the post. I can only repeat my reply. 100% saturation simply means a particular shade of red in the Rec. 709 gamut. I can think of no reason why THAT shade of red would be any more or less represented in the real world than other shades of red, 75% in particular. Think of the red one sees in a red rose. My guess is that that's more saturated than 100% Rec. 709, not less.

Using 75% as a target to minimize dE errors across the range of saturations within the gamut is a perfectly reasonable approach, but that strategy--even if successful--says nothing about what level of saturation is more typical in available content.

Maybe what he meant is that in most content 75% saturation is seen more than 100% stauration. Anyway, with my LG 47LH90 I had to choose between getting either 100% saturation right and getting 75% saturation and lower right. Adjusting for a correct 75% saturation response produced the smallest error at the remaining staurations. And, assuming that there was more content at 75% and lower made it easier to accept the tradeoff

tqn

04-19-10, 05:37 PM

That's how I understood the post. I can only repeat my reply. 100% saturation simply means a particular shade of red in the Rec. 709 gamut. I can think of no reason why THAT shade of red would be any more or less represented in the real world than other shades of red, 75% in particular. Think of the red one sees in a red rose. My guess is that that's more saturated than 100% Rec. 709, not less.

Using 75% as a target to minimize dE errors across the range of saturations within the gamut is a perfectly reasonable approach, but that strategy--even if successful--says nothing about what level of saturation is more typical in available content.

Thanks, Tom. So I guess the answer is that it depends on the display, and how closely any given display hits the targets at the various saturation levels.

Do you have any insight on how one can correct that funky red saturation-luminance response that I showed in the bottom two graphs in my message above? Here they are again for reference. First is default; second has Red Y lowered.

http://i123.photobucket.com/albums/o298/tqnn/RedSaturationdefaults.jpg

http://i123.photobucket.com/albums/o298/tqnn/RedSaturationColordefaultRed20.jpg

The bottom curve tracks more closely to the reference but is still too bright at 75% saturation levels and below.

spongebob

04-19-10, 07:24 PM

Do you mean 100% saturation/75% stimulus or 75% saturation/100% stimulus? I believe 100% saturation/75% stimulus is better than 100% saturation/100% stimulus since 75% stimulus produces about 50% light output, which is closer to most content than 100% light output.

That's probably what I meant. Just that we dont' look at colors at 100% saturation all the time or even most of the time when we watch TV?

bob

TomHuffman

04-19-10, 09:55 PM

Do you have any insight on how one can correct that funky red saturation-luminance response that I showed in the bottom two graphs in my message above? Here they are again for reference. First is default; second has Red Y lowered.

http://i123.photobucket.com/albums/o298/tqnn/RedSaturationdefaults.jpg

http://i123.photobucket.com/albums/o298/tqnn/RedSaturationColordefaultRed20.jpg

The bottom curve tracks more closely to the reference but is still too bright at 75% saturation levels and below.You can just lower red brightness more, but since the line isn't straight, any adjustments will have non-linear results.

tqn

04-19-10, 11:33 PM

You can just lower red brightness more, but since the line isn't straight, any adjustments will have non-linear results.

That sucks. Seems I'll never be able to get rid of the reddish skin tones on my Samsung plasma then unless I make big compromises elsewhere. Perhaps the evolution of CMS should include controls for making adjustments to luminance vs. saturation.

Joelc

04-20-10, 04:55 AM

tqn

04-20-10, 05:01 AM

It already exists, it is called a Lumagen RadianceXE...

Sweet. Now I just need to sell my TV...and my audio system...to pay for it.

Any chance of that tech being built into a TV's CMS? :p

TomHuffman

04-20-10, 11:49 AM

That sucks. Seems I'll never be able to get rid of the reddish skin tones on my Samsung plasma then unless I make big compromises elsewhere. Perhaps the evolution of CMS should include controls for making adjustments to luminance vs. saturation.Not sure I understand. You can make independent saturation and luminance adjustments with the Samsung CMS. What you can't do is make independent adjustments at different levels of saturation (25%, 50%, etc.) or amplitude. There is no commercially available product that I am aware of that offers this functionality.

tqn

04-20-10, 03:24 PM

Not sure I understand. You can make independent saturation and luminance adjustments with the Samsung CMS. What you can't do is make independent adjustments at different levels of saturation (25%, 50%, etc.) or amplitude. There is no commercially available product that I am aware of that offers this functionality.

Yes, I am looking to somehow, either directly or indirectly, make those independent adjustments at different levels of saturation. As it stands now, the Samsung CMS will allow me bring Red brightness down to the correct level ONLY at 100% saturation, leaving the 25/50/75 too bright. I believe that is why regardless of how hard I try to calibrate this display, I always see reddish skin tones.

I could bring the red brightness @ 100% saturation to well below reference to help bring down the curve at 50/75% saturation levels, but that will compromise deep reds, while still not helping at the 25% saturation.

PlasmaPZ80U

04-21-10, 06:29 PM

Yes, I am looking to somehow, either directly or indirectly, make those independent adjustments at different levels of saturation. As it stands now, the Samsung CMS will allow me bring Red brightness down to the correct level ONLY at 100% saturation, leaving the 25/50/75 too bright. I believe that is why regardless of how hard I try to calibrate this display, I always see reddish skin tones.

I could bring the red brightness @ 100% saturation to well below reference to help bring down the curve at 50/75% saturation levels, but that will compromise deep reds, while still not helping at the 25% saturation.

Try the Bruce Linbloom method for the brightness of red and hue of yellow. I used it simply for color/tint but I'm sure it can be used with the CMS as well.

http://www.avsforum.com/avs-vb/showthread.php?p=18476166#post18476166

ElwayLite

05-07-10, 07:21 PM

kjgarrison

05-08-10, 12:15 AM

This is a little hard to describe, but I've been mulling this over for awhile. So, I'll give it a try.

All within Rec709 specifications (say). Just talking about one color (red) at this time, but it applies to all of them.

The given: The red Y value at 100% saturation should be 21% of 100% white Y.

The question: What should the red Y value be at 75%, 50%, and 25% red saturation? It seems to me that each color's Y value is like a tent pole and the white Y value is the center pole at D65. Looking at it from the side, the red pole (Y) is 21% of the center pole (White/D65). As the spots for lesser amounts of red saturation proceed towards D65 why doesn't the proper Y value go up so that it is linear from the starting point 100% red to 0% red (100% white)?

It just doesn't make mathematical sense that red Y stays at 21% of white Y as the saturation of red progresses from 100% ->75% ->50% ->25% ->10% ->5% ->4% ->3%, etc. all the way to 0%, only to suddenly shoot up instantaneously to 100% white Y.

I guess another way to ask this if this isn't clear, is to point out that there is a straight line from white to each color's Rec709 spot on the 2D CIE chart. What does this line look like from the "side" where vertical relationships (aka Y values) can be seen? Is it linear, or is there some other function? It just can't be flat (21%); that would be too weird as it approaches white and approaches 0% red saturation.

Make sense?

TomHuffman

05-08-10, 01:52 AM

Rec. 709 is a specification for a full gamut of color and as such it is defined as only the outer limits of that gamut. Thus, within Rec. 709, I don't really think that less than full saturation has much meaning.

However, all else remaining equal, as you move a color towards the white point, it necessarily becomes brighter.

R100% G50% B50% (desaturated red) is brighter than R100% G0% B0% (full red).

This is a little hard to describe, but I've been mulling this over for awhile. So, I'll give it a try.

All within Rec709 specifications (say). Just talking about one color (red) at this time, but it applies to all of them.

The given: The red Y value at 100% saturation should be 21% of 100% white Y.

The question: What should the red Y value be at 75%, 50%, and 25% red saturation? It seems to me that each color's Y value is like a tent pole and the white Y value is the center pole at D65. Looking at it from the side, the red pole (Y) is 21% of the center pole (White/D65). As the spots for lesser amounts of red saturation proceed towards D65 why doesn't the proper Y value go up so that it is linear from the starting point 100% red to 0% red (100% white)?

It just doesn't make mathematical sense that red Y stays at 21% of white Y as the saturation of red progresses from 100% ->75% ->50% ->25% ->10% ->5% ->4% ->3%, etc. all the way to 0%, only to suddenly shoot up instantaneously to 100% white Y.

I guess another way to ask this if this isn't clear, is to point out that there is a straight line from white to each color's Rec709 spot on the 2D CIE chart. What does this line look like from the "side" where vertical relationships (aka Y values) can be seen? Is it linear, or is there some other function? It just can't be flat (21%); that would be too weird as it approaches white and approaches 0% red saturation.

Make sense?

kjgarrison

05-08-10, 11:55 PM

Thanks for the reply Tom. You were who I was hoping for an answer from. It seems that you are agreeing with what I am saying.

Your example confused me though (not that hard to do.) As you move along the line from Rec 709 red point toward D65 (increasingly desaturated), why would the percentages of Green and Blue increase, and why would the percentage of Red stay the same? I would think 50% red would be R50% G0% B0%. No?

Isn't this concept useful when you happen to have a TV that doesn't have all the colors on their Rec 709 spots? I downloaded a spreadsheet called Gamut, Saturation, Brightness, dE Chart from someplace, and I was recently looking at it. The target Y values for all colors remained basically unchanged for all saturations. I believe you and I have just agreed that they should go up, and significantly so.

