or Connect
AVS › AVS Forum › Display Devices › Display Calibration › Reading and interpreting calibration charts and data for dummies
New Posts  All Forums:Forum Nav:

Reading and interpreting calibration charts and data for dummies - Page 5

post #121 of 181
Ursa,

Would you be so kind to briefly run through the math for these various variables with a few examples with the output from i1 and CF (have both).
post #122 of 181
NiToNi - Here's the math conceptually for the 1976 dE calculation (highly simplified):

dE (1976)
- You can use either CIE Luv or Lab color spaces - I'm assuming Luv here
Step 1: Convert xy to u' and v' - this ends up being neutral with respect to light
Step 2: Convert u' and v' into u* and v*
- Converting to u' and u* starts by taking the difference between the measured u' and the white point u' (thus, you need to convert D65 xy into u' v')
- Then you scale this value by L* multiplied by 13.
- Converting v' into v* is similar.
Step 3: Since you are finding the difference between the measured grayscale point and the ideal grayscale point (expressed in CIE Luv values), the dE calculation is essentially the square root of the squares of dL*plus u* plus v* when calculating a difference in grayscale.

So, Delta C* ends up being dE, but without dL*, but even then, there is still influence from L* itself. To truly isolate dC* from the influence of grayscale issues, you need to use the ideal L* value, rather than the measured L*. Why? Here's a specific example:

Say your display has a gamma "hitch" (inflection point) such that it puts out less light at a particular percent stimulus than it is supposed to, but everything else is where it should be when normalized. For a given absolute deviation (e.g., your x and y are off 0.1 each), when your display puts out less light than it otherwise should be, your dC* is lower than if it is tracking gamma correctly. This is an odd result, but the idea is that color errors are more noticeable as relative light levels increase. The corollary is also true. That is, if your display is putting out more light than it should, then the dC* is larger than it would be if it were tracking a smooth gamma curve.

With me so far? Let's now look at setting L* to 100 for purposes of calculating u* and v* (again, using the 1976 formula). For blue, you are essentially forcing an artificial non-linearity into your equation, and overstate the error by roughly 4 times (blue should only be about 7% of total light output). Red is less dramatically overstated, but is still significant. Green is closest, but still overstated (by potentially a bit more than 20%, depending upon which color standard you are using).

If you pull the equations from Poynton's site/book, you will see what I am talking about. Other methods of calculating dE introduce additional complications (e.g., dH), so we would probably need a dedicated thread to discuss those issues.

HTH.

Later,
Bill
post #123 of 181
WOW,
what a monster of a thread.
Got some more ideas of the deal.
With the Metavision 576 projector, you can dial in primary and secondary. Once you've done both, you have to tweak the primary again. Playing with them is almost like maxing out the dynamic with brightness and contrast, where you also have to redo the settings.

Best
Armin
post #124 of 181
Quote:
Originally Posted by Armin2 View Post

WOW,
what a monster of a thread.
Got some more ideas of the deal.
With the Metavision 576 projector, you can dial in primary and secondary. Once you've done both, you have to tweak the primary again. Playing with them is almost like maxing out the dynamic with brightness and contrast, where you also have to redo the settings.

Best
Armin



Yeah, really! IF dummies are supposed to understand this I'm in deep trouble!
post #125 of 181
I am a dummy, when you are going to calibrate using the Avia Disc with a Z2 do you turn down the IRIS and do you turn the enhanced black feature off on the DVD Player.
post #126 of 181
wow, my head is reeling.
post #127 of 181
Getting back to the titled subject:

1. How do you turn on the display device?
2. Where should you sit when playing with calibration settings?
3. How light or dark should the room be?
4. How long should you use the set before trying to calibrate it?
5. If you set has calibration charts available through it service menu should you try to use them or should you stick to a DVD, broadcast or calibration generator?
6. Which calibration chart should you start with, black, max white, color saturation, sharpness?
7. When setting basic calibration on a less than perfect display (100% on the current market) what are the best compromise settings?

