What Photoshop fill for 18% grey card?

Started Dec 16, 2005 | Discussions
SimonG156 Forum Member • Posts: 61
What Photoshop fill for 18% grey card?

As subject, if I wanted to print a grey page at 18% grey, what should I use for the grey in Photoshop?
--
Cheers!
Simon

technoid Senior Member • Posts: 2,276
Re: What Photoshop fill for 18% grey card?
1

SimonG156 wrote:

As subject, if I wanted to print a grey page at 18% grey, what
should I use for the grey in Photoshop?
--
Cheers!
Simon

LAB. L=50*, a=0, b=0

Actually L=49.5, but 50 is the commonly accepted and used value. The difference isn't observable.

Make sure you select Absolute Colorimetric when you print or you could be off 10% or so due to paperwhite perceptual or relative scaling.

PSCS2 has a bug in the color fill mode when the image is rgb. Apply the LAB values to foreground then fill from foreground as a workaround.

Printer made "grey" cards are good for setting exposure, but not for white point balancing since they often have very spikey spectrums. The better printers for this have more neutral tone inks. For instance, the 2200 is a lot worse spectrally than the 2400 for this purpose. The 2400 is pretty good at neutral grey that stays neutral under different lighting.

Marty

MU84 New Member • Posts: 7
Re: What Photoshop fill for 18% grey card?
2

RGB: 124,124,124

#7C7C7C

Howard Moftich Veteran Member • Posts: 9,874
Re: What Photoshop fill for 18% grey card?
1

This is/was a 15 *year* old thread.

-- hide signature --

Online civility: Before you press 'Post', ask yourself if you'd say that to someone face to face.

Petruska Veteran Member • Posts: 8,939
Hey, we now have RGB numbers instead of only LAB....

Howard Moftich wrote:

This is/was a 15 *year* old thread.

Bob P.

technoid Senior Member • Posts: 2,276
Re: What Photoshop fill for 18% grey card?
2

MU84 wrote:

RGB: 124,124,124

#7C7C7C

No, it's not. First, no colorspace is given so I'll pick 3 common ones. RGB(124,124,124) is actually over 20% reflectance in both sRGB and Adobe RGB (1998) and way over that at 27% in ProPhoto RGB.

Using Adobe RGB the values should be (117,117,117). This is 18.01% reflectance when printed using Absolute Colorimetric. Absolute must be done because otherwise the card won't be neutral but shifted in the direction of paper white. Printing in any other intent other than Absolute will also shift the gray depending on how reflective the unprinted paper.  RGB (255,255,255) in Relative/Perceptual/Saturation Colorimetric Intent is unprinted paper white whether it is 85%, 90% or 95% reflective and smaller numbers are scaled to that paper white.

NAwlins Contrarian Veteran Member • Posts: 5,283
No--technoid is right

RGB: 124,124,124

#7C7C7C

I agree with technoid, RGB values are at best approximate indicators, not precise values, unless and until their color space is specified.

 NAwlins Contrarian's gear list:NAwlins Contrarian's gear list
Nikon Coolpix S30 Canon PowerShot S120 Sony Alpha DSLR-A580 Sony DT 16-50mm F2.8 SSM Tamron SP 70-300mm F/4-5.6 Di USD +5 more
Guillermo Shashte Contributing Member • Posts: 951
Re: No--technoid is right

>>Using Adobe RGB the values should be (117,117,117).

Does the exact value really matter? For production work it does but then you will not print your own card.

If I know the value of of my Kodak gray card I can pick a point that is close to that value; but what value do I pick if I don't know the actual value of a self-printed card.

I will have to pick a point that looks good to my eye anyway.

Using a Kodak card to set the white point

-- hide signature --

GS

NAwlins Contrarian Veteran Member • Posts: 5,283
Depends on use
1

Using Adobe RGB the values should be (117,117,117).

Does the exact value really matter? For production work it does but then you will not print your own card.

How much or even whether it matters really depends on what you are doing. It is not clear to me what the OP--way back in 2005--was trying to do.

IMO the more important point was that MU84's claim that "Photoshop fill for 18% grey card [is] RGB: 124,124,124" is simply wrong, and the two people who upvoted that are not aware of some important qualifiers, and are endorsing a misunderstanding / wrong statement. RGB values only define colors in reference to a specific color space (like sRGB or Adobe RGB), and most color spaces are not linear, instead having a gamma (some of which are a slightly-complex function). It would be better for people to know these things, and understand the limitations of all of this.

