10 step grayscale card with defined RGB values

RH
Posted By
Ronald_Hirsch
Feb 27, 2004
Views
2143
Replies
17
Status
Closed
Photoshop User Magazine had a very good article on color correction in Photoshop using Taz Tally’s 10 step grayscale card, with defined RGB values, in the Jan-Feb issue.

I’d like to get feedback from anyone who has used his card and approach.

Also, any feedback about availability of calibrated cards would be appreciated.

Ron Hirsch

Must-have mockup pack for every graphic designer 🔥🔥🔥

Easy-to-use drag-n-drop Photoshop scene creator with more than 2800 items.

JS
John_Slate
Feb 27, 2004
One would think you’d need the Lab values not RGB.
J
JasonSmith
Feb 27, 2004
Yup.
RH
Ronald_Hirsch
Feb 28, 2004
John,

While normal profiles may use LAB, the article in question uses RGB values. What I’m attempting to do is get feedback on the process, and locate where I can get such a card. To get the details, one should probably read the article, which is not an amateurish presentation. And, Taz Tally is a very respected source.

My IT-8 card for profiling seems like more card than needed for this use.

Ron Hirsch
J
JasonSmith
Feb 28, 2004
ALL profiles reference Lab, while RGB numbers are relegated to a particular monitor profile/setup.

Lab values are universal, RGB can and usually do vary from machine to machine.
RH
Ronald_Hirsch
Feb 28, 2004
Jason,

I take no issue with what you said. But the procedure in question is not really a profile per se.

What it does is to make color corrections similar to what the black and white eyedroppers in Levels and Curves do, except that it’s done at 10 levels of grey.

Have you seen the article???

Taz Tally is a Photoshop guru, and his procedure is certainly not whimsy.

Ron
J
JasonSmith
Mar 1, 2004
I have not seen the article – but to say that RGB values will give you definitive results across the board – across thousands of users and machines is FALSE.

If Taz is really such a guru – he should know that.

Now if he explains WHICH RGB space to be in – then that would be a start.
RH
Ronald_Hirsch
Mar 1, 2004
Jason,

OK, I’m pretty well convinced, and very surprised that Taz ‘s article is what it is. He also sells his calibrated greyscale card, and book.

From what I’ve also gathered in other text, Lab is the way to go. So, I’ll give that a whirl.

I have an IT-8 card with a file having all the values, and Gernot Hoffman was kind enough to send me some values in January for the card.

Ron
RH
Ronald_Hirsch
Mar 1, 2004
Jason,

I just noted that I had a reply from Gernot Hoffman in my thread of January 8th, entitled "10 step grayscale target". Gernot gave me Lab values for his IT-8 target, and I realized that I have the values for mine in the file that I got from Wolf Faust when I bought the IT-8 target from him last year.

So, I’m going to run a test photo session, using the IT-8 target. And then, using my Lab values, run a correction process to confirm that this will indeed do a good job at correcting things.

If this does indeed do what I want, I’ll look into using actions as much as possible to speed up the process.

QUESTION – Lets’ assume that I have an IT-8 image in one of numeruos images, all taken with the same lighting setup. What is the easiest way to apply the correction involved to other images.? I have tried using the new color matching process in CS, but I didn’t think it did that good a job.

I would appreciate it if you would read the message I posted to Gernot today in the referenced thread, and offer any suggestions you can.

I didn’t want to post it again in this thread, as I’d be chastised for doing that <G>.

Thanks again

Ron Hirsch
GH
Gernot_Hoffmann
Mar 1, 2004
Ronald,

on page 6 of this updated doc is a 100 patch gray chart with sRGB and CIELab numbers:
<http://www.fho-emden.de/~hoffmann/a3gencolortest.pdf>

One can see by PhS, using working space sRGB, that the RGB values are accurate within +-0.5 unit of 0..255
sRGB is a "calibrated RGB space" – the RGB numbers are functions of LAB values, only affected by rounding.

This gray chart can be used for adjusting printers by curves and of course for testing the gray balance of calibrated printers.

Best regards –Gernot Hoffmann
RH
Ronald_Hirsch
Mar 1, 2004
Gernot,

Thank you so much for the very nice document you’ve prepared – I’m printing it out now on my QMS 2300DL color laser. It appears to have lots of very good info in it.

The aspect of color correction is certainly not a simple thing. But, I’m working very hard to understand as much as I can about it

I don’t know if you referred back to my thread in January, where you replied on the 8th, and I just posted a response today. The task I’m trying to accomplish is using a test target in a digital photo as the basis for doing color correcting. Below I’ve repeated the workflow I’ve come up with. I have some questions, interspersed in the workflow, which you can certainly answer.

This is what I interpret as as the steps involved in using an IT-8 target in a digital photo, to come up with a color correction process

1. Open the digital image that includes the IT-8 target, in Photoshop, .

2. My DSLR uses a color space of sRGB – should I convert to Adobe RGB (1998) which is my default workspace, when opening it, or does this not matter – I would suspect not?

3. Convert the image to Lab mode.

4. Go to a given greyscale swatch on the target, and use the color sampler tool to record that sample – to keep a set of L-a-b values displayed on the screen for me to use.

5. Bring up curves. Start off with the "L" curve. Place the sampled point in question on the L curve by CTRL+clicking on that sample in the image.

6. Here’s where it’s a bit confusing – let’s say this is a middle point that shows as 52/52 in the x-y values. The value of "L" in the text file for the swatch in question is 50.3. Exactly how do I correct that point on the L curve? Do I merely change the output value so that it is "50.3"? Then do I repeat this process for "a" and "b" using the same approach?

