AVS Forum banner

1 - 20 of 23 Posts

·
Registered
Joined
·
156 Posts
Discussion Starter · #1 ·
Hi guys.

Anyone else experiencing problems with Sony 2pt and 10pt controls? (Sony X930E in my case)

It seems that as I work my way down the 10pt scale I start getting control misalignment problems beginning at IRE 60... For example, my IRE 60 control actually affects IRE 5.5, and my IRE 50 control actually affects IRE 4.3, and so on down to IRE 10.

My gamma brightness and RGB levels are set correctly and all processing is off so that's not it. Out of the box this TV is actually fairly close to targets. :)

Also, my contrast ratio is rather low compared to the Rtings report. I'm getting 3400:1 whereas Rtings is reporting somewhere around 5000:1 - (Native with dimming off)

Do you guys think this a defect that warrants an exchange at Best Buy. I'm a little afraid to exchange it as my current panel is pretty good in terms of light bleed, uniformity and so on. But the 10 pt controls and the contrast ratio really bothers me.

What do you think?
 

·
Registered
Joined
·
9,883 Posts
Hi guys.

Anyone else experiencing problems with Sony 2pt and 10pt controls? (Sony X930E in my case)

It seems that as I work my way down the 10pt scale I start getting control misalignment problems beginning at IRE 60... For example, my IRE 60 control actually affects IRE 5.5, and my IRE 50 control actually affects IRE 4.3, and so on down to IRE 10.

My gamma brightness and RGB levels are set correctly and all processing is off so that's not it. Out of the box this TV is actually fairly close to targets. :)

Also, my contrast ratio is rather low compared to the Rtings report. I'm getting 3400:1 whereas Rtings is reporting somewhere around 5000:1 - (Native with dimming off)

Do you guys think this a defect that warrants an exchange at Best Buy. I'm a little afraid to exchange it as my current panel is pretty good in terms of light bleed, uniformity and so on. But the 10 pt controls and the contrast ratio really bothers me.

What do you think?
Hi, it's better idea to use 21-Point Grayscale sweep to calibrate your display, to locate issues coming from calibration controls differencies per each point (RGB setting combo you have used) easier.

Idealy, run a complete 21-Point Grayscale sweep, look your RGB balance/gamma charts and decide what adjustments to apply to all controls you have available at once....then re-measure fully the 21-Point Grayscale...now you will see that your required adjustments will be less. Doing that keep an eye to remove or add the same ammount of (R,G,B values) per each control, so your adjustments to be linear per each color channel.

About your contrast ratio, Rtings use a checkerboard pattern for that measurement and they set the peak white to 100nits.

I have that pattern to my calibration disk (with meter placement pattern also) to the ChromaPure area of my disk, but you can use CalMAN/HCFR or LightSpace also for these measurements:



It's the Simultaneous Contrast (ANSI) pattern.

When you compare contrast ratio between displays/projectors, the peak white has to be configured to the same levels between the models you will use for that comparison, to have valid comparison.
 

·
Registered
Joined
·
156 Posts
Discussion Starter · #4 ·
Thanks Ted! And sorry for the late reply. I just saw your response today. I really appreciate your help and I respect your experience. I am now using Calman with 21 point sweeps to 'see what's going on'.

Regarding your idea of making RGB changes equal and all at the same time. Do you mean IRE 1 through IRE 10 should each receive the same changes, or do you mean each individual IRE's RGB should receive the same changes?

How do I 'flatten' the final result if all are adjusted the same? Will this not leave inaccuracies?

Finally, if I am changing all IRE's at the same time, how do I know how much to change them without the interactive guide to show me as I go? Just guess? Seems like a lot of a lot of back and forth.

Thanks again Ted for any advice you might share. :)

RJL.
 

·
Registered
Joined
·
156 Posts
Discussion Starter · #5 ·
The control displacement is normal.


Sent from my iPhone using Tapatalk
Thanks Chad... But I wonder why I've never seen this problem brfore on any of my less-expensive Samsungs. The IRE's always lined up before... Weird.
 

·
Registered
Joined
·
5,595 Posts
Finally, if I am changing all IRE's at the same time, how do I know how much to change them without the interactive guide to show me as I go? Just guess? Seems like a lot of a lot of back and forth.
RJL.
There's always back and forth since the controls interact. What Ted is saying is that you use the multipoint gray readings to determine any trends. Say your first run shows that Red is high across the board. Then you drop Red a click or two in the 2-point cuts/drives/bias/gains (whatever your set calls them), and do another multipoint grayscale read. Rinse and repeat. There's no shortcut.
 