 NAwlins Contrarian's gear list:NAwlins Contrarian's gear list
Nikon Coolpix S30 Canon PowerShot S120 Sony Alpha DSLR-A580 Sony DT 16-50mm F2.8 SSM Tamron SP 70-300mm F/4-5.6 Di USD +5 more
Petruska Veteran Member • Posts: 8,939
The bottom line......

is that I would like to meet the person that can print with an inkjet printer a “true” 18% gray compared to a commercial gray card (ColorChecker).  Just produce a color in PS using any one of the suggested numbers, print it with many different printers and you will have many different looking shades of gray prints.  Matte papers versus luster, satin, gloss are also going to look different.

Later today I will spot read my Xrite ColorCheckers 18% patch using my I1Pro’s just to see what Lab and RGB values are indicated.

Bob P.

NAwlins Contrarian Veteran Member • Posts: 5,283
Re: The bottom line......

is that I would like to meet the person that can print with an inkjet printer a “true” 18% gray compared to a commercial gray card (ColorChecker). Just produce a color in PS using any one of the suggested numbers, print it with many different printers and you will have many different looking shades of gray prints. Matte papers versus luster, satin, gloss are also going to look different.

Agreed. Trying to print a gray card (except as a curiosity / exercise) seems like a bad idea, probably relatively high complexity for low quality.

Also, I have multiple gray cards, and I'm not confident they are even super-close to each other.

Later today I will spot read my Xrite ColorCheckers 18% patch using my I1Pro’s just to see what Lab and RGB values are indicated.

One thing that surprised me to learn was that the gray patches along the bottom row / right column of a ColorChecker are not designed for being neutral enough to white-balance on. With the ColorChecker Passport, there are two tiles on the opposite side with dished-in sides that are intended for neutral white-balancing:

The other tiles on those columns (in this view) are for, e.g., warming a portrait.

 NAwlins Contrarian's gear list:NAwlins Contrarian's gear list
Nikon Coolpix S30 Canon PowerShot S120 Sony Alpha DSLR-A580 Sony DT 16-50mm F2.8 SSM Tamron SP 70-300mm F/4-5.6 Di USD +5 more
Guillermo Shashte Contributing Member • Posts: 951
Re: Depends on use
1

So for setting White Balance in Photoshop does it matter?

I asked Robin Myers, who used to sell the plastic DGC-150 digital card and this was his reply:

As for the L*a*b* and RGB values, they changed slightly for each batch of plastic we received, and the type of light exposure and handling will have an effect on the values. So any numbers I supply are NOT exact. Your card will be slightly different.

For one of the last batches of cards we made in 2010, here are the averages. L* = 65.1, a* = -0.4, b* = -0.4. The average sRGB is 156, 157, 155; Adobe RGB is 156, 157, 155.

-- hide signature --

GS

pixelgenius
pixelgenius Senior Member • Posts: 4,224
Re: What Photoshop fill for 18% grey card?
3

MU84 wrote:

RGB: 124,124,124

#7C7C7C

Actually no, not with many RGB Working Spaces. And that's the rub, it depends on the RGB Working Space. As seen correctly in that very old post, it's really better to use a Lab value which is device independent.

The 18% gray correspond to a L* of 49.5. So in Photoshop, paint a gray with a Lstar=49 or 50 (you cannot decimal places in PS) and assign the image to the Working Space of Adobe RGB (1998). It render a RGB value of 116/116/116 for Lstar=49 and 118/118/118 for Lstar=50. ProPhoto RGB however is 100/100/100 and 119/119/119 for sRGB. So YMMV for RGB values, depending on the Working Color Space.

pixelgenius
pixelgenius Senior Member • Posts: 4,224
Re: The bottom line......

Petruska wrote:

is that I would like to meet the person that can print with an inkjet printer a “true” 18% gray compared to a commercial gray card (ColorChecker). Just produce a color in PS using any one of the suggested numbers, print it with many different printers and you will have many different looking shades of gray prints. Matte papers versus luster, satin, gloss are also going to look different.

Later today I will spot read my Xrite ColorCheckers 18% patch using my I1Pro’s just to see what Lab and RGB values are indicated.

Bob P.

There is this too:

http://www.babelcolor.com/colorchecker-2.htm#Xxr_v172_CCP2_data

ColorChecker_RGB_and_spectra.xls (155 KB .zip) (April 2012)This Excel spreadsheet contains:oRGB tables for 19 spaces, and L*a*b* (D50) values derived from averaged spectrums;oRGB tables for 19 spaces, and xyY (D50) values derived from GretagMacbeth L*a*b* D50 data (2005);oRGB tables for 19 spaces, and L*a*b* (D50) values derived from CIE (1931) xyY data (1976);oAveraged spectral data with Standard Deviation tables (from 30 charts)

Petruska Veteran Member • Posts: 8,939
Nice data at that link....

pixelgenius wrote:

Petruska wrote:

is that I would like to meet the person that can print with an inkjet printer a “true” 18% gray compared to a commercial gray card (ColorChecker). Just produce a color in PS using any one of the suggested numbers, print it with many different printers and you will have many different looking shades of gray prints. Matte papers versus luster, satin, gloss are also going to look different.