You get my drift?

Richard.
post #128 of 181
alright need the experts to chime in and answer this one. to which standard do you calibrate to for a plasma tv in the united states do you choose NTSC or go with HD 709? i watch OTA some in HD c band satellite some in HD and DVD NOT HD. any help trying to set up my plasma using spydertv and would like to know which is my best choice. Thanks
post #129 of 181
Perhaps this question is answered somewhere here, and if so, I apologize. When using DVE, which is fed via component inputs, will the calibration also serve for HDMI. If not, how can the HDMI input be calibrated with this disc?
post #130 of 181
Quote:
Originally Posted by Feirstein View Post

Getting back to the titled subject:

1. How do you turn on the display device?

The power switch?
Quote:
2. Where should you sit when playing with calibration settings?

In your normal seating position, or wherever your gear requires you to sit. Only focus really matters, unless you are blocking the light.
Quote:
3. How light or dark should the room be?

As much or as little light as would be present when you are optimally watching your set. Measurements, though, work best in as completely dark of an environment as you can make it.
Quote:
4. How long should you use the set before trying to calibrate it?

If you have your own gear, it doesn't matter. If you are using a professional, then a rule of thumb is to wait a few hundred hours, plus or minusm, depending upon your display type.
Quote:
5. If you set has calibration charts available through it service menu should you try to use them or should you stick to a DVD, broadcast or calibration generator?

Some Samsung RPTVs need to use the interal patterns. However, you should generally use known references.
Quote:
6. Which calibration chart should you start with, black, max white, color saturation, sharpness?

These would be "patterns" in the calibration lingo. The charts to which this thread refers are generated by a calibration software package, so they would be histograms, tracking charts, etc.
Quote:
7. When setting basic calibration on a less than perfect display (100% on the current market) what are the best compromise settings?

Depends entirely upon the individual set.


Bill
post #131 of 181
Quote:
Originally Posted by spurge View Post

Perhaps this question is answered somewhere here, and if so, I apologize. When using DVE, which is fed via component inputs, will the calibration also serve for HDMI. If not, how can the HDMI input be calibrated with this disc?

For the average person, a PC's DVI port and a converter cable is their best option to calibrate and HDMI port to the HD standard. For SD, all that you need is a DVD player with a DVI or HDMI output. These can be had for <<$200 these days.

Bill
post #132 of 181
Some basic "guidance" regarding Calibration Charts and Data seems to have gotten very deep

So , as a Dummy , I'll ask...is there any negative reason to not place the sensor closer to the PJ than leaving it at the screen facing the PJ ? I think that there is a benefit from excluding ambient light and increasing the values measured at low IRE or % Stim levels . What are the other positives ?

Also , if I read correctly , the Gray Scale adjustment can or does affect the Secondaries?This is new to me . I understand that the Primaries are what they are for any given PJ excluding the ones that allow adjustments but I thought the Secondaries were a result of the Primaries and not easily affected by the Gray Scale?

Scott...................
post #133 of 181
The secondaries (if tint and color decoder are corrected) lie on the gamut triangle enclosed by the primaries. Their location is the intersection of the gamut triangle and the line thru the white point to the opposing primary.

The reason is that white is a specific balance of RGB. The secondary is nothing but White without one of the opposing primaries mixed in. In color math the vector equations are

W = R+G+B
Y = W-B = R+G
C = W-R = G+B
M = W-G = R+B

So grayscale correction is also correcting all of the colors enclosed within that secondary gamut. It is just called grayscale because it is done with grayscale patterns - per the math get that right and the secondaries will be right (assuming the video is first decoded properly)

So if you have a typical bluish-cyanish greyscale - cyan/magenta will be bluish and yellow will be greenish. This is why you want to do grayscale - even though the eye/brain is white adaptive without a D65 reference - it will not adjust to the secondary colors being off. If a banana is not banana yellow but the slightest bit green - our brain has the reference and says that ain't right!

