Help me understand relative vs. absolute greyscale - AVS Forum
Forum Jump: 
 
Thread Tools
post #1 of 19 Old 03-17-2013, 09:20 AM - Thread Starter
Advanced Member
 
Mavinwow's Avatar
 
Join Date: Jan 2007
Posts: 572
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 52 Post(s)
Liked: 63
Hello,

I am doing my best to calibrate the "standard" mode of my Samsung set. This is problematic in many ways, as I lose 10p calibration and 10% and 100% get a little crazy on me. However, this is the only way to get access to the microdimming features which are turned off in movie mode, and I just seem to like the picture better.

I dont seem to understand the difference between relative and absolute greyscale. When I pull up the workflow of 2p greyscale, with the 30% and 80% patterns, I can get fairly dialed in:

I am using an i1d3 with AVS09 patterns through a ps3 with calman 5.1 final.

78b0b35c_ScreenShot2013-03-14at12.21.21PM.png

However, when I go through the quickflow workup, the greyscale gets wonky for the low end on absolute, but stays decently ok for the relative. I guess I just dont understand what the absolute is trying to tell me.





Here are my colors incase that impacts the discussion. I have an element of blue in my red, and red in my blue, that I cant calibrate out as both are already set to 0...



Thanks very much, this is all new to me and trying my best to understand it all! Perhaps this is an effect of auto CE dimming at the lower greyscales?

My ES8000 settings, calibrated with an i1D3 and calman: Standard Mode, Movie Mode (out of date, will update soon)
CalMAN 5 Novice Walkthrough
Mavinwow is online now  
Sponsored Links
Advertisement
 
post #2 of 19 Old 03-17-2013, 10:09 AM
AVS Special Member
 
airscapes's Avatar
 
Join Date: Dec 2008
Location: Philadelphia
Posts: 4,619
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 54 Post(s)
Liked: 111
In a nut shell, you use Relative when doing a 2 point gray scale because you can not adjust gamma with a set of 2 point controls.
You are setting the RGB components of white, Relative to each other.

Absolute includes gamma in the the DeltaE and the single point chart.
When you set your gamma point in the settings of Calman, for instance to 2.2
Then it is important to start your 10 point gray scale run with 100% then 10% 20% 30%...
After the 100% is read, Calman will calculate how bright each point should be to achieve a 2.2 gamma.
You will notice the single point chart when doing Absolute starts at 0 and goes + or -. You goal is to get RBG at 0 for each point and if so, you will end up with a gamma of 2.2. If you are on the Plus side (to much light output) gamma will be below 2.2 as in 2.0 if you are on the negative side ( not enough light) you will above 2.2 as in 2.3 2.4 etc. Now if you have a plasma or have local dimming on, the outcome is going to vary with pattern sizes that could trigger the ABL or LD. I think a good rule of thumb for LCD that have local dimming is to use full field patterns to prevent the dimming.

Hope that helps.
airscapes is offline  
post #3 of 19 Old 03-17-2013, 11:10 AM - Thread Starter
Advanced Member
 
Mavinwow's Avatar
 
Join Date: Jan 2007
Posts: 572
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 52 Post(s)
Liked: 63
Thanks for the reply, starting to understand now. It did not make sense to me before why, during calibration of say movie mode with 10p adjustment, I was still having to adjust the 3 and 8 settings if they were already calibrated, but now it is to adjust the light output for the gamma.