Later today I will spot read my Xrite ColorCheckers 18% patch using my I1Pro’s just to see what Lab and RGB values are indicated.

Bob P.

There is this too:

http://www.babelcolor.com/colorchecker-2.htm#Xxr_v172_CCP2_data

ColorChecker_RGB_and_spectra.xls (155 KB .zip) (April 2012)This Excel spreadsheet contains:oRGB tables for 19 spaces, and L*a*b* (D50) values derived from averaged spectrums;oRGB tables for 19 spaces, and xyY (D50) values derived from GretagMacbeth L*a*b* D50 data (2005);oRGB tables for 19 spaces, and L*a*b* (D50) values derived from CIE (1931) xyY data (1976);oAveraged spectral data with Standard Deviation tables (from 30 charts)

I measured the #23 patch (supposed to be the 18% gray) on the my ColorChecker Passport with my 3, I1Pro Rev D UVcut spectrometers using I1Profiler.

L*a*b measurements, D50 illuminate....

#1 I1Pro   36.41   -0.58   -0.35

#2 I1Pro   36.34   -0.62   -0.46

#3 I1Pro   36.37   -0.61   -0.33

Not exactly the same numbers as in the spreadsheet data but close enough.  My measurements do indicate to me that my 3 I1Pro spectrometers track closely to each other.  The #2 spectro has 15 hours of lamp time, the #3 has less than 60 seconds.  Thus anyone looking to buy a used I1Pro shouldn't shy away from one with a lot of lamp hours on it.

Bob P.

Iliah Borg Forum Pro • Posts: 27,765
Re: Nice data at that link....
2

Petruska wrote:

I measured the #23 patch (supposed to be the 18% gray)

It's close to 9% gray. You can use Y100 to read percents.

As pixelgenius notes in https://www.dpreview.com/forums/post/64158780 - 18% gray is close to L 50% in Lab.

-- hide signature --
Petruska Veteran Member • Posts: 8,939
It's sad when "Xrite" tells you....

Iliah Borg wrote:

Petruska wrote:

I measured the #23 patch (supposed to be the 18% gray)

It's close to 9% gray. You can use Y100 to read percents.

As pixelgenius notes in https://www.dpreview.com/forums/post/64158780 - 18% gray is close to L 50% in Lab.

that patch #23 is 18% gray!!!

I wish that I had their ColorChecker Passport 2 to see what the large gray panel is.

What is Y100?

Bob P.

Iliah Borg Forum Pro • Posts: 27,765
Re: It's sad when "Xrite" tells you....

Petruska wrote:

What is Y100?

Y100 is Y coordinate of XYZ normalized to 100%,

see http://brucelindbloom.com/ColorCalculator.html

enter Lab values (for 18%: 49.4961, 0, 0)

check "scale XYZ"

press "Lab" button to calculate XYZ

-- hide signature --
technoid Senior Member • Posts: 2,276
Just Use RGB 117,117,117 in Adobe RGB (1998)

Iliah Borg wrote:

Petruska wrote:

What is Y100?

Y100 is Y coordinate of XYZ normalized to 100%,

see http://brucelindbloom.com/ColorCalculator.html

enter Lab values (for 18%: 49.4961, 0, 0)

check "scale XYZ"

press "Lab" button to calculate XYZ

Right.

Also XYZ values are linear, aka gamma=1.0

To get the RGB values for a given reflectance use this when the colorspace has a pure gamma.

255 * .18^(1/gamma)

Turns out Adobe RGB gives the closest results when rounded.

So just enter this in google and it will calculate the result for you

255 * .18^(1/2.2)

116.95739395

Which rounds nicely to 117.

Now enter RGB 117,117,117 and fill a region in Adobe RGB (1998). Then use the info tool to examine the L*a*b* value. You can set the info tool to 32 bits and Photoshop will display 2 places of precision. The L*a*b* value is 49.53. This is 18.01% reflectance.

Darkmatterx76
Darkmatterx76 Regular Member • Posts: 102
Re: Just Use RGB 117,117,117 in Adobe RGB (1998)

Even "decent" grey cards aren't that expensive these days. I don't know why anyone now a days (yes, I saw the date of the OP... lol) would try to print a grey card instead of just buying one.

On a completely separate note, why does colour for calibration and printing have to be so flippingcomplicated!?! lol

DM

Keyboard shortcuts:
FForum MMy threads