[Chruser]

Recently, I discovered that the PS desaturation does not only nullify the saturation of a coloured image, but it also alters the actual values. Grayscale does the same, but the value changes become significantly less. Take a look at the following picture, and check the values if you don't believe me:



Does anyone know a way around this? I know there are certain complicated ways using colour blending mode for layers, but this is rather impractical to check the exact values of your painting/photos etcetera. I haven't dabbed enough in Painter to find a grayscale mode just yet.

[aphelionart]

check on google... there's tutorials out there... i don't know of any automatic conversion, though. one way is to use the channel mixer, click on "monochrome", and play with the red/green/blue values depending on the colors in your image.. the percentages just have to add up to 100%

-matt

[drunken_muse]

I asked about this before with no real answer (that I can remember), so I'm also pretty curious. Maybe jfrancis knows though.

[jfrancis]

I think this is related to the issues discussed here:

http://forums.sijun.com/viewtopic.php?t=36391

...all I really know about the whole rgb value thing is thatcolor TV's desaturate using a certain equation that is based on human perception.

Google rgb to luminance, or just visit this link:
http://www.scantips.com/lumin.html

...you'll see that the supposedly "correct" way (based on human vision) to convert rgb to luminance is is compute the weighted average of r, g, and b as follows: 30% red + 59% green + 11% blue.

[Jin]

[Chruser]

Painter registers the values properly, but Photoshop doesn't. Actually, I think I'm just being confused by the HSB system as opposed to HSV.

[xflox]

No there�s nothing wrong in photoshop. it�s already been said.
The human perception of color is not "mathematical correct" so to say.
just look at these full saturated squares. the yellow appears much brighter than the blue, but mathematically they�re the same brightness!


(that�s what happens if you desaturate the colors, they all become 50% grey.)


if you convert it to grayscale, photoshop converts it in a way that fits to the perception of the human eye, because for example yellow stuff appears much brighter to our eyes. (grayscale photo film also works that way to prevent things from looking strange, for example a grayscale photo of plate of fruit would otherwise look as if they all had the same color.)

[Jorge]

Here's another thread wondering about the weirdness of Photoshop grayscaling. Rather useless, though, at least my takes on it. It's been ages since I last visited Sijun BTW (not that I ever participated a lot anyways...).
_________________
Jorge

[Godwin]

this is rather.... enlightening... never noticed this myself heh....
_________________
Derelict Studios|Godwin's Space

[ColdKodiak]

color theory 101
_________________
http://kontinue.com/coldkodiak/

[jfrancis]

[faB]

Photoshop's grayscale conversion may not be the most scientifically exact, in regard to the fact that the eye is most sensitive to light in the middle of the spectrum rather than at the ends; but I think it is close enough to what most artists refer to as 'value', and value relationships are what really matters anyway, so I'm not going to change software for this reason

However, what I find annoying in Photoshop is that while your image is in colour mode, the B component of Photoshop's HSB with the color picker , displays the intensity that would result from the Desaturation method, and thus is rather useless to me. Painter does the same 'error' by the way, only difference is Painter will show 50% V, while Photoshop will give 100% B for a fully saturated colour. So the only way I can check the "artists's value" on a colour image is to turn it to Grayscale first. I have tried using 'K' (grayscale) with the color picker but cant really understand it, neither can I really understand or use the Lab method

Here is a very complete page about the aspects of value & color vision:

http://www.handprint.com/HP/WCL/color10.html
_________________
"I'm not a shrimp, I'm a KING PRAWN !" -- Pepe.
selfportraits & stuff

[jfrancis]

In the L-A-B method, L is a sort of Lightness, and A is sometimes good for warming colors, since one end of the "A" slider is "towards red" - B is sometimes good for cooling colors, since I think one end of the "B" slider is "towards blue"

[fluO]

If I remember correctly the Value in Painter is skewed to cling to the rule that midtones are the most saturated. So for a given color you always get 50 % value at max saturation.

Yet it's not what you see I think :
- Maximum saturation is achieved when the light ray hitting your eye is monospectral ( one single wavelength )

- Depending on said wavelength the resulting Lightness computed by your eye is different because it relies mostly on the R and G output to determine it.
Roughly :
- Max saturated blue = lowest response
- Max saturated red = average
- Max saturated green = above average
- Max saturated Yellow = highest

It is a bit more complicated though, you can get pretty saturated color with some cliff shaped light profile across the whole spectrum , but you get the idea.

the L from CieLAB should reflect that ( it's a measurement of the color value againt 100 % white )
_________________
- They're in the kidney bowls, next to the colostomy bag with the chilli sauce in it.
...
- This isn't a meal - this is an autopsy!

[faB]

Talking about Lab, here's an interesting resource on turning images to black and white

http://www.carlvolk.com/photoshop21.htm

So it is possible also to turn your image into Lab mode, and then using the Lightness Channel for grayscale image.

The result seems a little lighter than the Grayscale conversion. As to which is the 'correct' one...
_________________
"I'm not a shrimp, I'm a KING PRAWN !" -- Pepe.
selfportraits & stuff