My ES8000 settings, calibrated with an i1D3 and calman: Standard Mode, Movie Mode (out of date, will update soon)
CalMAN 5 Novice Walkthrough
Mavinwow is online now  
post #4 of 19 Old 03-18-2013, 11:51 AM
AVS Special Member
 
Doug Blackburn's Avatar
 
Join Date: May 2008
Location: San Francisco - East Bay area
Posts: 3,453
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 20 Post(s)
Liked: 226
There's one thing about the explanation that I don't agree with... if you have only 2 adjustment points for grayscale (cut/offset/low and gain/high/bias), you can still "adjust" gamma by moving the three RGB controls (3 for each adjustment range) up and down together. This will raise or lower luminance over the whole adjustment range of the grayscale control in question. Of course, if the cut/offset/low control affects everything from 0% to 60%, that whole range will get brighter if you raise all 3 controls (RGB) or dimmer. If the display is very well-designed and there aren't a lot of gamma variations between 0-60, that makes changing all 3 controls up or down useful for adjusting gamma (but you'll likely have to re-adjust Brightness, then re-adjust RGB and re-adjust Brightness, etc. until you have the right gamma and the right setting for the Brightness control at the same time). But if the TV has 1 or more peaks or dips in the range adjusted by the grayscale controls, you'll have to pick the best overall settings and live with the non-linearities in gamma.

If you are using the Relative setting for the luminance graph (usually set to a single point so you are only seeing the point you are adjusting/measuring), you won't have any feedback in that particular graph about how close or far you are from the gamma target you set in CalMAN (usually 2.2-2.3, I set it to 2.25 for almost every display or projector) . If you use the Absolute option for the graph, the zero point in the center is your goal and each color you get to match the 0 setting (no bar visible), means that you have that particular color adjusted to the proper luminance and when you have that, if all 3 colors are adjusted to 0 (no bars for any of the 3 colors), you have a perfect "neutral gray" at d65 AND you have the correct luminance for the measured step (per the gamma target you set before starting the calibration.... 2.2, 2.25 or 2.3.

The problem with using 2-point grayscale controls is that you could have a different gamma at each of the 10 grayscale measurement points if the TVs linearity is not ideal. Many of Pioneer's old Kuro models calibrated spectacularly well with just 2 grayscale adjustment points (high and low). A few newer models also have very linear response, but most TVs have so many ups and downs in luminance across the grayscale that you simply can't make gamma uniform for each grayscale step.

Calibration is a 3-dimensional concept -- you have to visualize what you are doing that way to be successful at it. When we use xyY coordinates, we are using all 3 dimensions (typically for setting grayscale for accuracy). The x and y coordinates determine how close to d65 you are... the closer the better. When you look at dE in CalMAN and other calibration software for grayscale, you are seeing only the error for x and y. The Y is your luminance and gamma determines what the Y measurement (in fL or cd/m2) should be for each grayscale step. So when you enter 2.25 as your gamma target, when you measure 100% white and get 35 fL, the software then knows what luminance (fL) you should have for each grayscale step. 70% white, would be, for example, would need to be about 16.2 fL (an estimate). The luminance (actual and target) for each gamma step changes any time you change the luminance level for 100%. So if you calibrate for 2.25 gamma when you set 100% white at 35 fL, if you then make 100% white brighter or darker, your absolute graph for each point will show that your grayscale steps are now "off" from where they should be (if they had been perfect with 35 fL). This concept is important to grasp in order to be efficient and effective when calibrating.

"Movies is magic..." Van Dyke Parks
THX Certified Professional Video Calibration
ISF -- HAA -- www.dBtheatrical.com
Widescreen Review -- Home Theater & Sound
Doug Blackburn is offline  
post #5 of 19 Old 03-18-2013, 12:05 PM
AVS Special Member
 
sotti's Avatar
 
Join Date: Aug 2004
Location: Seattle, WA
Posts: 6,610
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 21 Post(s)
Liked: 167
Quote:
Originally Posted by Doug Blackburn View Post

Calibration is a 3-dimensional concept -- you have to visualize what you are doing that way to be successful at it. When we use xyY coordinates, we are using all 3 dimensions (typically for setting grayscale for accuracy). The x and y coordinates determine how close to d65 you are... the closer the better. When you look at dE in CalMAN and other calibration software for grayscale, you are seeing only the error for x and y.

That does depend a little on the page you're looking at.