·
Registered
Joined
·
156 Posts
Discussion Starter · #7 ·
There's always back and forth since the controls interact. What Ted is saying is that you use the multipoint gray readings to determine any trends. Say your first run shows that Red is high across the board. Then you drop Red a click or two in the 2-point cuts/drives/bias/gains (whatever your set calls them), and do another multipoint grayscale read. Rinse and repeat. There's no shortcut.
Thanks... So I have 2 point and 10 point controls. Would you suggest that I use this technique on both sets of controls, or just on the 10 point, and use the interactive feature in Calman to do the 2 point?
 

·
Registered
Joined
·
5,595 Posts
Thanks... So I have 2 point and 10 point controls. Would you suggest that I use this technique on both sets of controls, or just on the 10 point, and use the interactive feature in Calman to do the 2 point?
No. Since you've already said the 10-point doesn't line up, I'd reset it and leave it alone for now. Don't use the "interactive feature". Run a multipoint gray scale reading just as Ted has told you. Once you have all the points read, you can see overall trends. Make small changes on the two point and do another multipoint gray scale reading. If you have one color well above the other two, reduce that one first (the other two will come up as you do this) unless it's Green, which gives you most of your luminance. Do small changes only until you are done. Don't expect perfect. The controls just aren't fine enough. Get as close as you can, then take a break and watch some content. Chasing dEs can be frustrating, especially when you are starting out.
 

·
Registered
Joined
·
5,595 Posts
For instance, on a read like the one in this screenshot, you can see that Red is high nearly all the way across the gray scale. You would want to reduce Red one or two clicks on both the high and low ends, then remeasure. Because the controls have proportionately less impact as you go up the scale towards 100% (not IRE) White, you would probably need to take more off the drives/gains. As you do that, Blue and Green would begin to come up. Blue and Red will be a balancing act: as Blue drops, Red rises, and vice-versa. Actually, you have to juggle all 3 to get them as close to the target line as possible.
 

Attachments

·
Registered
Joined
·
156 Posts
Discussion Starter · #10 ·
No. Since you've already said the 10-point doesn't line up, I'd reset it and leave it alone for now. Don't use the "interactive feature". Run a multipoint gray scale reading just as Ted has told you. Once you have all the points read, you can see overall trends. Make small changes on the two point and do another multipoint gray scale reading. If you have one color well above the other two, reduce that one first (the other two will come up as you do this) unless it's Green, which gives you most of your luminance. Do small changes only until you are done. Don't expect perfect. The controls just aren't fine enough. Get as close as you can, then take a break and watch some content. Chasing dEs can be frustrating, especially when you are starting out.
Oh, I see. So I am using the 21 point graph in Calman to do the 2 point adjustments. This will show me where the 2 point controls are having the most effect... Try to get the RGB channels as close as possible to the target... Then try to do the 10 point the same way to finish the job... This allows for the most efficient and accurate use of the IRE controls rather than relying on the 'gray patches' to match up with controls, I'm letting the sweep show me where the controls actually are.
 

·
Registered
Joined
·
5,595 Posts
Oh, I see. So I am using the 21 point graph in Calman to do the 2 point adjustments. This will show me where the 2 point controls are having the most effect... Try to get the RGB channels as close as possible to the target... Then try to do the 10 point the same way to finish the job... This allows for the most efficient and accurate use of the IRE controls rather than relying on the 'gray patches' to match up with controls, I'm letting the sweep show me where the controls actually are.
Pretty much. The two-point controls overlap and control a wide range. Think of them as bass and treble controls with the 10-point as parametric EQ for finer adjustments.
 

·
Registered
Joined
·
156 Posts
Discussion Starter · #12 ·
Pretty much. The two-point controls overlap and control a wide range. Think of them as bass and treble controls with the 10-point as parametric EQ for finer adjustments.
Good analogy comparing the bass and treble controls with an EQ... I like audio stuff as well. :) And thanks for the screenshot. Really helps to see what you're talking about.

I managed to get a really flat response from my TV tonight using this procedure, and after doing a 4000+ patch color characterization using DispCAL, the results were spectacular! Color and gray scale DE's are all below 0.5! Amazing accurate and 3 dimensional picture. I do photography and my pictures have never looked to alive and balanced from near blacks to peak white.