Color HCFR, for one that I know of, has 75%, 50%, and 25% test windows. I think I've seen somebody suggest using 75% to do calibration adjustments. I'm thinking that if you use the same percentage of White Y as the target for 75% Red Y that you use for 100% Red Y (21%) the colors would look dull. If the relationship is linear, 75% Red Y value would be 1/3 of the difference between White Y and 100% Red Y [((100-21)/3) + 21]. This comes out to 47.3%

At 50% Red, it's Y value should be a little over 60% of White Y. Again if this is linear. It could be geometric. I don't know.

Make sense?

Rec. 709 is a specification for a full gamut of color and as such it is defined as only the outer limits of that gamut. Thus, within Rec. 709, I don't really think that less than full saturation has much meaning.

However, all else remaining equal, as you move a color towards the white point, it necessarily becomes brighter.

R100% G50% B50% (desaturated red) is brighter than R100% G0% B0% (full red).

bodosom

05-09-10, 12:21 AM

As you move along the line from Rec 709 red point toward D65 (increasingly desaturated), why would the percentages of Green and Blue increase, and why would the percentage of Red stay the same? I would think 50% red would be R50% G0% B0%. No?

No.

You're confusing saturation and relative brightness. Note that brightness (lightness, luma, luminance) is the unplotted z (or Y) axis of a chromaticity chart.

From the first post:

Terminology

xyY - A common method for precisely measuring color performance. x and y are the coordinates that plot out a color on the triangular CIE chromaticity chart mentioned above. This graphically represents the established definitions of the color spectrum. Y is the brightness of the color. This is not plotted by the xy coordinates.
Saturation - the colorfulness of the color relative to its own brightness. A color's saturation is displayed on the CIE chart as the distance from the white point. Add saturation to a color and it will begin to appear excessively deep and rich-red becomes crimson. Undersaturate a color and it will begin to appear as a less colorful version of a similar brightness--red becomes pink.
Hue - the primary characteristic of color that allows us to distinguish one color from another. A color's hue is measured by its angle to the white point. When a color's hue is off, its appearance will seem contaminated by other colors. For example, red that is too yellowish will begin to seem orange. Blue that is too reddish will begin to appear purplish.
Brightness - the luminance of color. The brightness of color (or white) can be measured by a simple light meter.

TomHuffman

05-09-10, 12:48 AM

As you move along the line from Rec 709 red point toward D65 (increasingly desaturated), why would the percentages of Green and Blue increase, and why would the percentage of Red stay the same? I would think 50% red would be R50% G0% B0%. No?No, that would be a fully saturated, but dim, red. White is R100% G100% B100%.

Isn't this concept useful when you happen to have a TV that doesn't have all the colors on their Rec 709 spots? I downloaded a spreadsheet called Gamut, Saturation, Brightness, dE Chart from someplace, and I was recently looking at it. The target Y values for all colors remained basically unchanged for all saturations. I believe you and I have just agreed that they should go up, and significantly so.For purposes of testing linear behavior at lower levels of saturation, it really doesn't matter what the brightness is. Many have used constant brightness for these tests, but that's just an optional convention.

I think I've seen somebody suggest using 75% to do calibration adjustments. I'm thinking that if you use the same percentage of White Y as the target for 75% Red Y that you use for 100% Red Y (21%) the colors would look dull.Again, since we are only testing saturation, the brightness doesn't matter. Some people do this because by using a CMS to get accurate saturation at the Rec. 709 specified points results in desaturated performance for the rest of the spectrum. Using 75% saturation as a target may result in lower average dEs throughout the entire spectrum than you would get if you used the standard Rec. 709 points.

docrings

05-10-10, 03:58 AM

Forgive me if this has been asked before.

Whenever I use to meter to set color and tint, the pic just seems a little more drab, which is probably because I come up with a slightly lower color setting than I would with the filters (Color 0 and Tint 8 with meter/Color 3 and Tint 11 with 709 disc and blue only mode).

Is the meter the definitive way to adjust, and it's prob that I'm used to over saturated colors?

Calman and Chroma5

fL---100%W----100%R
x---0.3135---0.6442
y---0.3279---0.3150
Y---40.1908---8.4599

21% of 40.1908 is 8.44

Just making sure Im not missing something, thanks.

I think you are indeed "used" to oversaturated television colors... it's easy to get hooked on all those vibrant neon reds and greens.

When you look at the screen, think about the colors of the faces and flesh tones on the screen? Hold your own arm up (if you're caucasian) and think: "Does the skin on the screen look real, like mine... or deeper orange or red?"

Look at the leaves on the trees on the television, and look out your window on a sunny day... do the leaves on the screen appear a natural green, not overly saturated? Or do they appear neon green (unnatural?)

If you do a search on AVSforum for "steaming rat" method of tweaking your set, this might help... it helps me "calibrate" TV's for which I don't have my photometer...

Cheers!

Doc

janos666

06-05-10, 06:55 PM

I am a little embarrassed about the calibration target settings and I hope somebody can help me.
I have an EyeOne Display 2 with Basiccolor 4.1 software and I would like to calibrate my LCD display for general end-user type PC usage, like web browsing, video games, movies, etc...

I have to choose between Gamma 2.2 (or 2.22 -> which one is theoretically better?) and sRGB (close but not similar to x^2.2) tonal response curves.

The default profile in the Windows CMS settings is sRGB IEC61966-2.1, so it suggests to set the calibration target to sRGB because the theoretically perfect (Windows-)PC monitor should have sRGB characteristics.

But there is a lot of other things that suggests that I should choose 2.2.
- Some calibration softwares (like OneMatch) do not even offer the sRGB response curve as a target.
- Cheap monitors usually do not have this kind of settings but the more expensive ones offer some gamma settings like 1.8, 2.2, 2.3, but I never saw if any of them offer sRGB gamma. (If there is an sRGB mode, that stands for a D65 WP or sometimes for sRGB gamut correction, but gamma is independent...)
- I would use the same settings that artists/producers/developers/engineers use to see the same what they saw during their works...

It is not clear for me. For example:
- Do web designers use their monitors with sRGB response curves as they should?
- If they use gamma 2.2 then do they convert their final product with care about the fact that end-users (theoretically) will watch it on sRGB monitors?
- Or is it the plan that they work and publish their products with gamma 2.2 and I should watch it on an sRGB display and it will be perfect because they work in controlled environment and end-users watch it on brighter rooms than a studio (or my PC-room...)?

What about game studios? Do (PC) video game developers care about this? Do they use displays with Rec709 or sRGB characteristics, or do they calibrate their displays to gamma 2.2?

Rec709 has it's own tonal response curve. There is a solution to convert it for PC monitors (either sRGB or gamma 2.2). But do HDTVs has this transfer function in the practice or do they have a gamma about 2.2?

And the last one: Here is this number 1/0.45=2.22222˙. Which one is more theoretically accurate number: 2.2 or 2.22? I saw the 0.45 in Rec709 documents but that system has a different transfer function than old CRTs (the mother of the x^2.2), so...

Sorry, if it was too much hotchpotch for a day. :)

demoniacal 360

06-07-10, 02:18 PM

hey guys i just was wondering my denon789 colour space is set to ycrbr on the panasonis bluray i set it too auto it readS 4.2.2 YCRBR IS THAT BETTER THAN ycrbr 4.4.4 i was playing aviatar it looked great i can set it to any of this settings ycrbr4.4.4///ycrbr4.2.2// which is better?? i was reading 4.2.2 is and older from of hdmi the reciever does all colourspace but rgb enhanced clippes the whites and black levels:eek:

TomHuffman

06-07-10, 03:28 PM

TomHuffman

06-07-10, 03:35 PM

I have to choose between Gamma 2.2 (or 2.22 -> which one is theoretically better?) and sRGB (close but not similar to x^2.2) tonal response curves.The difference between 2.2 and 2.22 is so small I wouldn't worry about it. For example, for a monitor with 120 nits peak output, the difference between the two options at 10% stimulus will be

2.2: 0.757
2.22: 0.723

This is a negligible difference and probably within the repeatability tolerances of your probe.

I believe that 2.22 is the correct sRGB spec, but many people simply round it off to 2.2.

janos666

06-11-10, 03:23 PM

The difference between 2.2 and 2.22 is so small I wouldn't worry about it.

I know that it won't change anything in the real world. I am just wondering which one is the theoretically accurate value.

I believe that 2.22 is the correct sRGB spec, but many people simply round it off to 2.2.

This is false. sRGB has it's own tonal response curve:
The overall gamma is approximately 2.2, consisting of a linear (gamma 1.0) section near black, and a non-linear section elsewhere involving a 2.4 exponent and a gamma (slope of log output versus log input) changing from 1.0 through about 2.3

It is clearly noticeable when I change my VGA LUTs on-the-fly between 2.2 and sRGB calibrations. But which one should I use...?
Most of the studios should follow the sRGB specifications but gamma 2.2 is a common calibration target. (Some popular calibration softwares do not even support a true sRGB calibration.) I think it is not always clear for the studios. How should I choose then? (I am using 2.2 now.)