For our two-point pages, we do factor the luminance error out, but on our grayscale multi-point page we include the luminance or big Y error in the calculations. On any of our dE charts you can right click-> properties, and then on the properties panel there is a checkbox to include/exclude the luminance error.

Joel Barsotti
SpectraCal
CalMAN Lead Developer
sotti is offline  
post #6 of 19 Old 03-18-2013, 12:20 PM - Thread Starter
Advanced Member
 
Mavinwow's Avatar
 
Join Date: Jan 2007
Posts: 572
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 52 Post(s)
Liked: 63
Quote:
Originally Posted by Doug Blackburn View Post

There's one thing about the explanation....

Thanks very much for your reply. I thought I understood greyscale with x,y, but then Y has added quit a bit to my confusion. For example, I am now trying to calibrate "movie" mode on my ES8000. I calibrated for a bright room with 50fL output at 100% white, with target gamme set at default 2.2 and "use native gamma for gamut targets" checked (default, no idea what that means). My confusion is, when comparing to other peoples settings (yes I know, dont' compare), my 10p white settings are very different, including some dramatic shifts to raise the gamma curve.

Example:

1: 7, 4, 4
2: 8 ,6, 7
3: 7, 6, 8
4: 6, 5, 7
5: 5, 4, 6
6: 4, 2, 3
7: 3, 2, 2
8: 2, 2, 3
9: 2, 2, 1
10: 1, 0, 4




I end up with very pretty results, but am concerned that I am inherently doing something wrong as other calibrators have commented on my results. My workflow has been to set the backlight, brightness, contrast via AVS709 patterns, then the relative 2p white balance, then the 10p. I start at reading 100% white, then go to 10% white and tweak the 10p and work my way up. I am using an i1D3 meter with AVS709 through a PS3.

Any idea why I am needing such high offsets on my 10p? I could adjust the gamma setting from its default of 0, but am not sure why I would need to do that. Most people say that the ES8000 has a flat gamma and Im confused why my panel is needing so much tweaking.

Thanks very much!

My ES8000 settings, calibrated with an i1D3 and calman: Standard Mode, Movie Mode (out of date, will update soon)
CalMAN 5 Novice Walkthrough
Mavinwow is online now  
post #7 of 19 Old 03-18-2013, 12:33 PM
AVS Special Member
 
airscapes's Avatar
 
Join Date: Dec 2008
Location: Philadelphia
Posts: 4,619
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 54 Post(s)
Liked: 111
Quote:
Originally Posted by Mavinwow View Post


Any idea why I am needing such high offsets on my 10p? I could adjust the gamma setting from its default of 0, but am not sure why I would need to do that. Most people say that the ES8000 has a flat gamma and Im confused why my panel is needing so much tweaking.

Thanks very much!

The numbers do not matter.
One thing you did not list is what color temp did you start with? You should choose the one closes to D65 out of the box as you will have to make less adjustments.
airscapes is offline  
post #8 of 19 Old 03-18-2013, 12:34 PM - Thread Starter
Advanced Member
 
Mavinwow's Avatar
 
Join Date: Jan 2007
Posts: 572
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 52 Post(s)
Liked: 63
Quote:
Originally Posted by airscapes View Post

The numbers do not matter.
One thing you did not list is what color temp did you start with? You should choose the one closes to D65 out of the box as you will have to make less adjustments.

warm 2, closest to D65 on my Samsung

My ES8000 settings, calibrated with an i1D3 and calman: Standard Mode, Movie Mode (out of date, will update soon)
CalMAN 5 Novice Walkthrough
Mavinwow is online now  
post #9 of 19 Old 03-18-2013, 02:48 PM
Advanced Member
 
hungro's Avatar
 
Join Date: Mar 2007
Posts: 590
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 12 Post(s)
Liked: 56
the charts look good, I wouldn't worry about what comments you have recieved check out refrence material how does it look? If you are lovin the results as I am on my d6500 samsung then you are done. Just enjoy watching movies smile.gif
hungro is offline  
post #10 of 19 Old 03-19-2013, 03:00 PM
AVS Special Member
 
Doug Blackburn's Avatar
 
Join Date: May 2008
Location: San Francisco - East Bay area
Posts: 3,453
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 20 Post(s)
Liked: 226
Most every test/setup disc I've ever seen provides nothing that helps you determine the best backlight setting and contrast combination. If you follow their instructions, you end up with the brightest possible Contrast setting rather than one producing an appropriate amount of light from the panel.

