Home Theater Forum and Systems banner

41 - 60 of 80 Posts

·
Registered
Joined
·
44 Posts
Discussion Starter #41
But that would bring us back to a real BOM difference. :bigsmile:
I didn't say that there wasn't a cost difference between more fully-featured sets vs. more basic sets. I only said that having a properly-engineered color decoder didn't necessarily require more expensive components, merely more expensive engineers. Given the economies of scale, I doubt that there are too many companies cranking out cheap sets using proprietary solutions, but I don't claim to know the entire market. :spend:
 

·
Plain ole user
Joined
·
11,121 Posts
Pretty good steering of a topic there Bill. I even caught a hint of some statistics about to creep in. I wonder why?
 

·
Registered
Joined
·
2,604 Posts
I see... I misinterpreted what you were saying from get-go.... typical :ponder::rolleyesno::doh::whistling:
So one last thought.... "features" like CMS controls, or gamma controls, different firmware? Or more expensive chips? (Let's not differentiate between good/bad full3D/2D CMS or 10-point gamma sliders...)

BTW, thanks for indulging me here:T
 

·
Registered
Joined
·
44 Posts
Discussion Starter #44
I see... I misinterpreted what you were saying from get-go.... typical :ponder::rolleyesno::doh::whistling:
No worries. It's a good thing I keep this thread subscribed, otherwise it might be months before I remember to check back in on it (seems like a flurry of activity, and then dormant...).

So one last thought.... "features" like CMS controls, or gamma controls, different firmware? Or more expensive chips? (Let's not differentiate between good/bad full3D/2D CMS or 10-point gamma sliders...)
Definitely more expensive chips and more of them. The math isn't hard, but it is more math. Where it gets maddeningly complex is translating the matrix real-number math into integer math. In theory, if you put the CMS right at the point where you convert the inbound signal into linear RGB (i.e., at the RGB LUT), then you just need to do a relatively simple transform. People who talk about a CMS not being fully-featured because it doesn't include independent adjustments for secondaries that is different than the primaries, among other things, don't actually know how color math works. When you move the CMS from this "sweet spot", you end up adding complexity and cost into the design.
 

·
Plain ole user
Joined
·
11,121 Posts
Thr sweetness of that spot also depends on assumptions about what it takes to get "linear" RGB. Add to that characteristics of the display that may require some additional consideration for spectral densities that are not consistent with the approximations on the camera and encoding side, or primary limitations, or extra color wheel segments, or having to modulate panel response in various ways, or frame rate interactions with display latencies or a dozen other variables and we end up with non-trivial challenges. But then, that is why we need tools like CalMAN, right?
 

·
Registered
Joined
·
44 Posts
Discussion Starter #46
Thr sweetness of that spot also depends on assumptions about what it takes to get "linear" RGB. Add to that characteristics of the display that may require some additional consideration for spectral densities that are not consistent with the approximations on the camera and encoding side, or primary limitations, or extra color wheel segments, or having to modulate panel response in various ways, or frame rate interactions with display latencies or a dozen other variables and we end up with non-trivial challenges. But then, that is why we need tools like CalMAN, right?
So that we are clear, the conversion from R'G'B' or YCbCr to linear RGB is very straight-forward. You do have to make an assumption (or have an explicit control) for gamma, but that's really it. To your point, how the signal is then translated into physical light is definitely dependent upon display-specific factors.
 

·
Registered
Joined
·
2,604 Posts
No worries. It's a good thing I keep this thread subscribed, otherwise it might be months before I remember to check back in on it (seems like a flurry of activity, and then dormant...).
I have the opposite problem, I subscribe to so many just to make sure I don't miss anything important that it takes me over 2 hrs a day just to browse most and read the few that I find interesting...
Definitely more expensive chips and more of them.
ARG. But THAT means increased BOM cost! (Sorry, I just couldn't resist:whistling:)
Seriously though, I do see your points.
People who talk about a CMS not being fully-featured because it doesn't include independent adjustments for secondaries that is different than the primaries, among other things, don't actually know how color math works.
I wasn't going to get to this for a few months, but since you mention it...
Why is it that people sometimes report that their secondaries don't fall on the line between their primaries? Or for that matter report secondary luminances that aren't the sum of the primary luminances? (that IS the correct math, isn't it?) Is this simply a bad LUT or improper calcuations or something more, er, sinister?