TomHuffman

06-11-10, 03:58 PM

This is false. sRGB has it's own tonal response curve.What, exactly, is "false?" That sRGB is 2.22 or that some people round it off? Not clear that your crib from Wikipedia comments directly on either.

Most of the studios should follow the sRGB specifications but gamma 2.2 is a common calibration target.Since sRGB was developed with viewing web pages on computer monitors in mind, it is not clear why you think that "Most of the studios should follow the sRGB specifications . . . ." The studios are much more concerned with home theater displays reproducing Blu-ray and DVD content.

The question as to the proper gamma for this has been hashed over here ad-nauseum. As Poyton points out, there is no universal standard currently followed for gamma, though there should be. He recommends 2.35. However, since telecine operators at the studios have not yet taken his advice, Blu-ray and DVD content is mastered on monitors using no official gamma response that I know of, though 2.2 is often cited.

janos666

06-12-10, 07:15 PM

What, exactly, is "false?" That sRGB is 2.22 or that some people round it off? Not clear that your crib from Wikipedia comments directly on either.

The quoted wikipedia paragraph says that the standard sRGB tonal response curve is not a simple y=x^c function (like gamma 2.22 or 2.2) but it is a complex curve (with linear and exponential sections).
And the difference is clearly noticeable when I change between them.

Since sRGB was developed with viewing web pages on computer monitors in mind, it is not clear why you think that "Most of the studios should follow the sRGB specifications . . . ." The studios are much more concerned with home theater displays reproducing Blu-ray and DVD content.

I mean game studios (PC games) here, not film studios. I know that Blu-Ray contents are made for Rec709 displays and that standard has it's own tonal response curve (which is not a simple y=x^2.2 and it isn't identical with the sRGB curve). sRGB is close to 2.2 but Rec709 is far away from that.
Rec709 is another issue. Somebody is already working on the solution for that problem (Blu-Ray movies on PC monitors...).

2.2 is often cited.

I could make a quick test today. I discovered that Source Engine 2007 v2 (PC games use that, like Half-Life 2) has some console commands to set custom levels and gamma settings for the renderer.
- I set the values to 0-255 and 2.2 and I repeatedly enabled and disabled this switch. Nothing happened.
- I set it to 2.22 and I could see a very slight difference (small, but real changes...)

So, it seems that Valve software uses gamma 2.2 and the difference between 2.2 and 2.22 is very small but not completely irrelevant in real world situations (at least with 8 bit output and rounding...).

I will calibrate my display with x^2.20 response curve in the future. (I used to do it with sRGB curve in the last few months.)

don_

08-06-10, 02:39 PM

how would you try to improve gamma if you could change only those parameters ? (no direct gamma changing option)

RGB drives and cuts (gains/offsets)
all drvs/cuts
sub-brightness
color
contrast
tint

if its possible please give me your thoughs, how would you try to do that for these two cases:
1) black has excellent shadow details, brightness in bright parts is too bright
2) black has poor shadow details, brightness in bright parts is excellent

I have Panasonic PY8 plasma and I used spyder3tv colorimeter, first of all I equaled the level of RGB to D65.
My theory is that for too bright parts I guess I should lower "all drv" and to get more shadow detail I should increase "all cuts" but it seems it affects each other (at least "drives" for greyscale)

best what I could without losing details in black, was moving gamme from 1,98 to 2,11 (but still too bright), I would prefer even more than 2,2, any ideas what should I try ?

my measurements after simple calibration:

http://img839.imageshack.us/img839/6038/gamma.th.jpg (http://img839.imageshack.us/i/gamma.jpg/)

http://img843.imageshack.us/img843/9471/cie.th.jpg (http://img843.imageshack.us/i/cie.jpg/)

http://img541.imageshack.us/img541/3407/luminance.th.jpg (http://img541.imageshack.us/i/luminance.jpg/)

http://img204.imageshack.us/img204/8382/measures.th.jpg (http://img204.imageshack.us/i/measures.jpg/)

MoonRabbit79

08-06-10, 03:41 PM

1) Try to decrease contrast and check if black level is still calibrated, if not then reajust it and go back and forth if needed and try to get a compromise to get a gamma as near as 2.22 for all IRE.

2) Try to increase brightness and check if white level is still calibrated, if not then reajust it and go back and forth and try to get a compromise to get a gamma as near as 2.22 for all IRE.

Your gamma seems pretty "flat" around 2.1 for all IRE, which is good, but what you could do is to "shift" your gamma up to 2.22 by adjusting the green offset and/or gain to get it closer to 2.22, which is equivalent to increase or decrease the contrast around 20 IRE (gain) and 80 IRE (offset). Then you'll have to reajust the blue and red offset/gain to reach D65 for all IRE.

I did that with my 42LG50 because the gamma was not flat (too high from 40 to 90 IRE). But it was easier because I have a 10-points white balance calibration option.

PlasmaPZ80U

08-06-10, 07:04 PM

don_

08-17-10, 04:20 AM

thanks guys for your reply

yes, you are right, using only these limited options its just not possible to set up exactly picture as I want
I tried many settings in last days but finally I gave up :)
I little improved cinema mode + colour calibration, and thats all, picture quality before was good - there is little better, but always could be better :)

1)

Your gamma seems pretty "flat" around 2.1 for all IRE, which is good, but what you could do is to "shift" your gamma up to 2.22 by adjusting the green offset and/or gain to get it closer to 2.22, which is equivalent to increase or decrease the contrast around 20 IRE (gain) and 80 IRE (offset). Then you'll have to reajust the blue and red offset/gain to reach D65 for all IRE.

hmm but wouldn't be this same thing like adjusting to D65 all R-G-B offsets/gains first and then changing "all offset/gains" ?
because if I move G like you said I have to readjust RB anyway after that
but I am not so sure is it really same thing :rolleyes:

romeoavs

09-19-10, 03:34 AM

I notice you list all of the Epson Home Cinema series as having proper CMS except the Epson HC 8100.

I have an Epson Home Cinema 8100 (non-UB model)...does this model qualify?

Considering getting an EyeOne LT (Display2 w/o the software as I understand it) and Chromapure for calibration.

Do I have proper CMS? Will the Display2 on a tripod be okay for my front projection? Is there a trial version of Chromapure 1 or 2?

Thanks

Jleach1

09-23-10, 04:18 PM

thx!

basement

09-28-10, 07:48 PM

Hello, I'm relatively new to calibration and was wondering if someone could suggest what I'm doing wrong. I'm using the i1 with hcfr on a jvc rs10. I'm basically looking to calibrate for greyscale, contrast, brightness, and trying to target gamma for 2.2.

I don't seem to have any trouble with greyscale, but brightness and gamma is a problem. I've set white level to the jvc default which gives me acceptable luminance. Then, following the objective black level setting procedure I display a 10 ire pattern (Tom's posted images) and adjust Y to .65% of Y at 100 IRE. I adjust for greyscale then do a measurement. Greyscale measures fine and luminance tracks the 2.22 curve. However, when I check the brightness setting against a pluge pattern (avshd basic settings), it turns out the brightness setting is set to about six clicks too low. When I view a video, there's a clear loss of detail in dark areas. The picture is clearly better with the brightness set using the subjective pluge pattern.

I then do a new measurement with the new brightness setting, greyscale is still good, but hcfr now averages 1.9 and gamma no longer tracks flat across the scale. It tracks linearly from 10 to 30% then starts to drop off, until at 90% the gamma reading is about 1.1.

So I seem to be in a situation where my settings measure well but the black level seems wrong. When black level is set using the pluge it looks better, but gives me a gamma that tracks to a non-linear lower gamma average.

I'm not sure if it's a factor or not but I'm calibrating against a computer source using kmplayer. I haven't been able to find postings from others with like circumstances.

Thanks in advance. Any help appreciated.

rfbrown94

09-28-10, 08:14 PM

TomHuffman

09-28-10, 10:57 PM

Please re-read the instructions carefully for setting black level at the beginning of this thread. If the gamma is not a standard 2.22, the objective method won't work.

The pluge pattern is the gold standard. Use the objective method only refine it. The AVSHD test pattern is so precise that no refinement is needed.

Hello, I'm relatively new to calibration and was wondering if someone could suggest what I'm doing wrong. I'm using the i1 with hcfr on a jvc rs10. I'm basically looking to calibrate for greyscale, contrast, brightness, and trying to target gamma for 2.2.

I don't seem to have any trouble with greyscale, but brightness and gamma is a problem. I've set white level to the jvc default which gives me acceptable luminance. Then, following the objective black level setting procedure I display a 10 ire pattern (Tom's posted images) and adjust Y to .65% of Y at 100 IRE. I adjust for greyscale then do a measurement. Greyscale measures fine and luminance tracks the 2.22 curve. However, when I check the brightness setting against a pluge pattern (avshd basic settings), it turns out the brightness setting is set to about six clicks too low. When I view a video, there's a clear loss of detail in dark areas. The picture is clearly better with the brightness set using the subjective pluge pattern.

I then do a new measurement with the new brightness setting, greyscale is still good, but hcfr now averages 1.9 and gamma no longer tracks flat across the scale. It tracks linearly from 10 to 30% then starts to drop off, until at 90% the gamma reading is about 1.1.