7. Once I’ve done that point, go on to other grayscale points and do the same thing??? Continue on until as many points as I want have been done.

8. When I’m done, The image should be quite well corrected for luminance and chrominance. I assume that I should then convert back to RGB mode, and then save the file.

So, do I have things reasonably understood? If not, any corrections and advice would be appreciated.

It would be very nice if this procedure could be converted into an action, but I suspect that the steps involved would not lend themselves to a proper recording. However, I will work on that, and see how much I can automate.

It woud be nice if some guru (which I am not) could come up with an action or script, that could do this process. Packaging that with an IT-8 greyscale chart would make a very nice product, which I would be happy to purchase. Including the greyscale in important digital photos would make the process of color correction just about as fool proof as it could get. As a senior citizen/amateur photographer with a zillion years of doing things the hard way in a color darkroom, I love the new found freedoms that digital photography has provided in recent years. And, I’m doing all that I can to avail myself of all the "tools" that I can use. My high end darkroom is now permanently closed. Photoshop is my new "darkroom".

I truly appreciate all the help and info you’ve provided in the past.

Thanks again for your current help.

Ron Hirsch
RW
Rene_Walling
Mar 1, 2004
QUESTION – Lets’ assume that I have an IT-8 image in one of numeruos images, all taken with the same lighting setup. What is the easiest way to apply the correction involved to other images.?

Adjustment layers. You can drag n drop them from one file to the next one.

My DSLR uses a color space of sRGB – should I convert to Adobe RGB (1998) which is my default workspace, when opening it, or does this not matter – I would suspect not?

Since the next step (#3) is to convert to Lab mode, your suspicions are correct, it does not matter.
RH
Ronald_Hirsch
Mar 1, 2004
Rene,

I would assume that if I create a layer set which encompassess all the adjustment layers involved, that I can then do this via a single "drag".

Ron
GH
Gernot_Hoffmann
Mar 2, 2004
Ronald,

thanks for the feedback. IMO your workflow seems to be a little complicated. I´m only an amateur photographer but I have experimented with our Nikon D100 and different light, mainly for portraits.
It boils down to this:
Use indirect flash (via paper reflector) plus a fluorescent Just Normlicht. Both sources have about 5000K (at least near to each other).
The background is white paper. The paper appears light gray in the image, if properly exposed.
Then I measure a small averaged area on the background, which may deliver RGB=220/235/225. In Curves I move only the right end of R,G,B individually until the test area has e.g. RGB=228/228/228 (an estimated average). This delivers always a very reasonable white balance.

The main issue is IMO that one doesn´t need a graycard – a white background or a white shirt is better.
A multipoint match, as you want to do, might be interesting as an experiment. Considering ALL the difficulties taking good photos the color match is just one problem.

Sorry for not being able to answer your questions.
Only this: you have either sRGB or AdobeRGB(98) data. It´s not wrong to use AdobeRGB as working space and convert sRGB data. For my little saturated images I´m working consequently in sRGB.

My document should help to find the relations between Lab and sRGB colors easily, and the prints can be measured by an instrument.

Best regards –Gernot Hoffmann
RH
Ronald_Hirsch
Mar 2, 2004
Hi Gernot,

Thanks for your reply. The setup in my "temporary" studio has been giving me some color shifts, which can probably be attributed to the yellowish-beige wall colors. Of course, in a perfect world/studio setup, everything should be black, but that’s not practical in the real world.

I’ve been using a card that I made up, with white, black, and grey on it. But I thought that I’d try to expand on that. It does seem however that PS actions cannot record the steps I want to do, so that a 10 step process would be a very tedious procedure. So, what I may try is to manually do a ten step process using adjustment layers. Then when I’ve carefully done the first one, and am satisfied with it, I can make an adjustment layer set of all the layers needed to do the job. For other images taken in the same setup, I can then drag the layer set onto those images to produce the same corrections.

Thanks again for your kind help and replies.

Ron
RH
Ronald_Hirsch
Mar 2, 2004
Gernot,

One further question –

When you make the adjustments in the curves, I’d like to know exactly how you make the moves. Do you slide the right end points for R, G, and B, which started out at 255/255? This will keep the curve as a straight line. Or, do you locate the original white point on the curves, and then go into the input/output fields, and make change there. This will put a strong curvature into the curve, and would seem to be NG.

If you were to set the white dropper to 228-228-228, and then just click it on the white area in question in your image, would this not do the same thing you do via the curves?

Ron
GH
Gernot_Hoffmann
Mar 2, 2004
Ronald,

actually I´m doing this in my own program.
I had tested the workflow before by PhS.
Yes, straight lines by moving the end points: this
causes a little clipping for some values.
A midtone correction (banana) would be wrong.

Important is the averaging of the test area. I´m
using 10×10 pixels. Can PhS average more than 5×5 ?

White point correction without numbers doesn´t work
– adaptation prevents from detecting minor shifts.

Maybe it´s possible by setting the eyedropper as you
say, don´t know.

Best regards –Gernot Hoffmann
RH
Ronald_Hirsch
Mar 2, 2004
Hi Gernot,

Thaks for the info. I assumed that it had to be "straight line correction".

Photoshop does not have a sample larger than 5X5. But, a way around this is to make a marquee of the desired size, and use FILTER>BLUR>AVERAGE. This makes every pixel in the marquee the same RGB value.

Ron

Master Retouching Hair

Learn how to rescue details, remove flyaways, add volume, and enhance the definition of hair in any photo. We break down every tool and technique in Photoshop to get picture-perfect hair, every time.

Related Discussion Topics

Nice and short text about related topics in discussion sections