Thr sweetness of that spot also depends on assumptions about what it takes to get "linear" RGB. Add to that characteristics of the display that may require some additional consideration for spectral densities that are not consistent with the approximations on the camera and encoding side, or primary limitations, or extra color wheel segments, or having to modulate panel response in various ways, or frame rate interactions with display latencies or a dozen other variables and we end up with non-trivial challenges. But then, that is why we need tools like CalMAN, right?
:rofl:Nice Len.<insert-shameless-plug-here> And you don't even work for them... (do you?):neener:
 

·
Registered
Joined
·
44 Posts
Discussion Starter #48
I wasn't going to get to this for a few months, but since you mention it...
Why is it that people sometimes report that their secondaries don't fall on the line between their primaries? Or for that matter report secondary luminances that aren't the sum of the primary luminances? (that IS the correct math, isn't it?) Is this simply a bad LUT or improper calcuations or something more, er, sinister?
Some of it could be bad information floating around. I have some issues with a certain often-quoted "guide" on another forum. Also, it could be misunderstanding of good information. Or it could be bad tools. Or ...

In other words, it is probably some combination of all of these issues. For a display that is worth calibrating, I doubt it is the display itself unless someone has just really screwed something up in a menu in which they didn't belong. :dumbcrazy::paddle::nono:

For example, unless someone finds a way around Grassman's law, the sum of the luminances of each primary is most assuredly equal to the total luminance.
http://en.wikipedia.org/wiki/Grassmann's_law_(optics)
 

·
Plain ole user
Joined
·
11,121 Posts
So that we are clear, the conversion from R'G'B' or YCbCr to linear RGB is very straight-forward. You do have to make an assumption (or have an explicit control) for gamma, but that's really it. To your point, how the signal is then translated into physical light is definitely dependent upon display-specific factors.
Yes, of course, the conversion is easy linear math. It is the assumption about gamma that is not so straightforward. Encode gamma can vary a great deal. There is no way for the display to know what it was, so the linearity is really an approximation. The error may not matter much, but in some cases can be significant. This is one reason that sources vary so much and the education of the user with regard to how to use controls even on a calibrated display is so important. It is also a reason that a properly calibrated display is essential. At least you have a starting point in the calibrated settings, which takes out much of the most objectionable variance in the results.
 

·
Registered
Joined
·
44 Posts
Discussion Starter #52
And here I thought all along that at least the movie studios stuck to the 0.45 encode... at least from what's "seen" on the monitor to the signal...
By and large they do, though you can't put too much of a universal law around this (the SMPTE and ITU standards are very clear). A lot of that is hard-coded into the capture devices themselves. Then, if the tech working on the production flow changes it, then it is probably intentional. Decode gamma is all over the map.
 

·
Registered
Joined
·
2,604 Posts
Then, if the tech working on the production flow changes it, then it is probably intentional.
Yes, they're free to "bake in" anything they want... sometimes I wonder if we couldn't make a distinction between what we call "gamma" for the encode/decode chain vs. "gamma" of a particular image...
 

·
Registered
Joined
·
44 Posts
Discussion Starter #54
Yes, they're free to "bake in" anything they want... sometimes I wonder if we couldn't make a distinction between what we call "gamma" for the encode/decode chain vs. "gamma" of a particular image...
You may be over-thinking the issue. When someone pushes the gamma on a scene intentionally, it is to create a particular look. If you correct for that push, then you lose rendering intent. What "ought" to happen is concrete guidelines that tie reproduction environment to recommended gamma in a way that accounts for the "desired" home environment (there are very strict/rigid guidelines for grading studios) so we can preserve rendering intent across heterogeneous environments.
 

·
Registered
Joined
·
2,604 Posts
I very often over think issues... it's what I do...:nerd:
But I think I'm OK on this one... I don't want to correct for any indended push... what I'm talking about is strictly language/usage to avoid what I see as some confusion among some of us less edumacated folks...
 