So I seem to be in a situation where my settings measure well but the black level seems wrong. When black level is set using the pluge it looks better, but gives me a gamma that tracks to a non-linear lower gamma average.

I'm not sure if it's a factor or not but I'm calibrating against a computer source using kmplayer. I haven't been able to find postings from others with like circumstances.

Thanks in advance. Any help appreciated.

millerwill

09-28-10, 11:44 PM

bodosom

09-29-10, 08:25 AM

Tom, When setting Contrast ( i.e., white level) with the S&M disk, do you set it to resolve the white bars all the way up to 255, or to you set it (higher) to resolve them only up to 235?

Assuming you can get your display to distinguish it's a personal choice. The only people I know of that have looked closely at this found no significant deliberate content above 235 (they looked at discs and DI sources). All of it could be explained by poor scaling (i.e. it was all ringing). There is a reason to go to 254 (which I forget so how important can it be) and a reason to stop at 235 (better contrast ratio).

TomHuffman

09-29-10, 10:45 AM

millerwill

09-29-10, 11:02 AM

The amount of program material this high is infinitesimally small, if it exists at all. I would not sacrifice a significant amount of peak output to maintain performance in an area that has little bearing on image quality.

OK, thanks to you and bodosom for the replies. What you say is what I had thought was preferable, but S&M on the guide for their disc says to set it to resolve the bars up to 255, which led to my confusion and question.

bodosom

09-29-10, 11:45 AM

... S&M on the guide for their disc says to set it to resolve the bars up to 255 ...

While I would never speak for Stacey or Don I suspect they're making some assumptions. If you're using a "typical" display (LCD) then you probably have the headroom to waste a bit of peak output. Some folks (PJ/plasma owners) need to be a bit more parsimonious with their limited output budget.

basement

09-29-10, 07:15 PM

Set black level by eye using the pluge pattern. Are you re-measuring your 100% white window after changing the brightness setting with the pluge pattern? Set your settings, take a 100% reading, then measure 0-90%. If that doesn't help, make sure your contrast isn't set too high. You might be clipping white.

I don't know about this "kmplayer," but computer sources are generally crap. Not saying yours is, but it could be the problem.

Ok. From your responses, and Tom's, I think I see the problem with my process. I should adjust using pluge then re-measure 100% and verify at 10%.

If after doing this, and it still does not track to 2.2, then the display gamma is not set correctly, or my contrast is too high, clipping luminance at higher ire.

TomHuffman

09-29-10, 09:01 PM

If after doing this, and it still does not track to 2.2, then the display gamma is not set correctly, or my contrast is too high, clipping luminance at higher ire.Or your display's gamma is not a natural 2.22 and cannot be fixed. This is not uncommon.

basement

10-05-10, 07:14 PM

Or your display's gamma is not a natural 2.22 and cannot be fixed. This is not uncommon.

I spent a bit more time on this. Taking a cue from comments about computer software video players, I tried calibrating the display using different software players and found that luminance does not track very well. Prior to the software players, I calibrated a Panasonic bluray player and easily had it adjusted properly for greyscale and with luminance following the 2.22 reference.

It turns out kmplayer was the best of the three I had. It tracks fairly well between 20 and 60, then from 70 to 90, luminance drops off - looking like contrast was set too high. I lowered contrast and got the same relative result. I played around a little bit with video card software settings rather than player settings. With no knowledge on what the settings do, gamma response can be adjusted but the results were not satisfactory. On a side note, even with changes in gamma, greyscale was always good, so colour was satisfactory. The htpc forum may have some answers.

Anyhow, thanks to all for your help.

delphiplasma

10-17-10, 05:35 AM

Hi Tom,

I notice in your introduction, regarding using blue filter for Plasma displays can result in a wholly inaccurate colour level, when using SMPTE colour bars. I've noticed this, as the SMPTE bars are not 'widows', it causes a deviation. The only method one can use, for a plasma display, is the 'windows' colour patterns, using 'blue only' mode.

There are no colour patterns for a plasma display that will allow the use of a blue filter, not that I know of? Unless someone knows better?

Regards

Delphi

TomHuffman

10-17-10, 02:02 PM

I don't know if it is all plasmas. I just noticed it on a couple of plasmas. I also don't know if the CAUSE of the problem is the need for window patterns. It may or may not be. However, I am skeptical about this explanation because it would apply equally to CRTs and I have not noticed the problem there.

In any case, I would always recommend direct measurement instead of filters and color bars. Even inexpensive colorimeters are reasonably accurate reading luminance, which is all that is necessary to set Color. Tint requires an accurate reading of hue, which cheap colorimeters are less adept at. Still, even here, I would expect the tolerances to be well within reason.

I notice in your introduction, regarding using blue filter for Plasma displays can result in a wholly inaccurate colour level, when using SMPTE colour bars. I've noticed this, as the SMPTE bars are not 'widows', it causes a deviation. The only method one can use, for a plasma display, is the 'windows' colour patterns, using 'blue only' mode.

There are no colour patterns for a plasma display that will allow the use of a blue filter, not that I know of? Unless someone knows better?

delphiplasma

10-18-10, 06:43 AM

Hi,

I come to this conclusion after measurements on Plasma displays with 'Blue only mode'.

The results are very different when choosing the size of the pattern, this would indicate to me that the results would be the same for 'blue filters'. I assume that this is the draw in current due to the 'White level'? However, you are quite right when you speak of CRT measurements not seemingly being affected by the size of the pattern?

Of course, as you say, direct measurement is the best way to do the measurement, although, as the direct measurement isn't 100%, I find it best to go through all the methods to get the best balance?

Not sure if you would agree?

Thanks

Delphi

Joelc

10-18-10, 09:59 AM

Why is this thread no longer a sticky?

KramerTC

10-18-10, 06:08 PM

Why is this thread no longer a sticky?

I hope this thread being unstickied has nothing to do w/AVS becoming a reseller of Calman.

Tom Huffman has been helping fellow AVSers for a long time and I thought his thread being stickied was a nice recognition of that.

buzzard767

10-18-10, 06:16 PM

It's now included in the stickies under "Links to this forums popular threads" making it and the others, ummmm, somewhat sticky.

mechman25

11-06-10, 04:34 PM

Tom

At the start of this thread for the brightness adjustment you stated to adjust the luminance of a 10% white test pattern to .6% of the luminance of a 100% white pattern.

So that is multiply by .006?? Correct??

TomHuffman

11-06-10, 05:03 PM

At the start of this thread for the brightness adjustment you stated to adjust the luminance of a 10% white test pattern to .6% of the luminance of a 100% white pattern.

So that is multiply by .006?? Correct??Yes, but I also wrote:
Since I originally wrote this, the AVSHD disc has released a very advanced version of the pluge pattern that allows the user to set black level with perfect accuracy by eye alone. If you have access to this disc, then use it. Because it is so precise, it makes using objective measurements to set brightness unnecessary.

spongebob

11-06-10, 05:18 PM

Yes, but I also wrote:
Since I originally wrote this, the AVSHD disc has released a very advanced version of the pluge pattern that allows the user to set black level with perfect accuracy by eye alone. If you have access to this disc, then use it. Because it is so precise, it makes using objective measurements to set brightness unnecessary.

Which pattern, Tom and where on the disk?

thx

bob

PlasmaPZ80U

11-06-10, 09:21 PM

Which pattern, Tom and where on the disk?

thx

bob

first basic pattern, "Black Clipping Pattern" AVSHD709

spongebob

11-07-10, 09:32 AM

first basic pattern, "Black Clipping Pattern" AVSHD709

That has never worked on my Elite Pro-101FD. Even cranking brightness all the way up just washes out the picture.

What about the near black windows? Which of those should be visible? I usually have to go to 4% or so.

thx

bob

kjgarrison

11-07-10, 12:09 PM

From the first post:

Of these definitions, only the xy coordinates for the primary colors and white point are absolute. The secondary colors and luminance values are DERIVED from the primaries and the white point. If your primary colors measure according to these standards, then this list correctly states the proper specifications for brightness and secondary hue/saturation. However, with a different set of primaries, you would want to shoot for a slightly different set of brightness and secondary hue/saturation targets. The math required to figure out these relationships is too complicated to go into here, but good calibration software should take all of this into account.

Tom can you elaborate a bit on the part I bolded above? Many of us, if not most, have "different set of primaries". I have struggled to know what the "different set of brightness and secondary hue/saturation targets" to aim for.

I have CalMAN and ColorHCFR and don't really see how either of these gives me the "different" targets, (unless you're suggesting to randomly change things looking for lower dE values?)

Thank you.

mechman25

11-07-10, 01:43 PM

Thanks Tom for the reply

Yes I have this disc, even downloaded the version 2 disc.
This pattern was very easy to use.

I recently bought a colorometer and have done some calibrations on my Plasma and LCD.

Greyscale coming out fine on both.
LCD coming out fine but the color gaumet on the plasma is way off on the green, so I am just trying to figure out if I am doing something wrong.

I am going to try some more calibrations on something other then the custom mode and warm2 settings I was using.

