How to match paint samples

Started by maf1909, May 11, 2016, 02:35:19 PM

Previous topic - Next topic

0 Members and 4 Guests are viewing this topic.

maf1909

I'm just starting out with keyshot, and am trying to get our catalog of paints into the system to be able to use them in future projects.  We've got 30+ metallic paint colors that I was hoping to match in a similar setup to our studio to get consistent color there, then be able to apply those paints to objects in more settings.

I've tried this technique (https://www.keyshot.com/forum/index.php?topic=11567.0), which looks great, but it's matching the colors that's proving to be difficult.  Even if I try to set up a "studio" environment similar to how we'd normally take a picture, with the same temperature lighting, it's a nightmare trying to guess at the colors to get them to match.

So my question is, what's the best way to go about matching paints like the attached picture?  For reference, that is a 1.75" square piece with a raised dome in the middle.

guest84672

Do you have access to the measured values of the color by any chance?


maf1909

Quote from: thomasteger on May 11, 2016, 05:27:57 PM
Do you have access to the measured values of the color by any chance?

I have access to a physical paint sample, a photoshop file containing a picture of that sample under studio lighting, and that's pretty much it.  I might be able to get the individual toners that go into the paint itself, but that's a long shot.

guest84672

What information does the PS file contain? Or is it similar to what you posted?

maf1909

Quote from: thomasteger on May 12, 2016, 06:51:00 AM
What information does the PS file contain? Or is it similar to what you posted?

the picture posted was exported from the photoshop file.  It was quite literally a photograph of the physical sample, taken in our studio.

guest84672

That's what I thought. It is difficult because you don't have consistent lighting. Otherwise you could just pick the color using the color picker.

If you had the measured values like Cie-Lab colors and gloss 60 value, it would be super-easy to match.

maf1909

My boss just dropped off a "Node" sensor that measures the lab values.  But how consistent can that get with a metallic paint?

guest84672

Give it a try. I'd be interested myself.

DMerz III

Matching colors from real life to digital is definitely a process, and it's very subjective because monitors vary in color output as well.

We haven't found a perfect system, but here is our process:

-Color calibrate our monitors, we have a small team and utilize over 10 monitors. We calibrate them on the same day, usually monthly if we can.

As a starting point, we take our physical sample and measure the CIE LAB #s

-We use an XRite Spectrophotometer (http://xritephoto.com/ph_product_overview.aspx?ID=1913)
(It's definitely not perfect, but it's consistently 'not perfect' if that makes sense, better than a photograph, which changes depending on light).

-Take those #s and put them into Keyshot's color picker, it gives us a great starting point.

-Using color balanced light (We use a D65 color temp lightbox) we put our physical sample next to our desk under the lights and turn everything else off except our monitor, which is turned away from the lightbox and physical sample.

-We manually tweak the #s in the color picker in Keyshot by eye.

-For Keyshot environment, we have a standard that we custom made ourselves and decided would be our "material" creation environment. Keyshot has an environment called "material ball" if you want to use that one. All of the lights are neutral.

Like I said, it's a bit of a process and not perfect, but the goal for us is consistency. So we're constantly trying to remove variables.

I am very interested in the new "measured" material rumored in Keyshot Beta. Although, I am not sure how we can capture the full appearance yet.

Thomas, does XRite have a tool for sale, or is it a service only type of thing?

maf1909

I pretty much eyeballed them all after building a simple studio environment in Keyshot.  My monitors were just calibrated last week.  I had a light setup that was reasonably close to D65, and got most of them as close as I can for now.  At least until I get more experience and work with some more detailed finishes to really get things to look the way I want.  If I get some flat samples I'll try out the color sensor we have again, the samples I have at the moment don't have enough flat space for the sensor to work properly.

Do you worry about the color space in Keyshot, or leave it default?  At the moment I plan on just rendering out pngs for our web guys to work with, but I imagine it'll be a future requirement to send them into photoshop and make changes in there.

guest84672

Quote from: dmerziii on May 18, 2016, 04:58:38 PM
Matching colors from real life to digital is definitely a process, and it's very subjective because monitors vary in color output as well.