I have a couple of years of self-taught experience with calibration, but I still have many questions and a lot to learn... Like HDR or choosing different gammas for different uses, or choosing initial settings of the TV making calibration easier or harder, etc.

I should have mentioned that my HDTV is connected to a Windows computer as it's sole source.
 

·
Registered
Joined
·
9,883 Posts
I am now using Calman with 21 point sweeps to 'see what's going on'.

Regarding your idea of making RGB changes equal and all at the same time. Do you mean IRE 1 through IRE 10 should each receive the same changes, or do you mean each individual IRE's RGB should receive the same changes?

How do I 'flatten' the final result if all are adjusted the same? Will this not leave inaccuracies?

Finally, if I am changing all IRE's at the same time, how do I know how much to change them without the interactive guide to show me as I go? Just guess? Seems like a lot of a lot of back and forth.
Hi, when you calibrate displays, it can happen to experience some issues which will try to make your calibrate time more painful.

There many issues you can experience like:

1) Mismatch of calibration controls adjustments of the display related with the patterns you are displaying and correcting, so an un-align of calibration controls affecting area.

2) When you display a static pattern for over some xx seconds, the display will enter to a dim mode, since it will see that you display the same static frame, to prevent temperature increase of panel or to power save to reduce the power consumption etc...it can dim the pattern.... so this can affect your adjustments decisions and not make them stable, becasue when you will re-measure with grayscale sweep you will see different results with the ones you had when you were calibrating a specific xx% gray pattern.

3) An adjustment of one point, can affect nearby points. So it can uncalibrate nearby when they are already calibrated.

4) OSD menus open while you measure a specific grayscale patch can affect the display output, so to have different results with OSD On/Off.

5) Large adjustment differences between 2 points; for example between 50 and 40% Gray RGB balance, it can increase the error to 45% Gray; so when you will look a grayscale ramp after the finished the 10-Point Grayscale calibration you can see a lot of issues to the gradation.

6) Extensive usage of calibration controls with random combinations of large scale can add processing issue to the image, with real content or color ramps etc.

For all these problem (I'm thinking right now) I recommend to take a full grayscale sweep and then examine the charts of RGB balance and gamma and imagine what adjustments you believe it will be better to apply to the whole grayscale adjustments controls at once and then re-measure the whole grayscale again.

For that reason to my calibration disk, all the gray/color patterns have maximum of 10 seconds duration (while I have chapters where the patches are autochange per 2 or 6 seconds also), I don't have 2 minutes for example of the same patch, to prevent the users for doing calibration OSD open (sometimes it's required, so you can pause) and stay to the same pattern for a long time. I haven't used numbers inside or outside the patch window, to say what pattern are you measuring (I use pattern announcement screens before each pattern)...external pattern generators don't use text notes for each pattern for example.

With that method, when you will be more familiar, it will provide you better visual results (beyond measurements...where they can be perfect looking by using any method).

Start from measuring all color temp modes (and them gamma modes) to see which one is closer to your target White Point and Gamma, the selection which will require less adjustments from you side.

For example when you will measure with grayscale sweep and you will see that you have ~10% more Red to most of the grayscale points you measure, you can correct that with many ways.....someone can:

1) Reduce for example -5 from Red slider and leave untouched the Blue/Green

2) or keep Red untouched but add +2 Green and +5 of Blue.

3) or reduce -2 Red but add +1 Green and +2 to Blue.

There many combinations you can do, also about what range of controls you have to tweak is related with each display design....its something you have to check....for example add or remove only 5 to a whole colorchannel and then re-measure with grayscale sweep to see how the RGB balance affected by these changes.

So when you will calibrate keeping the same 'strategy' of adjustments and keep a logic between grayscale RGB balance controls, the correction will be smoother and look better visually, since the processing of the display will be more linear.

A note for those who are addicted with Color Temperature charts.....when you see the RGB Balance Chart of a calibration software and you see the three (R/G/B) Channels Bars at exact 100% = 0 dE; doesn't mean that you have used equal percentage of each color channel.

The calibration software it's doing the normalizing internally according to the selected colorspace target options to give you better presentation for easier calibration.



D65 White Point for REC.709 (BD Movies) Color Space is using Red 21.27%, Green 71.52%, Blue 7.22% which gives 6504K.

REC.601 (PAL...EU DVD) D65: Red 22.20%, Green 70.67%, Blue 7.13% which gives 6504K.

REC.601 (NTSC... US DVD) D65: Red 21.24%, Green 70.11%, Blue 8.66% which gives 6504K.