TomHuffman

11-07-10, 02:44 PM

However, with a different set of primaries, you would want to shoot for a slightly different set of brightness and secondary hue/saturation targets. The math required to figure out these relationships is too complicated to go into here, but good calibration software should take all of this into account.

Tom can you elaborate a bit on the part I bolded above? Many of us, if not most, have "different set of primaries". I have struggled to know what the "different set of brightness and secondary hue/saturation targets" to aim for.

I have CalMAN and ColorHCFR and don't really see how either of these gives me the "different" targets, (unless you're suggesting to randomly change things looking for lower dE values?)ColorHCFR doesn't offer this ability. I don't know if CalMan does or not. ChromaPure does.

It is certainly not random. The same math that yields the secondary and brightness targets for the standard gamuts results in different targets for non-standard gamuts.

TomHuffman

11-07-10, 02:46 PM

LCD coming out fine but the color gaumet on the plasma is way off on the green, so I am just trying to figure out if I am doing something wrong.This is a very common reaction. The simple fact is that many commercial displays are just not designed with accuracy in mind.

mechman25

11-07-10, 02:52 PM

Once I do a couple more calibrations I will post results.

Can there be hardware problems that cause one color to go way off?

TomHuffman

11-07-10, 03:32 PM

Once I do a couple more calibrations I will post results.

Can there be hardware problems that cause one color to go way off?Sure, but the most likely hardware problem is the design of the display.

bodosom

11-07-10, 03:52 PM

That has never worked on my Elite Pro-101FD. Even cranking brightness all the way up just washes out the picture.

9G Kuro monitors clip BTB and WTW in CE space unless you enable DRE. The easiest way to verify this is to switch to IT space (RGB:0-255) with the input in CE space. You'll see all the bars flashing unless clipping is happening elsewhere. Of course if the input signal is YCC the colors will be wrong but the bars will still be visible.

marc55

11-14-10, 09:44 AM

For adjusting contrast only, would just the light meter and a test pattern work?

buzzard767

11-14-10, 10:13 AM

For adjusting contrast only, would just the light meter and a test pattern work?

All you need is a test pattern and your eyes.

LordOfThundera

11-23-10, 05:17 PM

Hey Tom,

Just started reading this thread from the start a few hours ago and holy cow that's a lot of serious technical information right there. I don't presume to understand half of it yet, but hopefully with some practical application it will become clearer.

Thing is, I don't have a colorimeter of any kind. Do you know of anything under £100 (not sure what that is in dollars, somewhere around 130 maybe?) that is anywhere near accurate enough to do a half decent job? I was originally looking at the SpyderTV but heard that they can be poorly calibrated out of the box and I don't know anyone with more accurate equipment that I can use to calibrate it. Not that I know yet what any of that really means.

I recently bought a Samsung UE32C4000 LED Edge-Lit LCD, and I've been having a few issues getting a decent picture using just DVE and some White Balance settings I found on the internet. From what I've read on here am I right in thinking that those controls (RGB offset and RGB Gain, 6 controls total) would be changing the colour saturation but also the colour intensity, thus adversely affecting the general picture quality even if the saturation is correct? So, unless I'm willing to go into the service menus (which may not even contain full colour controls anyway and will void my warranty I think) would you advise leaving White Balance settings on factory defaults?

Apologies for all the questions, I hope it's not presumptuous or pushy of me (especially as I have another thread vaguely concerning this in a different part of the forum but nobody has responded to it as far as I'm aware). I'm just a little frustrated, especially as I'm very satisfied with the picture on my tiny little Samsung LE19R36, but also very interested in learning how to better calibrate a display generally.

Thanks in advance for any advice, and sorry in advance if I am indeed being too pushy with my line of questioning.

Paul.

TomHuffman

11-23-10, 08:31 PM

http://www.amazon.co.uk/Pantone-EODLT-Eye-One-Display/dp/B000CR78CE/ref=sr_1_2?ie=UTF8&s=software&qid=1290565850&sr=8-2

Hey Tom,

Just started reading this thread from the start a few hours ago and holy cow that's a lot of serious technical information right there. I don't presume to understand half of it yet, but hopefully with some practical application it will become clearer.

Thing is, I don't have a colorimeter of any kind. Do you know of anything under £100 (not sure what that is in dollars, somewhere around 130 maybe?) that is anywhere near accurate enough to do a half decent job? I was originally looking at the SpyderTV but heard that they can be poorly calibrated out of the box and I don't know anyone with more accurate equipment that I can use to calibrate it. Not that I know yet what any of that really means.

I recently bought a Samsung UE32C4000 LED Edge-Lit LCD, and I've been having a few issues getting a decent picture using just DVE and some White Balance settings I found on the internet. From what I've read on here am I right in thinking that those controls (RGB offset and RGB Gain, 6 controls total) would be changing the colour saturation but also the colour intensity, thus adversely affecting the general picture quality even if the saturation is correct? So, unless I'm willing to go into the service menus (which may not even contain full colour controls anyway and will void my warranty I think) would you advise leaving White Balance settings on factory defaults?

Apologies for all the questions, I hope it's not presumptuous or pushy of me (especially as I have another thread vaguely concerning this in a different part of the forum but nobody has responded to it as far as I'm aware). I'm just a little frustrated, especially as I'm very satisfied with the picture on my tiny little Samsung LE19R36, but also very interested in learning how to better calibrate a display generally.

Thanks in advance for any advice, and sorry in advance if I am indeed being too pushy with my line of questioning.

Paul.

dpnaylor

11-24-10, 06:55 AM

Been a while since I have done this but let me see if I can shed a little light on the subject for you as simple as possible.. Whatever you do without a meter is going to be pointless and most likely make things worse for you. Especially if you start getting into the Offsets and Gains...

RGB Offset adjusts red green and blue individually from medium gray down to black.

RGB Gain adjusts the red green and blue individually from a medium gray up to white...

a perfect gray scale from black to white will have no tint of color to it at all. for example to much Red gain and your whites will start looking pinkish.

The eye one display LT that Tom suggested is the one and only meter to get if you don't want to spend a lot of money. It will give you respectable results and hours of fun learning. My only question about the LT is, Will it work with a new LED TV? Tom can probably answer that for sure. I guess it depends on the software you are using if it has a look up chart for that particular meter..

Try looking up your TV model number here and see if anybody has some calibrated settings you could try. Every unit is different but it may get you closer than you are getting by eye.

TomHuffman

11-24-10, 12:01 PM

My only question about the LT is, Will it work with a new LED TV?There's nothing special about LED backlighting that poses special problems for colorimeters.

LordOfThundera

11-24-10, 02:40 PM

Hey guys,

I've decided to take it back and get the LE32C530 instead, as it seems to be a pretty decent set from what I've seen on here (and I'm on my second set with build quality issues in a row, missing screws etc.). Either of you guys have experience with it? Even though it's only about £20 cheaper I'd rather have a set that's been tried and tested (knowledge about the UE32C4000 seems practically non-existent). Unless you guys would recommend something else in that price range of course (sorry, wrong thread for that I know, but nobody over in the recommendations thread seems interested).

I'll definitely be picking up the Eye One though, and working through the tutorial on here.

angryht

12-17-10, 03:55 PM

rfbrown94

12-17-10, 07:43 PM

Isn't the insignia a 720p display?

PlasmaPZ80U

12-17-10, 08:51 PM

I see that the Samsung PNxxB650 is listed, within the first post, as a plasma with CMS. I have the Insignia NS-P501Q-10A, which is supposed to be the same as the Samsung PNxxB650. I have not dug too far into the service menu, but I was wondering if anyone knows if there is a CMS within the service menu.

I don't think it is the same as the Samsung B650. Anyway, the CMS in the Samsung is only in the user menu (advanced picture settings, custom color space).

703

12-17-10, 09:53 PM

http://www.spectracal.com/documents/CalMAN%20Doc-V2010-12-06.pdf

Is a very good document on what controls to adjust and in what direction to fit the CIE charts! Note Page 26.

angryht

12-17-10, 10:44 PM

I don't think it is the same as the Samsung B650. Anyway, the CMS in the Samsung is only in the user menu (advanced picture settings, custom color space).

You're right. I guess it's a PN50B430P2D. Never mind. I saw that the service menu said: "adjustments and controls are identical to Samsung PNxxB650" That must just be for the grayscale.

thomasl

01-02-11, 06:07 PM

kjgarrison

01-02-11, 10:24 PM

Tom et al, do you have any experience with the CMS implementations in some of the newer LG or Sharp flat panels? I've been looking at their manuals seeing what calibration capabilities that they had since I may be in the market for a new flat panel. I'm wondering if their CMS implementations are as good as the Samsung's.

--tom

Some of the reviews on the high end LG models seem to show that in THX mode at least they are already pretty spot-on.

Knd

01-03-11, 01:35 PM

Tom et al, do you have any experience with the CMS implementations in some of the newer LG or Sharp flat panels? I've been looking at their manuals seeing what calibration capabilities that they had since I may be in the market for a new flat panel. I'm wondering if their CMS implementations are as good as the Samsung's.

--tom

We own a LG that I bought back in June 2010 and we also bought a LG for our son about a month earlier so I spent some time calibrating them. See here for what I discovered about the CMS:
http://www.avsforum.com/avs-vb/showthread.php?t=1261013&highlight=

Nutdotnet

01-27-11, 01:46 PM

Hey Tom. I apologize in advance as I am sure this has been answered in this thread. But my search turned up squat.