Very easy to see this for yourself - most displays have advanced menus with RGB grayscale controls - just throw up a standard colorbars pattern and watch what happens when you adjust.

Legend is that standard calibration source material came about because someone was using a fruitbowl as a calibration image - and since there is no blue fruit - someone painted a banana blue. So the person on the image display side reversed tint so it was yellow!
post #134 of 181
Quote:
Originally Posted by Scott_R_K View Post

So , as a Dummy , I'll ask...is there any negative reason to not place the sensor closer to the PJ than leaving it at the screen facing the PJ ? I think that there is a benefit from excluding ambient light and increasing the values measured at low IRE or % Stim levels . What are the other positives ?

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

The negative reason is that PJ screens may not be neutral. But since you get better results with a lens reading by getting the light closer to the design target of the sensor - what you want to do is take a bright white reading of the screen and apply it as an offset to the lens reading.
post #135 of 181
Quote:
Originally Posted by krasmuzik View Post

The secondaries (if tint and color decoder are corrected) lie on the gamut triangle enclosed by the primaries. Their location is the intersection of the gamut triangle and the line thru the white point to the opposing primary.

The reason is that white is a specific balance of RGB. The secondary is nothing but White without one of the opposing primaries mixed in. In color math the vector equations are

Thanks Kraz ,

I was able to see this today after spending some quality time on my PJ . After getting the D65 point almost smack on , I remeasured the Secondaries and they had improved dramatically . My Primaries are just outside of the Gamut co-ordinates and are not adjustable from within the User Menu and thus will always affect the ultimate location for the Secondaries , yes ?

Is it not also true that you can have a 6500K calibration but not be on or near the D65 co-ordinate ? I think that this was my original problem and misconception . I'm learning

Scott...................
post #136 of 181
Quote:
Originally Posted by Scott_R_K View Post

Thanks Kraz ,

I was able to see this today after spending some quality time on my PJ . After getting the D65 point almost smack on , I remeasured the Secondaries and they had improved dramatically . My Primaries are just outside of the Gamut co-ordinates and are not adjustable from within the User Menu and thus will always affect the ultimate location for the Secondaries , yes ?

You've got it. Your color error will increase as you move out from the white point towards the perimeter of your display's gamut due to the primary locations being incorrect. Whether this is objectionable or not is a different question.

Quote:


Is it not also true that you can have a 6500K calibration but not be on or near the D65 co-ordinate ? I think that this was my original problem and misconception . I'm learning

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

You are two for two. 6500K is essentially a line segment on the gamut. D65 is merely one point on that line segment. You can actually be fairly green and still technically be at 6500K.

Bill
post #137 of 181
Yes check out the last WSR of a Sharp LCD - nailed 6500K - but the flourescent backlight made it 30dE off D65! Surely made earthlings look like martians!

The strange thing is this is something every newbie calibrator must learn - because the professional reviewers still report 6500K differences not D65dE. Despite it being the only failing question if they were to submit their ISF test. Their excuse is that is what the reader expects - and I disagree - customers only know what calibrators tell them - and getting a physics degree in blackbody curves is something they do not need to know - especially when it turns out it is not even a relevant measure - and certainly not a comparable metric!

Take a stand - abolish Kelvin charts from chartset!


As far as primary colors affecting secondaries - it depends. Certainly it will affect color depth if primaries are too deep. If just one is off rather than the other - then they should at least get the hue right by aligning with the white point in the color decoder - rather than just blinding mixing two primaries. I know Infocus has six axis color correction - even if primaries are a bit off - they nail the secondaries on hue relative to D65 (even if grayscale preset is different - they use D65 secondaries)

But there have been displays that cannot get secondaries right even when primaries are close and D65 is right. Red pushed color decoders are common - for a reddish magenta/yellow - as are color decoders that assume 9300K and align secondaries with that despite grayscale correction - giving blueish cyan/magenta.
post #138 of 181
Thanks Kraz and Bill ,

I've just seen posted in the Oppo Thread that someone had success bringing the Primaries in line by reducing "Color Saturation" in the DVD Player Menu . I haven't had time to explore this but will try this weekend . I'm also not clear on which Gamut is correct . The HDTV triangle better matches the location of the Primaries and the NTSC is definitely "smaller" . The Oppo outputs 720p RGB over HDMI/DVI into my Infocus SP5000 . I can select "native" and pass through the signal without further processing.