REC.2020 (UltraHD Movies) D65: Red 26.27%, Green 67.80%, Blue 5.93% which gives 6504K.

All these colorspaces are using D65 as reference white point.

It happens (the most times) that the Warm1/2 of the consumer TVs (because coming not calibrated from their factory) to be closest mode when you have D65 White Point as a target, so for someone without measuring instruments, he is choosing one of those modes, these factory modes as selections are still providing uncalibrated picture.

When you have instruments/software, you can even make the 'cool' preset to match D65 White point, just it's better make less adjustments while you calibrating because some times large adjustments of the sliders can introduce problems like distortions, discolorations, clipping, banding, posterization....etc.

We use D65 (which has been created with specific mixture of RGB) which has 6504K because this is the white point the movies has been mastered (BD/UHD).

Each colorspace (REC.709 for BD / REC.2020 for UHD) while they have the same xy cordinates to create the D65, it's using different mixture of RGB colors channels, while all they report 6504K.

You can have 6504K temperature with different RGB channels mixture.

When you are using a meter/calibration software but you check only the Color Temperature Chart for the Grayscale, while the Color Temperature Graph can be perfect, the same time the RGB Balance Chart can be off.

This is happening because just a number 6504K is not the same as when we say D65 for a specific colorspace.

So when you are adjusting RGB balance controls, the Green channel adjustment is affecting more the luminance compared to the other 2 color channels, also when you reduce equally a specific point RGB sliders (-5 to all RGB sliders of 40% control for example), it can undo your RGB balance, all these stuff after practicing, testing, measuring, the user is improving it's calibration skills.....there no end for calibration, every day you learn something new ;)
 

·
Registered
Joined
·
9,883 Posts
I managed to get a really flat response from my TV tonight using this procedure, and after doing a 4000+ patch color characterization using DispCAL, the results were spectacular! Color and gray scale DE's are all below 0.5! Amazing accurate and 3 dimensional picture. I do photography and my pictures have never looked to alive and balanced from near blacks to peak white.
Display characterization using 3D LUT's is the ultimate correction you can have. In that case you don't need all that manual parametric pre-calibration stuff, only select your display largest gamut colorspace, do contrast/brightness/color clipping check, pre-calibrate with RGB-High (Gains) the 100% White only (while you will look your desired peak output level you want) and then start the measurements.

For those where they use HTPC for movie playback, they can use madVR as render which support 3D LUT also with DisplayCAL, for users with stand-alone players, they can use eeColor 3D LUT Box with DisplayCAL (since not all 3D LUT Boxes are supported by DisplayCAL), about these 2 options I have posted there: LG 2017 OLED Calibration Thread and Settings

About a 3D LUT post-verification here: CMS controls broken

For difference about normal calibration vs. 3D LUT here: Sony A1E 4K HDR OLED TV Thread (No Price Talk)
 

·
Registered
Joined
·
1,080 Posts
Display characterization using 3D LUT's is the ultimate correction you can have. In that case you don't need all that manual parametric pre-calibration stuff, only select your display largest gamut colorspace, do contrast/brightness/color clipping check, pre-calibrate with RGB-High (Gains) the 100% White only (while you will look your desired peak output level you want) and then start the measurements.
Ted, if generating a lut for REC 709 would you recommend having the Sony in REC 2020 color space mode even though its is a much larger gamut than 709 in order to give the software more to work with?
 

·
Registered
Joined
·
9,883 Posts
Ted, if generating a lut for REC 709 would you recommend having the Sony in REC 2020 color space mode even though its is a much larger gamut than 709 in order to give the software more to work with?
Hi, it's better to have some space to work, sometimes a display can have large gamut coverage to higher luminance levels but smaller to lower luminance levels, so when you will use the larger (native gamut) initially as a selection for 3D LUT it will help.

To the animated charts below (for Samsung + LG), the white triangle is the REC.709, the gray triangle is the measured. The animation showing the gamut coverage between 20-Points Luminance (5%-100%) of primary color with 100% Saturation, from 100% (255.255.255) down to 5% (13.13.13).

This is a measurement from Samsung 65JS9000 @ SDR mode:



This is AGIF Image (Animated GIF), this may playback to all browsers.

Also when you use native mode, there no active internal display gamut correction, since a fixed gamut options, lets say it's called 'REC.709' its a correction offset applyed by the display to the native gamut mode of the display. To some displays this is not so good, so it can affect display linearity, you can compare colorspace options when you look LightSpace's RGB Separation chart, so see how a display colorspace option affefting the display output to 20-point luminance range of primary colors.