Backlight setting is tricky at best... you want the Backlight as dark as possible while still achieving an image that's bright enough for the viewing conditions you are calibrating for. BUT!!!! Many backlights misbehave at low settings where they produce the blackest black levels... you end up with really bad color problems that are difficult or impossible to eliminate. So the lowest available blacklight setting may not be the best backlight setting... and some low backlight settings may reduce image brightness so much that you can't achieve 35 fL for dark-room viewing nor 50+ fL for bright room viewing. So you have to play with the backlight measuring different settings while re-setting Contrast for each different backlight setting to keep your peak white level where you want it to be.

You get color tints out of grayscale points by getting x and y correct. You get gamma right by picking the right gamma target in CalMAN, then manipulating the Y for each step until you get to the right luminance level for each grayscale point... that's your third dimension of calibration. As far as grayscale is concerned. CMS cal is also 3D if it is to be done completely.

"Movies is magic..." Van Dyke Parks
THX Certified Professional Video Calibration
ISF -- HAA -- www.dBtheatrical.com
Widescreen Review -- Home Theater & Sound
Doug Blackburn is offline  
post #11 of 19 Old 03-19-2013, 08:19 PM
10k
Advanced Member
 
10k's Avatar
 
Join Date: Sep 2005
Posts: 687
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 4 Post(s)
Liked: 38
Quote:
You get color tints out of grayscale points by getting x and y correct. You get gamma right by picking the right gamma target in CalMAN, then manipulating the Y for each step until you get to the right luminance level for each grayscale point... that's your third dimension of calibration. As far as grayscale is concerned. CMS cal is also 3D if it is to be done completely.
I have the same TV as op. The CMS on our samsung offers R, G, B controls for CMS. As you mention, CMS is 3d just like grayscale. With grayscale I have no problems thinking about things in terms of xyY because most of any Y adjustment comes from adjusting G in the whitepoint balance, and then adjusting R and B to get x and y back in line.

With CMS it is different though because depending on the color, G can impact x, y, Y, or some combination of all 3. Is there an excel spreadsheet or some way of approaching CMS that makes translating RGB controls into xyY adjustments more straightforward?
10k is offline  
post #12 of 19 Old 03-19-2013, 08:48 PM
AVS Special Member
 
sotti's Avatar
 
Join Date: Aug 2004
Location: Seattle, WA
Posts: 6,610
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 21 Post(s)
Liked: 167
What meter/software are you using?

CalMAN has it's CMS pages setup with RGB absolute charts that make it obvious what needs to be adjusted.

You could build a spread sheet, but it would be fairly involved.

Joel Barsotti
SpectraCal
CalMAN Lead Developer
sotti is offline  
post #13 of 19 Old 03-20-2013, 05:51 AM
10k
Advanced Member
 
10k's Avatar
 
Join Date: Sep 2005
Posts: 687
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 4 Post(s)
Liked: 38
Quote:
Originally Posted by sotti View Post

What meter/software are you using?

CalMAN has it's CMS pages setup with RGB absolute charts that make it obvious what needs to be adjusted.

You could build a spread sheet, but it would be fairly involved.
HCFR and a colormunki display. HCFR seems good, it reports xyY reference, and dE, and has an interactive CIE chart that updates in realtime in continuous measurement mode. Using the above I basically look at the CIE and "drag" around a color using RGB controls in CMS. Then check back that xyY is ok and dE is low. I was just wondering if perhaps there is a more scientific method to take an xyY measurement and translate that into "needs more R and less G", for example.