The other item I will go back and address is the effect each of the three Gammas and Color Temps have on the shape of the final Gamma curve . I'm now only getting a 2.92 Gamma with my current settings but these were set when I was using my eyes and I'm learning that they can be deceived

With the new tool (thanks Bill) everything is so much faster than the Smart-III that I can afford to experiment a little now .

Scott..................
post #139 of 181
Any advice on calibrating the PE7700 pj with the Spyder2pro. How to set it up, and be able to calibrate DVD's and also HI-Def tv from over the air antenna. I'm totaly new at this and all I want to achieve is a better picture over all. It doesn't have to be perfect, just better. I can't afford to pay an expert to achieve an optimal picture. I have to try to do this myself. Thanks!!
post #140 of 181
Quote:
Originally Posted by CRS59 View Post

Any advice on calibrating the PE7700 pj with the Spyder2pro. How to set it up, and be able to calibrate DVD's and also HI-Def tv from over the air antenna. I'm totaly new at this and all I want to achieve is a better picture over all. It doesn't have to be perfect, just better. I can't afford to pay an expert to achieve an optimal picture. I have to try to do this myself. Thanks!!

These questions are probably better handled in this thread:
http://www.avsforum.com/avs-vb/showthread.php?t=697072

Bill
post #141 of 181
Scott

Don't reduce color saturation to bring in primaries...

http://www.avsforum.com/avs-vb/showthread.php?t=767155
post #142 of 181
Quote:
Originally Posted by krasmuzik View Post

Scott

Don't reduce color saturation to bring in primaries...

http://www.avsforum.com/avs-vb/showthread.php?t=767155

Hmmmm OK ? I read through the Thread and I gather that reducing the overall Saturation , with non-PJ controls , risks changing the ratio of the Primaries which is a bad thing ? Is this correct ?

Both the Primaries and Secondaries have specific xy locations for either REC601 or REC709 Gamuts . Would it be possible to calculate a dE for each of the six positions from theoretical to measured and determine whether a Display is likely to be getting or showing either of the two Gamuts ?

I know that you use the Green in the Pluge pattern to tell which Space is being processed , but you have an experienced , trained eye . I'm looking for a non-experienced user "tool"

Thanks for sticking with me on this ,

Scott............................
post #143 of 181
It changes the brightness ratio of the primaries w.r.t. White - and since video is colorized B&W it is necessary to preserve this ratio. To eyeball color brightness is correct all you need is AVIA and the RGB filters and RGB colorbars.

Yes it is possible to measure the dE of color error - instead of using D65 white as your target - you just use REC709 RGBCMY as targets. I prefer to use CIELCH to break down dE into lightness (dL), colorfulness (dC) and hue (dH) errors - as that is much more informative and matches perceived errors. I discuss this earlier in this thread how I chart that.

You need a visual reference to truly identify color of color errors just as you do with grayscale errors - it is best done with colorimeter calibration gear and software that can measure the dE of colors. The point of the earlier thread on the Sony Pearl - is that sometimes you may need to make a choice of dC vs. dL when it comes to minimal dE. Which is worse - undersaturated video or candy primary colors?

Also just today I did a post on charting color which merits repeating again in this thread...