On your calibration DVD, what size are the windows? I am in need of windows smaller than AVS/Getgray/DVE. AVS does have the small APL, but that actually makes voltage clamp down in the mid-range.

TomHuffman

01-27-11, 07:14 PM

poiuiop

01-28-11, 08:26 AM

Good info

pbc

01-30-11, 08:33 AM

Setting White Level (Contrast)

The Contrast control determines the peak output your display will provide. Set too low you lose image punch and lower contrast ratios. Set this too high and you lose color accuracy and detail in bright scenes and you may suffer from eye strain.

The standard method for setting Contrast requires that you look at a test pattern that has a just-below-white stripe against a white background. You are supposed to set Contrast as high as you can without losing the ability to distinguish the just-below-white stripe from full white. I included such a pattern on the calibration DVD.

However, there are a couple of problems with this method.

Some displays, especially LCDs, will never suffer from loss of high level detail even with Contrast set to 100%. This method will recommend a setting that is much too high.
This method does not take into consideration color performance. Many displays will lose their ability to track a neutral white at high output levels with Contrast set as high as this method recommends.

Thus, I think that a better method for setting Contrast is to just set it at a level consistent with good color performance and reasonable light output for a given display device. What's a reasonable level?

CRT tubes: 30-40 fL
Plasma: 30-40 fL
LCD flat panel: 30-50 fL
Digital rear projection: 30-40 fL
Digital front projection: 12-16 fL

Tom, most/all LCD's and many plasmas also have Cell Light or Back light controls embedded. How do these factor into the setting of White Level as they also impact the measured fL. I.e., one can likely get to a fL of 40fL on an LCD with a contrast setting of say 70 and a backlight of 20, and at the same time a contrast setting of 100 and a backlight setting of under 10.

Would one simply pump contrast up as high as possible (keeping in mind the WTW pattern) and then drop back light until you reach the desired fL?

TomHuffman

01-30-11, 12:19 PM

Tom, most/all LCD's and many plasmas also have Cell Light or Back light controls embedded. How do these factor into the setting of White Level as they also impact the measured fL. I.e., one can likely get to a fL of 40fL on an LCD with a contrast setting of say 70 and a backlight of 20, and at the same time a contrast setting of 100 and a backlight setting of under 10.

Would one simply pump contrast up as high as possible (keeping in mind the WTW pattern) and then drop back light until you reach the desired fL?They work together. The biggest difference is that raising the backlight will also raise the black level. The rule of thumb is to use the lowest backlighting you can to get adequate light output.

TheWolfe

01-30-11, 10:20 PM

Thanks for the post. I'm sure it will come in handy. Wish I had seen something like this sooner.

bishopt

03-31-11, 01:32 PM

So I'm about to jump in and calibrate my first projector, well actually my first of anything. So I purchased an eyeone LT and I am planning on using the HCFR software. The projector that I will be playing with is a Mits HC4000 (which I believe to be very close to the HC3800), so I need to do some more reading using your original post and planning to use that in combination with the claibration for dummies guide with the Blu Ray DVE disc..So what I was wondering is my goals is to try to have some presets for watching movies, HDTV(ambient light) and maybe a sports setting. Any suggestions or recomendations on best approach to achieve this, should I use a cooler temp like 7500 for Sports which would probably allow me to be a little brighter? Any thoughts/guidance would be appreciated...Thanks.

dpnaylor

03-31-11, 02:37 PM

I have an older Mits the HD1000U. That has 3 presets to it. I calibrated one of them then all I did was adjust the brightness, contrast and brilliant color for the last two presets.. Bumping each of those up a bit for the second preset then a bit more for the last preset. preset two for gaming really then preset three had the highest settings for sports. I found I could bump the brightness almost all the way up for sports. The Contrast you cant go to crazy with or the Whites really get washed out.

bishopt

03-31-11, 03:31 PM

dpnaylor

03-31-11, 03:39 PM

bumping the brightness and contrast some wont mess up the grey scale to much once it is already calibrated. Some projects have presets that bump the green levels to try and get more lumens but my projector is pretty bright already. I have not really had the need to go back in and adjust the grey scale again.. My projector only has grey scale adjustments anyway. I cant adjust my primaries at all..

tjacks84

04-11-11, 10:19 AM

Hi Tom,

It appears I'll be joining the ranks of your customers with ChromaPure. I am purchasing a meter and software package for a new TV I am to receive this week. This brings me to my question. I noticed at the beginning of the thread some select Samsung plasmas were listed with a full CMS. I am taking delivery of their new PN64D7000 plasma and I cannot tell whether or not it has a full 3D CMS; however, I suspect it does not. I know compared to a lot TV's it has excellent PQ controls, which was one of the reasons I chose it. For example, I know it has 10 pt white balance for greyscale, and as best I can tell from the manual it has a great deal of control over color, as well. It appears that control is over the primaries however. If it turns out the CMS is a good one, but not a full 3D CMS, what video processors would you recommend to provide such, or do you think the Samsung plasmas are close enough to avoid an independent video processor (from a value perspective)? If the video processor would result in a noticeable PQ improvement and full 3D CMS for another $1k-$2k, then I think it would be worth it. I know it's a subjective question, but I'd appreciate your opinion along with anyone else. I intend to calibrate this TV and very much want to get as close as I can to reference.

Thanks,

Tjax

buzzard767

04-11-11, 10:32 AM

It has Gamma, 2 pt., 10 pt., and full 3D CMS. The whole shebang!

VP = DVDO iScan Duo especially now that we have a new controller thanks to a forum member. http://www.avsforum.com/avs-vb/showpost.php?p=20263758&postcount=28

tjacks84

04-11-11, 10:48 AM

Thanks for the fast response, Buzz. Long time no talk, huh? So your take is that Samsung has the whole Shebang? Glad to hear it. I had held off from pulling the trigger on CurtPalme due to their total package deals. I figured it best to sort out whether I'd need a video processor to get total control over calibration, before I bought the equipment to calibrate. After all, great equipment won't matter if I can't control the picture. You've been a lot of help in my decision process. Just when I think I'm done reading & thinking, I find there's even more. Thanks for taking the time.

TomHuffman

04-11-11, 01:45 PM

I noticed at the beginning of the thread some select Samsung plasmas were listed with a full CMS. I am taking delivery of their new PN64D7000 plasma and I cannot tell whether or not it has a full 3D CMS; however, I suspect it does not.It does. Any Samsung display that has a Custom Color Space option has a full 3D CMS.

jmoakk

05-19-11, 08:31 PM

dominickwok

05-20-11, 07:07 AM

I adjusted the greyscale now I'm attempting to adjust the color using CMS on an epson 8350. I have Hue, Saturation, and Brightness controls for each color. Adjusting the Hue affects both x,y. Adjusting Saturation also affects both x,y. Due to that I can come up with the correct x,y value with a number of different Hue & Saturations settings. How do I know which setting is correct? Which one should i adjust first?

I'm by no way an expert but I had invested quite a lot of time to adjust the CMS of my Panny 4000 to achieve fairly good result.

So, when it comes to the basic concept, adjusting Saturation is equivalent to adjusting its distance from the D65 white point; and adjusting Hue is equivalent to adjusting its swinging angle with respect to the D65 white point. I assume you know about this part (don't you?).

Therefore, THEORETICALLY speaking, for a particular primary/secondary color, there is only one Saturation/Hue/Brightness combination that can bring the color back to the correct xyY. To put it into another aspect - take 100% saturated Yellow as an example, it should be measured as 107% Red, 107% Green, and 0% Blue for the perfect alignment in the correct xy position (just ignore the Brightness (Y) part for the sake of simplicity). If your measurement comes out as 104% Red, 108% Green, and 5% Blue, that means it is both undersaturated (because it contains a little bit of Blue) and deviated towards Green a little bit (Unbalanced Red & Green). Adjusting Hue alone may allow you to get Red & Green balanced at 106% / 106%, but it is still undersaturated. So you need to increase Saturation to get the perfect RGB combination (i.e. xy coordinate) of 107% / 107% / 0%. You get it?

However, poorly implemented CMS will usually cause you much time to settle down at a correct setting because, e.g. when you adjust one aspect (e.g. Saturation) it also affects the other aspects (e.g. Hue and/or Brightness) unexpectedly.

Hope that helps.

TomHuffman

05-20-11, 09:22 AM

TomHuffman

05-20-11, 09:25 AM

So you need to increase Saturation to get the perfect RGB combination (i.e. xy coordinate) of 107% / 107% / 0%. You get it?This will provide excessive luminance. 100%, 100%, 0% is correct. RGB percentages are useful only for a human interface that provides RGB controls (Lumagen Radiance and Samsung CMSs come to mind). The Epsons have an HSL interface.

jmoakk

05-20-11, 05:01 PM

Where do I get the correct color % combination for each color? Like 100%, 100%, 0 for yellow? Also I didn't know I could use the cie chart in live reading mode when displaying one color windows. That alone helped me a lot last night when I did another calibration. It let me see which way the color was moving when adjusting hue and sat.

