Seems that I missed a few posts while visiting Atlanta during the day…
Agreed...I can look at a CIE chart and see which colors (i.e. xy co-ordinates) are further way from the white point.
I will give your analogy of hue/intensity to frequency/amplitude some thought...as to saturation I am, as you, not going to be of much help..that is, at least not yet...
Zoyd, much thanks, the first part of your posting may prove helpful
Agreed and thanks for the suggestions re: books.
While I do not have any technical background – I am neither an engineer nor a scientist -- I am nonetheless hoping that my actuarial / applied math / statistics education/training will enable me to get beyond the paint numbers…that said, based on your experience am I over-reaching?
Agreed and understood…hard work has NEVER scared me away and I will not let it do so now…that said, I do want to know the hows and whys as painting by number, in anything, is just not me!
Apologies to ALL…it never has been nor never will be my intention to scare people away…I was simply trying to push the discussion on the basis that others would benefit as well…that said, I will cease and desist unless other posters indicate otherwise.
Understood mostly…much appreciate your time…hopefully after much more study and research on my part we can better resume our exchange…
Quote:
Originally Posted by TomHuffman 
Something that has always puzzled me is the concept of saturation..."distance from the white point" is a serviceable operational definition if you are familiar with CIE charts, but it doesn't really explain the concept.
Something that has always puzzled me is the concept of saturation..."distance from the white point" is a serviceable operational definition if you are familiar with CIE charts, but it doesn't really explain the concept.
Agreed...I can look at a CIE chart and see which colors (i.e. xy co-ordinates) are further way from the white point.
Quote:
Originally Posted by TomHuffman
What is difficult about the concept is that hue and intensity can each be understood as part of a wave phenomenon. Hue is like frequency and intensity (or brightness) is like amplitude or loudness. But is there an analogy in nature that fits saturation? "Excitation purity" doesn't really doesn't mean anything to me.
What is difficult about the concept is that hue and intensity can each be understood as part of a wave phenomenon. Hue is like frequency and intensity (or brightness) is like amplitude or loudness. But is there an analogy in nature that fits saturation? "Excitation purity" doesn't really doesn't mean anything to me.
I will give your analogy of hue/intensity to frequency/amplitude some thought...as to saturation I am, as you, not going to be of much help..that is, at least not yet...
Quote:
Originally Posted by zoyd
A purely saturated color contains one wavelength (intensity or lightness doesn't matter here). 550 nm is a purely saturated green, and so is 551 nm. This is equivalent to laser at 1 excitation wavelength which I gather is what Gregr was alluding to. You can think of hue as telling you where you are in the spectrum(frequency or wavelength) but it's always relative to some reference system (like RGB) to make any sense. Intensity is the area under whatever spectral curve you're looking at.
A purely saturated color contains one wavelength (intensity or lightness doesn't matter here). 550 nm is a purely saturated green, and so is 551 nm. This is equivalent to laser at 1 excitation wavelength which I gather is what Gregr was alluding to. You can think of hue as telling you where you are in the spectrum(frequency or wavelength) but it's always relative to some reference system (like RGB) to make any sense. Intensity is the area under whatever spectral curve you're looking at.
Zoyd, much thanks, the first part of your posting may prove helpful
Quote:
Originally Posted by krasmuzik
You need at least two textbooks - Poynton would be best until Gregr writes one. He will cover the video engineering extensively especially SMPTE formats with a smattering of color science. But you also need a solid textbook on color science especially CIE formats with a smattering of video engineering. Start your research at a good engineering school with a decent library that you can use before you waste your time buying "How Stuff Works...." or googling clueless ad-sponsored blogs spewing ignorant drivel based on something they read on a forum from someone who writes a blog. You would be hard pressed to find any blog or forum with thousands of pages whose information content is as high as a well regarded textbook!
You need at least two textbooks - Poynton would be best until Gregr writes one. He will cover the video engineering extensively especially SMPTE formats with a smattering of color science. But you also need a solid textbook on color science especially CIE formats with a smattering of video engineering. Start your research at a good engineering school with a decent library that you can use before you waste your time buying "How Stuff Works...." or googling clueless ad-sponsored blogs spewing ignorant drivel based on something they read on a forum from someone who writes a blog. You would be hard pressed to find any blog or forum with thousands of pages whose information content is as high as a well regarded textbook!
Agreed and thanks for the suggestions re: books.
Quote:
Originally Posted by krasmuzik
If you do not have an engineering/science background - you will never get beyond the paint by number automation of the charted sensors.
If you do not have an engineering/science background - you will never get beyond the paint by number automation of the charted sensors.
While I do not have any technical background – I am neither an engineer nor a scientist -- I am nonetheless hoping that my actuarial / applied math / statistics education/training will enable me to get beyond the paint numbers…that said, based on your experience am I over-reaching?
Quote:
Originally Posted by krasmuzik
Tom's Basic Guide thread here is an excellent paint by number system that can get people a long way- but if you desire to know why you measure Red at 21% to set Color Decoding...it requires significant learning on your part
Tom's Basic Guide thread here is an excellent paint by number system that can get people a long way- but if you desire to know why you measure Red at 21% to set Color Decoding...it requires significant learning on your part
Agreed and understood…hard work has NEVER scared me away and I will not let it do so now…that said, I do want to know the hows and whys as painting by number, in anything, is just not me!
Quote:
Originally Posted by krasmuzik
So your questions really do not belong in this thread - because you are scaring away those who are thinking about trying Tom's simple approach to calibrating a display - though certainly 'ask gregr' would make an excellent sticky.