Thanks for the reply.
10k is offline  
post #14 of 19 Old 03-20-2013, 11:03 AM
AVS Special Member
 
sotti's Avatar
 
Join Date: Aug 2004
Location: Seattle, WA
Posts: 6,610
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 21 Post(s)
Liked: 167
Quote:
Originally Posted by 10k View Post

HCFR and a colormunki display. HCFR seems good, it reports xyY reference, and dE, and has an interactive CIE chart that updates in realtime in continuous measurement mode. Using the above I basically look at the CIE and "drag" around a color using RGB controls in CMS. Then check back that xyY is ok and dE is low. I was just wondering if perhaps there is a more scientific method to take an xyY measurement and translate that into "needs more R and less G", for example.

Thanks for the reply.

There absolutely is a more scientific way, it just takes quite a bit of know how to implement in a spreadsheet. Alternatively there are applications that can do it for you, granted I don't think any of them support the colormunki display.

Joel Barsotti
SpectraCal
CalMAN Lead Developer
sotti is offline  
post #15 of 19 Old 03-20-2013, 12:48 PM
AVS Special Member
 
Doug Blackburn's Avatar
 
Join Date: May 2008
Location: San Francisco - East Bay area
Posts: 3,453
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 20 Post(s)
Liked: 226
Joel,

I can't recall if I've tried this before or not, but it SHOULD work...

If the TV or processor's CMS controls are adjusting hue, saturation, and luminance, wouldn't it be more intuitive to set the CIE chart to uv mode (rather than the default xy mode) so that the hue and saturation controls would change u and v directly rather than having to do the xy to uv moves inside your head (I can do that OK, but it makes my head hurt after 30 minutes or so, LOL!).

If the CMS controls are adjusting Red, Green, and Blue (like Lumagen Radiance processors), it doesn't really matter whether you use xy or uv.

"Movies is magic..." Van Dyke Parks
THX Certified Professional Video Calibration
ISF -- HAA -- www.dBtheatrical.com
Widescreen Review -- Home Theater & Sound
Doug Blackburn is offline  
post #16 of 19 Old 03-20-2013, 07:01 PM
AVS Special Member
 
sotti's Avatar
 
Join Date: Aug 2004
Location: Seattle, WA
Posts: 6,610
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 21 Post(s)
Liked: 167
Quote:
Originally Posted by Doug Blackburn View Post

If the CMS controls are adjusting Red, Green, and Blue (like Lumagen Radiance processors), it doesn't really matter whether you use xy or uv.

xy, u'v', U*V* or even a*b* are all equally tricky to do CMS with. What you want is actual Hue and Saturation values, which of course we give you.



You can see our Quick View of gamut has RGB balance, directional hue, saturation and luminance indicators and x,y,Y values to help you adjust whatever kind of CMS you have.

Joel Barsotti
SpectraCal
CalMAN Lead Developer
sotti is offline  
post #17 of 19 Old 03-21-2013, 07:43 AM
10k
Advanced Member
 
10k's Avatar
 
Join Date: Sep 2005
Posts: 687
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 4 Post(s)
Liked: 38
Quote:
Originally Posted by sotti View Post

xy, u'v', U*V* or even a*b* are all equally tricky to do CMS with. What you want is actual Hue and Saturation values, which of course we give you.



You can see our Quick View of gamut has RGB balance, directional hue, saturation and luminance indicators and x,y,Y values to help you adjust whatever kind of CMS you have.
That looks very nice - too bad I cheaped out and got the colormunki display instead of the i1 rolleyes.gif

First, thank you for your informative response, even to someone who is not a customer of your software, respect.