Kilgore

05-20-11, 05:04 PM

jmoakk

05-20-11, 05:07 PM

hcfr, I know where the reading is in the software. I just dont know what the reading should be for each color.

dominickwok

05-20-11, 10:58 PM

hcfr, I know where the reading is in the software. I just dont know what the reading should be for each color.

You can refer to the following post:

http://www.avsforum.com/avs-vb/showthread.php?p=16166537#post16166537

In that post, it points you to the following link to download an Excel spreadsheet:

http://www.avsforum.com/avs-vb/showthread.php?t=1111066

The "Calibration Aid" tab of the spreadsheet lists out the correct RGB% for each 100% saturated primary/secondary (as well as for each 75% saturated primary/secondary). Please note that the RGB% (e.g. 107% / 107% / 0% for 100% sat. yellow) is the % shown by HCFR - that's nothing to do with anything else such as external video processor).

Wizziwig

05-25-11, 12:25 PM

Hi Tom,

Excellent guide. Any chance you could update it for Lumagen Radiance owners?

I'm mostly confused as to why the Radiance offers both a "Color Gamut" and "Color Decoder" setting under it's CMS system. Can't you adjust everything using just the Color Gamut controls? Why would you ever need to mess with the color/hue controls under the ColorDecoder menu?

Also, any general tips for calibrating the JVC RS40? It's difficult to decide what to adjust on the projector vs. on the Radiance.

Thanks.

TomHuffman

05-25-11, 12:44 PM

The color decoder controls are legacy features that the CMS renders obsolete.

Use the PJ's controls to get the gamma/grayscale close, and then refine with the Lumagen's controls.

Excellent guide. Any chance you could update it for Lumagen Radiance owners?

I'm mostly confused as to why the Radiance offers both a "Color Gamut" and "Color Decoder" setting under it's CMS system. Can't you adjust everything using just the Color Gamut controls? Why would you ever need to mess with the color/hue controls under the ColorDecoder menu?

Also, any general tips for calibrating the JVC RS40? It's difficult to decide what to adjust on the projector vs. on the Radiance.

ericlhyman

05-27-11, 03:50 PM

zoyd

check out the front projector forum - Hugh2/hughman did that very study of changes to color control impact on the CIELCH measures (same dE as CIELUV) of his JVC - but it is rumoured the firmware version may behave slightly different. I would say it is generally the case that you have to shift color brightness significantly before color saturation moves with video color controls - he posted several sets of numbers on that starting here....

http://www.avsforum.com/avs-vb/showthread.php?p=10426396&&#post10426396

I have observed similar with Sony Ruby/Pearl - the dL hit to move in the primaries with the color control per spec resulted in much worse dE than leaving primaries be and only shorting dL a bit under spec.

Of course Tom's post was how to use a CMS and video decoder that is capable of getting you the perfect result - as his other posts on color demonstrate - sometimes this can be an exercise in futility in picking the worst vs. best compromises.

The only thing I would add is to repeat the measures with 100% patterns - as sometimes displays can have settings impacting bright colors like DLP's BrilliantColor(tm) or any dynamic iris - which maybe can be turned off - but if not - compromise between light/dark colors time!

Also if doing tint with this method of sensoring rather than filtering - I would do it after greyscale not before - because secondaries shift with greyscale - with filters this is irrelevant because you have bluescale not greyscale!

Using red brightness to set color levels will only work if your set has no red push - if it does then the color will probably be too low overall - so again measure blue/green levels to find the compromise in between.

I also use 100% white for greyscale to first establish my clipping limits - I find hot white colortemp blowouts much more annoying than the tweak to contrast they give.

Did you change your settings for the Ruby in the service menu or regular user menu? Do you have any instructions on how to use the Ruby's color management system in the user menu? I find it difficult to use.

ericlhyman

05-27-11, 04:17 PM

dpnaylor

09-08-11, 02:45 PM

Greetings all.

I have been doing calibration now as a hobby on my own TV's as well as a few families TV's for a few years and have never really run into something I couldn't figure out.. It has been a while since I have done this and I ran into a question on a friends LG Plasma.
This LG Plasma is about 7yrs old. When I go into the color settings for the unit it has a USER option. Selecting the user option then brings up a RGB screen with a slider for each R, G, B, from minus 30 to plus 30 I believe for each color. Now normally there are two sets, one for Contrast and the other for Brightness to be able to set the grey scale properly but this set only has one. I know most computer monitors only have one set of sliders. So would I put up a 100% white pattern on the screen set my ref in Calman to D6500 and adjust the sliders till I am as close as I can get, probably starting with a warm setting?
I am trying to get the model number from the TV so I have a bit more info. I might be able to get into a service menu to give me some more options on this thing. I am using an eye-one pro meter with the Calman software.

Thanks guys! Any help appreciated!

serialmike

09-08-11, 03:43 PM

Originally Posted by TomHuffman

Thus, I think that a better method for setting Contrast is to just set it at a level consistent with good color performance and reasonable light output for a given display device. What's a reasonable level?

CRT tubes: 35 fL
Plasma: 35 fL
LCD flat panel: 50 fL
Digital rear projection: 35 fL
Digital front projection: 15 fL

If you use HCFR, you can determine fL by multiplying the peak output (100% Y) by 0.29

--------------------------------------
What would the recommendation be for an SXRD (LCOS) rear projection set?

I dont believe this chart is correct. There is no reason whatsoever not to calibrate an LCD to 35ftl also. None the black level on bad lcd sets is .015 at worst generally.

35 is the standard and all sets should be calibrated to this. This is one reason why people dont liek the look of LCD. They are being calibrated incorrectly.

This made sense when black levels of lcd were .05 and higher in order to get some kind of contrast. It is no longer needed.

TomHuffman

09-08-11, 04:27 PM

I agree. This is out of date.

Edit: In fact, this was edited some time ago to 30-50fL. I just edited it again to 30-40fL.

I dont believe this chart is correct. There is no reason whatsoever not to calibrate an LCD to 35ftl also. None the black level on bad lcd sets is .015 at worst generally.

35 is the standard and all sets should be calibrated to this. This is one reason why people dont liek the look of LCD. They are being calibrated incorrectly.

This made sense when black levels of lcd were .05 and higher in order to get some kind of contrast. It is no longer needed.

serialmike

09-08-11, 05:05 PM

I agree. This is out of date.

Edit: In fact, this was edited some time ago to 30-50fL. I just edited it again to 30-40fL.

Good deal :) When I saw you posted I got worried I drew your ire. So often when I post something against what a sticky says or goes against the grain I get into.

Whew :)

SecondaryColors

02-18-12, 07:54 PM

1) Setting Contrast
Why is 30 - 40 ftL the recommended LCD contrast setting? If movie theaters use 12 - 14 ftL settings, then I want to try those settings to see if they'll replicate the movie theater experience on the LCD screen. But if the 30-40 ftL are the standards used by post production people, then I'll give it a shot because accuracy is important.

2) Setting greyscale:
If my television is missing greyscale controls like RGB Cuts or Gains, then what can be done? Can I influence greyscale with other controls -- such as color temperature settings, dynamic range standard / out of standard settings, black expansion for shadow detail settings, contrast control, and C.M.S settings?

PlasmaPZ80U

02-18-12, 07:58 PM

dominickwok

02-19-12, 05:31 AM

2) Setting greyscale:
If my television is missing greyscale controls like RGB Cuts or Gains, then what can be done? Can I influence greyscale with other controls -- such as color temperature settings, dynamic range standard / out of standard settings, black expansion for shadow detail settings, contrast control, and C.M.S settings?

If your set has any setting related to color temperature, try the setting of warm / warm1 / warm2. Usually this sets it closer to D65 than any setting of normal / standard / cool, etc.

rickardl

02-20-12, 01:02 AM

35 is the standard and all sets should be calibrated to this. This is one reason why people dont liek the look of LCD. They are being calibrated incorrectly.

Isn't 35 only for dimly lit/dark rooms?
I would love to have some kind of equation/calculation which takes into account the ambient lighting of the room when coming up with the target contrast ftL.
Now, it seems 35 ftL is just the lower end of the scale which probably is too dark for most people's viewing conditions.

TomHuffman

02-20-12, 02:28 AM

Isn't 35 only for dimly lit/dark rooms?No. 35 ft-L is the recommended peak luminance of color monitors under ordinary viewing conditions. See SMPTE RP 166.

rickardl

02-20-12, 03:24 AM

No. 35 ft-L is the recommended peak luminance of color monitors under ordinary viewing conditions. See SMPTE RP 166.
I don't have RP 166, but does the doc specify what "ordinary viewing conditions" are?

den110

02-20-12, 12:20 PM

airscapes

02-20-12, 12:55 PM

Hi everyone

I have a quick question regarding calibration. How often should you recalibrate your projector? I have an Epson 8700UB. I calibrated it when it had 0 hrs on the bulb using Eye One Lt, AVSHD disc and HCFR software. It now has 500 hrs. Should I do a recalibration?

Yes, typically if you are going to pay someone to do the calibration you would wait at least 200 hours to allow the lamp to settle in. I now have my own equipment and plan to check my own projector every 500 hours (6 months)

TomHuffman

02-20-12, 02:09 PM

I don't have RP 166, but does the doc specify what "ordinary viewing conditions" are?