Read the textbooks - then ask questions regarding what is not clear. The standard learning/teaching method works the same even on the internet. You are not helping casual readers by posting everything you have yet to read on the forum - you are just confusing them. Whereas if you point to the book and say this is the question I have not resolved - they can go check out the same book and see if they come to a different conclusion - then have the intellectual debate/discussion moderated by those who know more than the both of them. I of course vote gregr to keep that sticky in line - kicking out those who don't do their homework before they waste the time of a master.[/
So your questions really do not belong in this thread - because you are scaring away those who are thinking about trying Tom's simple approach to calibrating a display - though certainly 'ask gregr' would make an excellent sticky.
Read the textbooks - then ask questions regarding what is not clear. The standard learning/teaching method works the same even on the internet. You are not helping casual readers by posting everything you have yet to read on the forum - you are just confusing them. Whereas if you point to the book and say this is the question I have not resolved - they can go check out the same book and see if they come to a different conclusion - then have the intellectual debate/discussion moderated by those who know more than the both of them. I of course vote gregr to keep that sticky in line - kicking out those who don't do their homework before they waste the time of a master.[/
Apologies to ALL…it never has been nor never will be my intention to scare people away…I was simply trying to push the discussion on the basis that others would benefit as well…that said, I will cease and desist unless other posters indicate otherwise.
Quote:
Originally Posted by gregr
In an RGB display a color is 100% saturated if there are only one or two RGB components in a color, but desaturated if all three RGB color components exist in the color. If you have two RGB color components, say red and green, those two color components produce a "yellow" color with a hue somewhere along the line between the red and green primary colors. So the "yellow" hue shifts according to the proportion of red and green, but the color stays on the line connecting the red and green primaries. But if the color also includes some blue (i.e. it now has some proportion of all three RGB components) then the color moves toward reference white and it becomes less saturated. It is contaminated by white, i.e. it is less pure, it is less saturated. For instance, if the RGB color is (80%,80%, 0%) we have a 100% saturated, 80% amplitude, yellow hue color. But if the RGB color is (80%, 80%, 10%) it is equivalent to adding together two colors (70%,70%,0%) and (10%,10%,10%). The first color is a 100% saturated, 70% amplitude yellow hue and the second color is a 0% saturated, 10% amplitude white (i.e. gray). Together they form a color that is still a yellow hue, but it moves toward the reference white color. It is now desaturated by the white "contamination", i.e. less purity.
Now here's the tricky part. What is the saturation of the (80%, 80%, 10%) yellow color? Well it depends on how you decide to model the human visual system's perception, because saturation is a perceived quality. There are lots of models (HSL, HSV, etc.), but each is device dependent since different displays have different RGB primaries. Or we could take into account the CIE x,y or CIE u',v' positions of the RGB primary colors and the reference white color to figure out how much the yellow moved toward the reference white point compared to where it was without the blue component. Or use some other CIE perceptual model. Or we could completely ignore the human visual system and the actual colors and just define a saturation value based on the RGB signal levels (taking into account gamma). That's what we normally do if we define color bar test signals with less than 100% saturation.
The point I'm trying to make is that none of these "measures" of saturation will tell us what happened to the colors without knowing exactly what model was used, and what the original device primary colors were, and then reversing the saturation calculations (and reverse similar hue calculations) to get back to CIE coordinates. Or we could simply use the device independent dominant wavelength and Purity which define a unique CIE point as I described earlier.
In an RGB display a color is 100% saturated if there are only one or two RGB components in a color, but desaturated if all three RGB color components exist in the color. If you have two RGB color components, say red and green, those two color components produce a "yellow" color with a hue somewhere along the line between the red and green primary colors. So the "yellow" hue shifts according to the proportion of red and green, but the color stays on the line connecting the red and green primaries. But if the color also includes some blue (i.e. it now has some proportion of all three RGB components) then the color moves toward reference white and it becomes less saturated. It is contaminated by white, i.e. it is less pure, it is less saturated. For instance, if the RGB color is (80%,80%, 0%) we have a 100% saturated, 80% amplitude, yellow hue color. But if the RGB color is (80%, 80%, 10%) it is equivalent to adding together two colors (70%,70%,0%) and (10%,10%,10%). The first color is a 100% saturated, 70% amplitude yellow hue and the second color is a 0% saturated, 10% amplitude white (i.e. gray). Together they form a color that is still a yellow hue, but it moves toward the reference white color. It is now desaturated by the white "contamination", i.e. less purity.
Now here's the tricky part. What is the saturation of the (80%, 80%, 10%) yellow color? Well it depends on how you decide to model the human visual system's perception, because saturation is a perceived quality. There are lots of models (HSL, HSV, etc.), but each is device dependent since different displays have different RGB primaries. Or we could take into account the CIE x,y or CIE u',v' positions of the RGB primary colors and the reference white color to figure out how much the yellow moved toward the reference white point compared to where it was without the blue component. Or use some other CIE perceptual model. Or we could completely ignore the human visual system and the actual colors and just define a saturation value based on the RGB signal levels (taking into account gamma). That's what we normally do if we define color bar test signals with less than 100% saturation.
The point I'm trying to make is that none of these "measures" of saturation will tell us what happened to the colors without knowing exactly what model was used, and what the original device primary colors were, and then reversing the saturation calculations (and reverse similar hue calculations) to get back to CIE coordinates. Or we could simply use the device independent dominant wavelength and Purity which define a unique CIE point as I described earlier.
Understood mostly…much appreciate your time…hopefully after much more study and research on my part we can better resume our exchange…