·
Registered
Joined
·
44 Posts
Discussion Starter #56
I very often over think issues... it's what I do...:nerd:
But I think I'm OK on this one... I don't want to correct for any indended push... what I'm talking about is strictly language/usage to avoid what I see as some confusion among some of us less edumacated folks...
All that the "average" enthusiast who cares about calibration and preserving rendering intent (which already puts you way, way above average) needs to know is that he or she needs to start with a 2.5 average gamma and adjust it downward to correct for ambient light. You then also need to potentially tweak specific parts of the gamma curve for display-specific limitations (and a bit of taste/content preference).

For most people with non-dedicated environments, something between a 2.0 and a 2.3 works, depending upon the specifics. In a bat cave, you can make a 2.5 work well.
 

·
Registered
Joined
·
2,604 Posts
(which already puts you way, way above average)
Awhhh! :heehee:
OK, granted the average guy doesn't need to be specific about language, but I've seen so much confusion between so-called experts (albeit not on this forum) when one's talking about encode, one's talking decode, one's talking about rendering... but I digress as usual...
You then also need to potentially tweak specific parts of the gamma curve for display-specific limitations (and a bit of taste/content preference).
I'm under the impression that without an outboard processor, this really can't be done, right? What displays have this resolution on their gamma adjustments, even in the service menus? Is there some way to adjust gamma I'm missing?
 

·
Registered
Joined
·
2,604 Posts
OK... I've given you guys a break for a while... time for another one... yes, that means I've been thinking again... (sorry about that...)...
CCFL backlights experience a color shift if they're set too bright, yes? Do LED edgelights do the same thing? How about backlights if you're left local dimming off? Local dimming on?
Assuming you've turned off all CE dimming...
Would you ever turn the backlight down to achieve a better D65? OF course, you still have your normal cuts/gains...
 

·
Registered
Joined
·
44 Posts
Discussion Starter #59 (Edited)
I'm under the impression that without an outboard processor, this really can't be done, right? What displays have this resolution on their gamma adjustments, even in the service menus? Is there some way to adjust gamma I'm missing?
My RS10 and Pioneer Kuro both have the ability to select different gamma curves. I believe that the Kuro can adjust its gamma curve at a more granular level with the ISF controls turned on (I've never bothered with mine since it is the "daily driver" TV). In general, though, yes, you will need an external processor for most of this.

OK... I've given you guys a break for a while... time for another one... yes, that means I've been thinking again... (sorry about that...)...
CCFL backlights experience a color shift if they're set too bright, yes? Do LED edgelights do the same thing? How about backlights if you're left local dimming off? Local dimming on?
Assuming you've turned off all CE dimming...
Would you ever turn the backlight down to achieve a better D65? OF course, you still have your normal cuts/gains...
Sorry I missed this one from a while back. To the best of my knowledge, LEDs do not color shift. The biggest "hit" on them is the relatively "spiky" (i.e., narrow) frequency range at which they operate, making some measurement instruments less useful in measuring performance. In order to measure them successfully, you need a spectroradiometer with a sufficiently narrow measurement interval or a colorimeter that has specifically been calibrated for use with LEDs.

As for CCFLs, I cannot attest to what happens in a lab, but my experience in the field where I've seen color shifting from an LCD panel seems more of an issue with the LCD side of the panel, rather than the backlight.

Net answer: use the backlight control to get the white point close to where you want your maximum luminance (i.e., not torch mode unless you have really specific reasons/issues). If the backlight control measurably affects grayscale, see if you can return the panel and get a different one.
 

·
Registered
Joined
·
2,604 Posts
Sorry I missed this one from a while back.
No worries!
or a colorimeter that has specifically been calibrated for use with LEDs.
This one has always perplexed me a little... different displays presumably have a choice of what LEDs they use, as such, how does a single profile matrix make a colorimeter accurate enough across the range of possibilities?
Net answer: use the backlight control to get the white point close to where you want your maximum luminance (i.e., not torch mode unless you have really specific reasons/issues). If the backlight control measurably affects grayscale, see if you can return the panel and get a different one.
OkeeDokey!
 
41 - 60 of 80 Posts
Top