Solved How to calculate RGB values to simulate a specific color temperature?

Hello!

Assuming I have a film camera and I want to take pictures of my monitor. Films are rated either as daylight (5500K or 6500K) or as tungsten (3200K or 3400K). In order to get the more accurate colors I have to match the color temperature of the monitor to the color temperature of the specific film type. However most monitors don't have values such as 3400K or 5500K, they only have values such as 6500K or 3200K. So I would like to know if I can simulate a specific color temperature by adjusting the R,G,B values from the monitor controls or from software such as nVidia Control Panel, Desktop color section. I guess I can, but I have no idea what values I must set to simulate a specific color temperature. Is there a formula or something to calculate it? Even better, is there a utility to choose the desired color temperature and set the values automatically, or at least tell me what they are?

Thanks a lot.

PS: Please don't suggest to use a digital camera! I am just curious how I would do it with a traditional film camera.

There are a couple of applications that can do it with decent results
f.lux
redshift

For the best results you would want a colorimeter and use something like DisplayCAL

This web page has a simple table: RGB to color temperature – Andreas Siess

But surely the disply will change if you change the monitor temperature - so where to start? To answer myself, the monitor should be at 6500K.

This page provides an approximate algorthm: How to Convert Temperature (K) to RGB: Algorithm and Sample Code

Wow! Thanks. I just need to simple Maths to convert 255 max to 100 max (we call it Method of three in Greek) and set the values in the monitor. Alternatively I can set them in nVidia Control Panel.

This batch script generates two sets of RGB numbers, one using Andreas Siess' lookup table and one using the Tanner Helland's algorithm. Because the lookup table only runs from 1000K - 12000K, the script is limited to that range.

garlin said:

This batch script...

Click to expand...

Please tell me you didn't knock this up after the question was posed? You had it up your sleeve already, I hope? I was in the process of creating my own little script to use the values in the tables (for my own amusment), but now I may give up... Martin

I don't plan to use any script. My programming skills stop at GW-Basic (or Q-Basic), very little Pascal before University and very little C at University. Now 30 years later, I barely remember my GW-Basic... I will just use the values from the table to change a photo's or video's colors to simulate the desired color temperature. Thanks.

mngerhold said:

Please tell me you didn't knock this up after the question was posed? You had it up your sleeve already, I hope? I was in the process of creating my own little script to use the values in the tables (for my own amusment), but now I may give up... Martin

Click to expand...

I just adapted the presented code from Tanner's article. His code example can be readily translated into any high-level programming or scripting language by fixing the if-then constructs, and the syntax for variables. The article notes how many other folks have translated it into their favorite languages,

Andreas' table of lookup values can be rewritten as a PowerShell array of arrays, using two global search & replace edits. I lucked out there in not having to do much work.

There's no reason not to convert it yourself, as a learning exercise. At least you have another script to confirm the output results are the same. It probably took me less time simply because in my previous line of work, I've had to convert random code examples to other languages.

spapakons said:

I don't plan to use any script.

Click to expand...

Speaking on behalf of the author, you only have to run the batch file, passing the colour temperature you want. The programming has been done.

It's there for anyone who's looking for a simple tool and isn't a coder. I found it interesting that one approach presents a table without much explanation about how it's calibrated (except to hint at another source), and the other dev explains his approximation method.

But without a colorimeter, you're supposed to run a LCD monitor calibration test by adjusting brightness, etc. Otherwise all this data is irrelevant if your monitor isn't somewhat calibrated, and doesn't have its native ICC color profile loaded.

garlin said:

I just adapted the presented code from Tanner's article.

Click to expand...

When I wrote my comment, I had copied the data table from Mitchell Charity’s web page, and used Excel to come up with curve fits - which not surprisingly produced the same (almost) coefficients for blue - I stopped there, as I was re-inventing the wheel, and not adding anything. If your batch file didn't work so easily, I would probably have coded it in vbs for simplicity.

But I was intrigued by your method of embedding the code in powershell, inside a batch file. I now know had seen that before a couple of years ago on TenForums, but forgotten about it. But I didn't note who posted the code back then, only the date - so I plugged 'timestamps' into the tenforum search window, and, hey presto, first up came the very thread, dated 22/12/23 - and guess who supplied the code? You probably already know it was you!

I had better think about using that method next time... Thanks, Martin