See what is happening with LG 2016 BTW, while the Normal Gamut provides REC.709 primaries to the high luminance rarge users usually use to calibrate (100 or 75% Luminance patterns), at mid-low end the gamut is larger:



This is AGIF Image (Animated GIF), this may playback to all browsers.

The complete post for that LG issue see here: http://www.avsforum.com/forum/139-d...-oled-calibration-thread-30.html#post51051753

About your question about what to select to the Sony, you have to measure all gamut options to see how affecting gamut, and then decide (or post there).
 

·
Registered
Joined
·
156 Posts
Discussion Starter · #17 ·
Thanks again Ted. Once again your explanations prove invaluable! I appreciate you taking the time to teach. :)

I would love to use the 'Native' colorspace of my HDTV (REC 2020) during calibration\profilling, but because I use a Window's computer as my source, and the desktop isn't colormanaged, I will end up with a highly over-saturated desktop environment. And I can't afford a LUT box quite yet to correct this issue.

So I guess my choices are either to use MadTV (with a 3D LUT created using DispCAL) for movies, giving me an oversaturated desktop, or calibrate to the REC709 colorspace emulation (created by the TV) by way of the TV controls and creating a monitor color profile with DispCAL. (*.icc)

P.S. Are you suggesting NOT to manually calibrate the TV at the hardware level, (it's own internal controls) when creating a 3D LUT?
 

·
Registered
Joined
·
5,595 Posts
Thanks again Ted. Once again your explanations prove invaluable! I appreciate you taking the time to teach. :)

I would love to use the 'Native' colorspace of my HDTV (REC 2020) during calibration\profilling, but because I use a Window's computer as my source, and the desktop isn't colormanaged, I will end up with a highly over-saturated desktop environment. And I can't afford a LUT box quite yet to correct this issue.

So I guess my choices are either to use MadTV (with a 3D LUT created using DispCAL) for movies, giving me an oversaturated desktop, or calibrate to the REC709 colorspace emulation (created by the TV) by way of the TV controls and creating a monitor color profile with DispCAL. (*.icc)

P.S. Are you suggesting NOT to manually calibrate the TV at the hardware level, (it's own internal controls) when creating a 3D LUT?
Nope, he isn't. But it is a simpler basic calibration. You'll set the color of 100% White, set Brightness correctly, set Contrast/Backlight for your desired luminance along with no clipping (color shifts), and normally set gamut for widest possible. In your circumstance, Auto may be preferable. You are essentially putting the display "in a good place" for the 3D LUT to be calculated and to correct the areas where it is most needed.
 

·
Registered
Joined
·
156 Posts
Discussion Starter · #19 ·
Nope, he isn't. But it is a simpler basic calibration. You'll set the color of 100% White, set Brightness correctly, set Contrast/Backlight for your desired luminance along with no clipping (color shifts), and normally set gamut for widest possible. In your circumstance, Auto may be preferable. You are essentially putting the display "in a good place" for the 3D LUT to be calculated and to correct the areas where it is most needed.
Thanks RR. I do realize that the per-calibration requirements for a 3D LUT are simpler, but wouldn't a 3D LUT also benefit from a well-performed 10 point adjustment as well? I was under the impression that it is usually preferable to adjust at the hardware level (TV's controls) as much as possible before handing it over to a software correction.

Please correct me if I am mistaken, I want to learn. :)

Thanks.
 

·
Registered
Joined
·
5,595 Posts
Thanks RR. I do realize that the per-calibration requirements for a 3D LUT are simpler, but wouldn't a 3D LUT also benefit from a well-performed 10 point adjustment as well? I was under the impression that it is usually preferable to adjust at the hardware level (TV's controls) as much as possible before handing it over to a software correction.

Please correct me if I am mistaken, I want to learn. :)

Thanks.
Doing a 10-point before a 3D LUT is usually a waste of time. A large 3D LUT has far more adjustment range than the built-in controls, giving you better potential accuracy. Also, a 3D LUT will sometimes have to sacrifice a small bit of its gray scale accuracy in order to produce the most accurate gamut corrections. Spending a lot of time to achieve a sub-.5 dE on a manual gray scale before creating a LUT only to verify and find gray scale dE back up over .6 or so in places because of gamut considerations can be an unpleasant surprise and more than a little frustrating. Let the LUT do it.
 
1 - 20 of 23 Posts
Top