All light sources in use during picture assessment or adjustment should approximate D65.
The ambient light reflected from the screen should be the lowest possible level.
Reflections on the monitor screen should not cause a perceptible impairment from the normal viewing position.
Illumination at the viewing area should be between 20-100 lux.

Other than warning against direct light reflecting off the screen surface, there is no requirement for "dimly lit/dark rooms".

gregr

02-20-12, 04:04 PM

rickardl

02-21-12, 09:28 AM

Thanks!

All light sources in use during picture assessment or adjustment should approximate D65.
The ambient light reflected from the screen should be the lowest possible level.
Reflections on the monitor screen should not cause a perceptible impairment from the normal viewing position.
Illumination at the viewing area should be between 20-100 lux.
Other than warning against direct light reflecting off the screen surface, there is no requirement for "dimly lit/dark rooms".
The most important factor (other than minimizing ambient reflected light) related to the brightness of a direct view monitor is the brightness of the surround - i.e. the lighted area around the screen that is visible to the viewer. It should be D65 and 10% of the screen brightness, i.e. 3.5 fL. The brightness of the surround changes the optimum gamma and optimum brightness of the monitor. Those are the specified conditions for telecine rooms, so you would ideally duplicate those conditions in your direct-view monitor viewing room.

SecondaryColors

02-21-12, 11:53 PM

...

tbaudoin

03-05-12, 06:00 PM

The most important factor (other than minimizing ambient reflected light) related to the brightness of a direct view monitor is the brightness of the surround - i.e. the lighted area around the screen that is visible to the viewer. It should be D65 and 10% of the screen brightness, i.e. 3.5 fL. The brightness of the surround changes the optimum gamma and optimum brightness of the monitor. Those are the specified conditions for telecine rooms, so you would ideally duplicate those conditions in your direct-view monitor viewing room.

So, what's the preferred method, measure the fL avg behind the screen with a handheld, add ~10%, adjust monitor backlight or contrast to be at +10% (38.5 fL) as the starting point. Or, is there a valid way to use a colorimeter for this? T

SecondaryColors

03-14-12, 02:29 AM

If I warmed up my tv for 60 minutes before calibrating, does that mean that I have to warm up my tv for 60 minutes before watching movies afterwards?

zoyd

03-14-12, 06:07 AM

no, 1 hour warm-up is overkill. I've never measured any changes in color/gray scale even after 5 minutes of on time. warm-up is more important for the meters but 20 minutes for those are adequate.

edit: [was referring to plasma display, measured change in Y over 1 hour was 1%, maximum shift in gray scale 0.3 dELuv ]

breadvan

03-14-12, 06:31 PM

If I warmed up my tv for 60 minutes before calibrating, does that mean that I have to warm up my tv for 60 minutes before watching movies afterwards?

A good article if you have not read it yet

http://www.spectracal.com/downloads/files/Website/Website%20Articles/Display%20Warm%20Up%20Rates%20-%20How%20Long%20is%20Enough.pdf

TomHuffman

03-14-12, 09:18 PM

I just tested a Samsung LED, and I got very different results.

http://www.chromapure.com/colortest/led.gif

I would agree that CCFL LCDs take forever to warm up. You may still get meaningful increases in luminance 30 minutes after turning it on. However, LED LCDs--at the least this Samsung--are much more stable than the SpectraCal test suggests. An hour after turning it on, its luminance was down only 2.8%, which is a negligible change.

More interesting to me is the change in chromaticity, which, although small, is probably worth considering.

In my opinion, 30 minutes is adequate warm-up time, but this is obviously a matter of judgment and practical time constraints.

breadvan

03-15-12, 07:12 AM

Tom what about projectors? And how far do you place your meter off the screen?

TomHuffman

03-15-12, 01:58 PM

Even the SpectraCal test showed that the fluctuation of a UHP bulb is tiny after the first couple of minutes of turning it on, hovering between 41 and 42 nits in their example.

Regarding distance the meter should be placed from the screen, this is one of the most common questions I am asked. In one sense the correct answer is, "It doesn't matter." The reason for this is that as you pull the sensor back from the screen the intensity of the light hitting the sensor decreases. However, the size of the reading area on the screen increases. These two cancel each other out. So, it really doesn't matter how far the probe is from the screen.

However, there are some practical constraints guiding meter placement. The most important one is that you don't want the size of the measurement area to be larger than the test pattern. This is especially important if you use window patterns. It makes getting a good aim point harder. For this reason, I generally recommend using full field test patterns for front projectors.

You also don't want the meter reading its own shadow, so you have to angle the meter a little. For high gain screens, such as the HP High Power, this is a problem because they color shift off axis. So you have to be very careful with screens like this to keep the angle as small as possible.

SMPTE recommends taking readings from the viewing position, but that will not be practical for most consumer probes. You need a professional tool with a good aiming system and narrow viewing angle for this. Readings from a couple of feet away should be fine in most cases.

Tom what about projectors? And how far do you place your meter off the screen?

airscapes

03-15-12, 02:42 PM

I have posted this in other places but will do it again since it really does work work with the HP 2.8 screen.

Obviously not everyone has their projector table mounted so please just use this as a guide.

I place the meter directly over the projector about 6" above (table mount) this is just below my eye level when seated. If it is shelf mounted then just under it so you have no shadow to deal with.

I then put up a pattern 80% or higher is fine.
Do a continuous read and look at the gray scale. You want to adjust the angle of the meter so red in the RGB mix is at it's lowest and the luminous is at it's highest. You will find very small amount of movement are required.

The reason you look at red is because the screen shifts red off angle so by adjusting the meter so the red is at it's lowest you will be as on axis as you can possibly be.

You will find that if you sit on axis this will provide the best possible calibrated image. Also, when you come back to tweak in 500 or 1000 hours, you will know that you are in exactly the same spot.

SecondaryColors

03-31-12, 05:34 AM

...

cinema mad

03-31-12, 06:36 AM

I suspect its got to do with the age of your TV, showing signs of retirement..

thebox

04-26-12, 12:39 PM

Tom,

I just found your topic. Very good indeed. :cool:

Please I just have a comment/question:

I have my display calibrated (Panasonic GT30 brazilian model) using Gamma 2.45 and found it more pleasure and good for the eyes than the 2.2-2.35 range. My room is fully dark. Source is blu-ray movies only.
I have contrast calibrated at default (50) and brightness at default either (0). Color and Tint have a few notches below from the default point (calibrated using a colorimeter - math graphics and not using the blue filter). Sharpness down to the minimum (0).
Grayscale is linear (RGB at 100% with Delta average under 4).

Now my question is... Once I had Gamma = 2.35 I didn't have the "pop" effect that I have now when using 2.45 value. Would I be running "off" using that Gamma valuefor my specific environment? I see it's been a long debate debate between Gamma 2.2 and 2.5 tough.... I just want to have my picture similar as the theater. :)

Thanks in advance for any thoughts and best regards,

Marcos

PlasmaPZ80U

04-26-12, 12:51 PM

thebox

04-26-12, 02:52 PM

What is the black level and peak white of this TV at current settings (Y readings in fL for 0% black and 100% white)?

The real (measured) contrast ratio of the TV will determine whether you can use very dark gammas without crushing shadow details together near black.

Hi PlasmaPZ80U !

Sorry- I should mentioned that. The contrast ratio mensured was about 3000:1 (not sure on top of my head now). Black Level was 0.02 cd/m2 and Luminance (peak white) was about 20 ftl (70-80 cd/m2).
I used Spears & Munsil, DVE and AVSHD 709 calibration disks.

Thanks!
Marcos

PlasmaPZ80U

04-26-12, 02:55 PM

Hi PlasmaPZ80U !

Sorry- I should mentioned that. The contrast ratio mensured was about 3000:1 (not sure on top of my head now). Black Level was 0.02 cd/m2 and Luminance (peak white) was about 20 ftl (70-80 cd/m2).
I used Spears & Munsil, DVE and AVSHD 709 calibration disks.

Thanks!
Marcos

seems like you probably have enough CR to avoid losing shadow detail with the higher gammas but I personally wouldn't go past 2.35-2.4 and something in the 2.2 to 2.3 range might be best

ultimately, there is no one correct answer and you should use whichever gamma looks best to you within the recommended range (which probably is between 2.2 and 2.4 for a dim or dark room)

zoyd

04-26-12, 05:17 PM

Hi PlasmaPZ80U !

Sorry- I should mentioned that. The contrast ratio mensured was about 3000:1 (not sure on top of my head now). Black Level was 0.02 cd/m2 and Luminance (peak white) was about 20 ftl (70-80 cd/m2).
I used Spears & Munsil, DVE and AVSHD 709 calibration disks.

Thanks!
Marcos

Given a peak white of 75 cd/m^2 and gamma 2.4 you will hit your mll floor at video code 23 so if you were to set your black level there you would crush levels 17-23. Most people don't calibrate that way though and if you follow the typical method of getting level 17 to be the first level to show some response the effect is actually to lower near-black gamma. So if you follow this method for setting brightness you don't have to worry about crushing near-black levels and if 2.4 is appropriate for your viewing environment you should be ok.