We haven't found a perfect system, but here is our process:

-Color calibrate our monitors, we have a small team and utilize over 10 monitors. We calibrate them on the same day, usually monthly if we can.

As a starting point, we take our physical sample and measure the CIE LAB #s

-We use an XRite Spectrophotometer (http://xritephoto.com/ph_product_overview.aspx?ID=1913)
(It's definitely not perfect, but it's consistently 'not perfect' if that makes sense, better than a photograph, which changes depending on light).

-Take those #s and put them into Keyshot's color picker, it gives us a great starting point.

-Using color balanced light (We use a D65 color temp lightbox) we put our physical sample next to our desk under the lights and turn everything else off except our monitor, which is turned away from the lightbox and physical sample.

-We manually tweak the #s in the color picker in Keyshot by eye.

-For Keyshot environment, we have a standard that we custom made ourselves and decided would be our "material" creation environment. Keyshot has an environment called "material ball" if you want to use that one. All of the lights are neutral.

Like I said, it's a bit of a process and not perfect, but the goal for us is consistency. So we're constantly trying to remove variables.

I am very interested in the new "measured" material rumored in Keyshot Beta. Although, I am not sure how we can capture the full appearance yet.

Thomas, does XRite have a tool for sale, or is it a service only type of thing?

They do have a tool for sale, but it is $150,000. I believe they provide it as a service, though. And keep in mind, it doesn't work for translucent materials.

Another thing that you can do, and of course it depends on the material, but it will get you close, is by a gloss meter to measure the Gloss 60 value, and a color meter that will give you the Cie-Lab colors. You can plug all of these values directly into the color picker inside KeyShot.

Both devices are less than $1,000 (each). This is definitely a good investment.

And then you can use physical lights in KeyShot to simulate D65, D55 etc so you get an exact digital representation of colors and materials.

Let me know if you have further questions with regards to this.

maf1909

Looks like most of them came out pretty good just by matching to my screen.  Still have to get a couple other pairs of eyes to check them. 

I did have an issue with creating a consistent matte finish with the way I created the metallic paints.  I had been hoping just adjusting the IoR and the clearcoat roughness would suffice, but it tended to brighten the color way too much.  Ended up adding a translucent label and adjusting the colors to match a sample.  I'm not happy with those yet, so I'm going to have to look into different ways to do that, but it works for now.

Will Gibbons

Quote from: maf1909 on May 24, 2016, 11:52:42 AM
Looks like most of them came out pretty good just by matching to my screen.  Still have to get a couple other pairs of eyes to check them. 

I did have an issue with creating a consistent matte finish with the way I created the metallic paints.  I had been hoping just adjusting the IoR and the clearcoat roughness would suffice, but it tended to brighten the color way too much.  Ended up adding a translucent label and adjusting the colors to match a sample.  I'm not happy with those yet, so I'm going to have to look into different ways to do that, but it works for now.

I was curious about this myself and conducted a bit of an experiment. See the image below.

I noticed that by changing roughness only, there is no global brightening or darkening effect of the diffuse color at all. Then, while keeping the roughness values the same, I increased and decreased IoR and noticed some unfavorable loss in detail and contrast within the material as well as a value change of the diffuse color. I think your paints will look the best if you keep the default IoR consistent between them all as it's a measured value based upon the actual IoR of that material.

I do find your solution of using a translucent label for fine-tuning a clever one though. Hopefully you'll be able to find a result you're happy with without too much effort.

DMerz III

That's a great experiment Will! Thanks for posting this. Next time someone takes a shortcut and tries to justify changing the IoR needlessly, I've got this bookmarked as reference.

Also, the label idea is a pretty creative solve, going to have to back pocket that one in a pinch as well.

Thanks guys

Will Gibbons

Quote from: dmerziii on May 25, 2016, 01:56:26 PM
That's a great experiment Will! Thanks for posting this. Next time someone takes a shortcut and tries to justify changing the IoR needlessly, I've got this bookmarked as reference.

Also, the label idea is a pretty creative solve, going to have to back pocket that one in a pinch as well.

Thanks guys

Of course, I realize that it can be hard to track the results of all your changes. Although it's a bit time-intensive, creating reference charts like this make it easier to grasp the effects of certain features or settings.