I've figured a method for the above with HCFR which is less elegant, but should work, and which I will post here in case anyone else has the same question I had.
There are a few colorspace calculators online, here's an example - http://www.easyrgb.com/index.php?X=CALC#Result

If all you have is RGB controls in your CMS, in HCFR what you can do is go ahead and measure your primaries and secondaries and white. In the datatable screen you now have your xyY measurements and your dE and delta luminance amounts calculated against xy reference and your white measurement and reference gamma.

Click "editable data" at the top.
Change each of the primary and secondary color xy coordinates in the top box based on their rec709 reference positions rec709 references (Click to show)
Red primary: x=0.640 / y=0.330
Green primary: x=0.300 / y=0.600
Blue primary: x=0.150 / y=0.060
Yellow secondary: x=0.419 / y=0.505
Cyan secondary: x=0.225 / y=0.329
Magenta secondary: x=0.321 / y=0.154
Then in the Y coordinate, adjust the measured Y value by your "delta luminance" to get your target Y and input that as well. Now you should have deltaE = 0 and delta luminance = 0 in your primary and secondary measurements screen.
Uncheck "editable data" and click one of the colors in the table. The box in bottom left will now have your target RGB amounts (among other information). Put a pattern up and go into free measure mode in HCFR and adjust your RGB cms controls so that the RGB free measurements match the RGB reference you just calculated.

Certainly not as easy as what was posted above but it should work for the rest of us cheapskates out there biggrin.gif
10k is offline  
post #18 of 19 Old 03-21-2013, 06:12 PM
Advanced Member
 
eghill1125's Avatar
 
Join Date: Apr 2007
Location: Butler, PA
Posts: 917
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 11
"Then in the Y coordinate, adjust the measured Y value by your "delta luminance" to get your target Y and input that as well. Now you should have deltaE = 0 and delta luminance = 0 in your primary and secondary measurements screen.
Uncheck "editable data" and click one of the colors in the table. The box in bottom left will now have your target RGB amounts (among other information). Put a pattern up and go into free measure mode in HCFR and adjust your RGB cms controls so that the RGB free measurements match the RGB reference you just calculated."

What do you mean by adjust the measure Y by Delta Luminance here? If my red is showing -1.8% right now, do you mean then take my Y value of red and times that by .018 and then add that onto the existing Y value?

How does this change the way the target looks then? is it something similar at that point to Calmans above?

I am not I am understanding..

Ed

PSN - eghill1125
eghill1125 is offline  
post #19 of 19 Old 03-22-2013, 11:50 AM
10k
Advanced Member
 
10k's Avatar
 
Join Date: Sep 2005
Posts: 687
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 4 Post(s)
Liked: 38
Quote:
Originally Posted by eghill1125 View Post

"Then in the Y coordinate, adjust the measured Y value by your "delta luminance" to get your target Y and input that as well. Now you should have deltaE = 0 and delta luminance = 0 in your primary and secondary measurements screen.
Uncheck "editable data" and click one of the colors in the table. The box in bottom left will now have your target RGB amounts (among other information). Put a pattern up and go into free measure mode in HCFR and adjust your RGB cms controls so that the RGB free measurements match the RGB reference you just calculated."

What do you mean by adjust the measure Y by Delta Luminance here? If my red is showing -1.8% right now, do you mean then take my Y value of red and times that by .018 and then add that onto the existing Y value?

How does this change the way the target looks then? is it something similar at that point to Calmans above?

I am not I am understanding..
Yes basically that, although your maths are wrong. You would do current Y measurement / (100% - luminance error ) = target Y. So in your example you would divide your measurement by .982 to get the 100% target value.

It's basically my poor mans attempt at converting xyY targets into RGB targets that you can then compare against your live RGB measurements and manipulate with the samsung RGB CMS controls. I'm sure a lot of the more experienced folks are rolling their eyes at this point, but I'm a total amateur so please cut me some slack. smile.gif
10k is offline  
Reply Display Calibration

User Tag List

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off