For the past year or so i have had to answer different questions on how the intensity response of a monitor behaves, nearly every week. Many of you probably know that the monitor intensity response to RGB values is nonlinear. Even if you didn't you can quite quickly verify so from reputable sources.

However, there is a small subset of people, who get completely stuck with the concept. Simply because they need to accept the nonlinearity as a fact by faith alone. So I have been thinking about ways to experimentally prove this, with help of things that can be found in most households.

The experiment that I've come up with is as follows:

  1. Take a piece of paper and fold it so that you have one side where you have paper layered 2 times and one side where you have the paper non layered.
  2. Place this paper against a monitor with on top of 2 color swatches and adjust the color in the other swatch so that its intensity is visually the same as the one that need to pass 2 layers of paper.
  3. Repeat for a few values plot the results on a graph paper and so on.

Ok, so I'm aware that the paper folded is not entirely linear either as theres 2 times more surface interfaces, and so on. But I can eliminate that nonlinearity by a additional experiment where I keep the monitor value constant and place different folds on top if O must. I'm just hoping it is sufficiently linear to prove my point without this step.

Now since I am on a vacation I have no calibrated screen or a colorimeter with me. So, I would like your input on this experimental setup, before I drop the idea onto the infinite list of things to do that i probably wont have time to resume later.

If you can come up with a better setup id be happy to hear it.

PS: I am aware that i can use a halftoned/pulsed color swatch to compare against but this has some drawbacks for certain viewers and is not so cool.

  • $\begingroup$ You have some good thoughts. But using vision to test brightness does not work well. The current edition of American Scientist has an article on the appearance of the solar corona you see during a total eclipse. You see optical illusions because of how your eye responds to brightness. You need instruments, which alas, are not cheap. Also you are right. brightness is not linear with the number of sheets of paper. $\endgroup$ – mmesser314 Jun 30 '16 at 13:46
  • $\begingroup$ If you don't want to rely on human vision (not quantitative) you could use an ambient light sensor which has been adapted to match human eye responsivity. More light = higher voltage output. They can be had for <$5 sparkfun.com/products/8688 $\endgroup$ – pentane Jun 30 '16 at 14:19
  • $\begingroup$ I would use a cheap encapsulated solar cell and a cheap multimeter. $\endgroup$ – CuriousOne Aug 2 '16 at 5:49

An idea:

Compare grayscale values to space-interpolated ones.

Give #808080 RGB-color to the left half of the desktop.

To the right half, make a 2D alternating pixel structure like this:

#fff #000 #fff #000

#000 #fff #000 #fff

#fff #000 #fff #000

#000 #fff #000 #fff

You can do this very simply with a browser younger as around 5 years, and with a little bit of css transition trickery. Sorry I won't fill out the PSE with HTML/CSS code :-)

Now see the monitor far, far away. By changing the left half, you can easily find what is the real greyscale equivalent of 50% black + 50% white.

By tuning the #808080, #fff and #000 values, you can even show the non-linearity graphically. You can do this also for the different R, G, B colors.

You will find surprising differences, if you do this with different type of monitors.

SO is very fast and very nice to answer easy CSS questions.

  • $\begingroup$ The gammacorrectors of many visual adjusters work just like this. But many people have problems doing this for some reason. But yes this does work. $\endgroup$ – joojaa Aug 2 '16 at 8:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.