"
Referenced are charts from my Infocus SP7210 review. Based on the 1939 CIE chart which uses the spectral sensitivity of the eyes - you would think the Greens are bad, Reds are OK, Blues are great.

http://krasmuzik.biz/reviews/SP7210_demo_CIE39.jpg

But look at the CIE1976 chart which is based on psychovisual perception in the brain - and you see that RGB are perceived to be about the same amount of error - despite spectrally measuring different error.

http://krasmuzik.biz/reviews/SP7210_demo_CIE76.jpg

But we need to get a closer look so we zoom into the target - as well as consider color brightness biases perception...(in my newer charts I further break out color brightness)

http://krasmuzik.biz/reviews/SP7210.pdf

And now we can see that perceptually the stronger deep Reds are worse perceived error than the paler yellower Green! Since we also account for color brightness - Blues are not as bright - so on this chart there is less perceived error than on the previous chart.

And indeed anyone who has seen the Infocus SP7210 would probably agree with the last chart - your critical perception based on informed comparison is deeper crimson Reds, OK but paler yellow Greens, and great Blues. An uniformed comparison would simply say "Wow! Bright! Colorful!" But I think you would find few that would say the greens wanted to make you vomit for comparison....that only happens when reviewers look at charts to form subjective comments that are not based on objective perceptual measures.
"
post #144 of 181
Ok , Saturation is going back to default on the OPPO

Somebodies Server is very slow tonight so I can't see the jpg's . The PDF came through but I have no idea on how that data is to be interpreted . It's very new to me . Looks very intriguing !

"But I think you would find few that would say the greens wanted to make you vomit for comparison....that only happens when reviewers look at charts to form subjective comments that are not based on objective perceptual measures."

Very funny I'm glad your sense of humour is still with us .

My CC30R filter arrived today so I'll spend some time with it and see if I can curb the excessive Blue/Green this unit has . Someone mentioned that some manufacturers will drive the UHP lamps a little harder to increase brightness . I wonder if there is anything as simple as a "jumper" on the power supply that would put the lamp back into a more normal operating range . This would surely prolong lamp life too .

Merry Christmas ,

Scott.................
post #145 of 181
Scott the review format section of my website discusses the chart interpretation - but the data in the charts are standard CIE1976 colorspaces (CIELCH and CIELUV) which is covered in many color science textbooks.

http://krasmuzik.biz/reviews/reviewformat.htm

My mail server is out as well so try again I guess - I do not have a dedicated server as those cost big moolah!
post #146 of 181
Thanks Kras ,

The Server was passing the jpg's later and I went to your site for all the explanations . Very interesting . I like the approach .

Back to more "interpreting" of graphs , how might you explain the differences in the RGB Tracking graphs where say the Blue and Green track fairly linearly but the Red has a very visible "hump" in the middle of the graph ? I know that this is hard without a graph to look at , but I've been unable with the controls I have to "flatten" out the hump for the Red .

If we use Figure 7 from the "Guide to Basic display Calibration for New Users" , page 16, my graph looks very similar except the Red tracking is the inverse of the example . I see that Bill changed Gamma options to remove the "hitch" , but of all the Gamma Options I have , three , I can't seem to change this response .

I do have individual Gamma controls for the RGB channels but even using them doesn't seem to affect this problem . I don't know what to try next .

Scott....................
post #147 of 181
Quote:
Originally Posted by Scott_R_K View Post

Thanks Kras ,

The Server was passing the jpg's later and I went to your site for all the explanations . Very interesting . I like the approach .

Back to more "interpreting" of graphs , how might you explain the differences in the RGB Tracking graphs where say the Blue and Green track fairly linearly but the Red has a very visible "hump" in the middle of the graph ? I know that this is hard without a graph to look at , but I've been unable with the controls I have to "flatten" out the hump for the Red .

If we use Figure 7 from the "Guide to Basic display Calibration for New Users" , page 16, my graph looks very similar except the Red tracking is the inverse of the example . I see that Bill changed Gamma options to remove the "hitch" , but of all the Gamma Options I have , three , I can't seem to change this response .

I do have individual Gamma controls for the RGB channels but even using them doesn't seem to affect this problem . I don't know what to try next .

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

Scott - In many cases, you have to go with what the display provides. Where you have a large dip/hump, then this is a candidate for a parametric control like what Lumagen provides in the HDP/HDQ processors. This is also about the only time I endorse using these vs. continuing to work with the display's native controls. I'm much more inclined to use controls like these to change the incoming source information, rather than correcting the display. If you don't have a video processor, then once you have exhausted all of your possibilities, well, you are out of possibilities.

In other words, if you've tried various image modes, various gamma options and tried tweaking your gamma controls, then chances are you are out of options for correcting this. The only other thing to consider is to dive deep into the undocumented wild of the service menu and start playing there (e.g., some DLPs will expose the DLP panel controls in the service menu). The red hump at 20% is a bit of a signature for Optomas, by the way.

Bill
post #148 of 181
These humps are intentionally designed into the gamma tables. I see no perceptual reason for it. A hump at 80-90% they also use - that has the perceptual reason of detouring through magenta before you get to the bluish whites. Blue/Green pops at white are easy to get rid of - just lower the contrast (or better the Blue/Green gains) If you have a hump in the middle you can usually dial it out - that is just the gain/bias interacting badly.

You have to weigh the tradeoffs of using an external processor limited to 8b of adjustment on a limited number of steps (causing banding for sure) vs. living with it.

Calibration is the art of compromise applied to color science & video engineering - and some manufacturers have to be compromised more than others!
post #149 of 181
I have a simple question for anyone that might have an answer for me, thanks!!
If I want to calibrate my Projector with Syder2pro, do I need to have the Projector as the primary monitor comimg from the Computer, or can I just use my regular monitor for viewing the software details while I am reading the projector with the Spyder2?
post #150 of 181
Quote:
Originally Posted by Bear5k View Post

Scott - In many cases, you have to go with what the display provides. Where you have a large dip/hump, then this is a candidate for a parametric control like what Lumagen provides in the HDP/HDQ processors. This is also about the only time I endorse using these vs. continuing to work with the display's native controls. I'm much more inclined to use controls like these to change the incoming source information, rather than correcting the display. If you don't have a video processor, then once you have exhausted all of your possibilities, well, you are out of possibilities.

In other words, if you've tried various image modes, various gamma options and tried tweaking your gamma controls, then chances are you are out of options for correcting this. The only other thing to consider is to dive deep into the undocumented wild of the service menu and start playing there (e.g., some DLPs will expose the DLP panel controls in the service menu). The red hump at 20% is a bit of a signature for Optomas, by the way.

Bill

Thanks Bill . The only good news is that I saw the same "hump" when I was using Smart-III and was convinced that it was that particular package . Now with CalMAN , I know it's the PJ . Did I say that was good news I may try fiddling with the Oppo Gamma settings with the PJ at neutral . Nothing left to try except a new PJ and we all know that ain't gonna happen I have been making eyes at the DVDO product and I'll have to re-check and see what kind of video parameters are available for adjustment .

Quote"These humps are intentionally designed into the gamma tables. I see no perceptual reason for it. A hump at 80-90% they also use - that has the perceptual reason of detouring through magenta before you get to the bluish whites. Blue/Green pops at white are easy to get rid of - just lower the contrast (or better the Blue/Green gains) If you have a hump in the middle you can usually dial it out - that is just the gain/bias interacting badly.

You have to weigh the tradeoffs of using an external processor limited to 8b of adjustment on a limited number of steps (causing banding for sure) vs. living with it.

Calibration is the art of compromise applied to color science & video engineering - and some manufacturers have to be compromised more than others!"Kras

Thanks Kras . You say that the Gain/Bias interacting badly may be causing the "hump" and that it can be dialed out . That's encouraging as well . Any guidance here ? I'll be back at it tomorrow and Friday and see what I can learn .

Scott.........................
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Display Calibration
AVS › AVS Forum › Display Devices › Display Calibration › Reading and interpreting calibration charts and data for dummies