Color Management Experiment Kit #1: If seeing is believing, does your monitor profile matter?
There are two approaches to learning anything: Learn theory and then try to apply it. Or (my preferred approach) start experimenting and worry about theory later. A "Color Management Experiment Kit" has ICC monitor profiles to play with, experiments to try, and images to try them on, sort of like the digital darkroom equivalent of being handed a kid's chemistry set, except the only thing you'll risk blowing up is a few pixels. This set of experiments is for exploring the difference a monitor profile makes in what you see on your monitor screen.
Written March 2013. Updated December 2017.
Color management pundits can preach at you all day long about why you should use color management. But you aren't going listen until you have a darn good reason.
These experiments allow you to see for yourself how very small differences in monitor profiles produces very visible alterations in the colors displayed on your monitor screen. The downloadable "profiles pack" includes six synthetic monitor profiles, each of which deviates from sRGB by a very small amount, plus a couple of real monitor profiles.
Likely none of the profiles in the profiles pack are better than sRGB as a profile for your particular monitor, and most of them will be worse. If you can find a profile for your monitor on the internet, download it and add it to the test profiles for the experiments on this page. And of course if you normally use a custom monitor profile, add that profile to the list.
The following two websites offer profiles for a surprisingly large number of monitors and laptop screens. If your monitor isn't on either list, as always, search the internet:
- TFT Central: ICC Profiles and Monitor Settings Database
- DigitalVersus: Colour Calibration Profiles for your Monitor: If you've never thought about using a custom monitor profile, this article is well worth reading. Here's an excerpt to tempt you to read the whole article:
By default, monitors display colours that range from relatively accurate at best to absolutely terrible. One shade often dominates, and recently, we've seen a tendency for blue to be too visible everywhere except in white, which looks red instead.
The cold hard truth is that if you don't use color management, or if you do use color management but you use sRGB as your monitor profile, then you don't know what your images really look like. You only know what sRGB makes your images look like on your monitor. (Unless, of course, you've calibrated your monitor to exactly match sRGB — but another cold hard truth is that LCD monitors can't be calibrated to exactly match sRGB.)
Note: These experiments assume that you have a standard monitor. If you have a wide-gamut monitor, the colors will be extremely oversaturated unless you use your monitor's sRGB preset.
Set-up for the experiments
Here's how to do the experiments:
- Open a test image with a color-managed image editor.
- Make some changes to the Color Management settings.
- Watch what happens to the test image.
It really is that simple. It will longer to read through the experiments than do them. What may surprise you is how much you can learn about color management theory just by watching what happens to your images. Here's what you need to do the experiments:
Here's everything you need to do the experiments:
A color-managed image editor that allows you to easily change your monitor profile: Gimp, Krita (as of 2.6), and Cinepaint (not using Oyranos) all allow you to easily change your monitor profile. (Unfortunately, changing your monitor profile using digiKam 2.9 or Krita 2.5 requires that you exit the program, change the system monitor profile, then restart the program.)
This page assumes you are using Linux with Gimp 2.8 or 2.9. However, the second experiment won't work using Gimp 2.8 because of a bug in the color management plugin; Gimp 2.9 uses LCMS2 and isn't affected by the bug.
Test images and profiles: Download and extract this Color Management Experiment Kit monitor profiles pack and test images. Put the test images in any convenient location. Put the ICC profiles wherever your editing program looks for ICC profiles. Gimp looks wherever you want it to look (good for Gimp, I wish all editing programs were this obliging). Most other image editing programs only search for ICC profiles in particular folders, but usually you can add a new folder.
Your workspace environment: Because you want your eyes to accomodate to your monitor rather than to any surrounding light sources, dim your workspace lights so they are quite not as bright as your monitor. It's hard to judge colors on a monitor when your eyeballs are trying to ignore strong directional light, so keep extraneous light reflections off your monitor and out of your eyes. If you don't have specialized D50 lighting (I don't), window light works well as long as it's not reflecting off your screen.
As you'll be judging color changes, there shouldn't be any strong colors in your field of view or on your desktop (my own desktop is a solid middle gray). Wear a dark gray or black shirt to avoid creating color reflections on your monitor. The experts recommend neutral gray walls, but spouses tend to object.
Gimp Color Management settings: If you've never opened the Gimp Color Management settings, now is a good time to get acquainted. Open Gimp. Open the Gimp Color Management settings by clicking "Edit/Preferences/Color Management". Then make the Color Management settings match the picture below:
- Set Mode of operation to Color managed display.
- Set RGB profile to none.
- Set CMYK profile to none.
- Set Monitor profile to sRGB-IEC61966.icc (included in the color management kit profiles pack). Note: the Gimp profile boxes show the profile description, which is a bit of descriptive text embedded inside the profile; all the other profiles in the profiles pack have descriptions that exactly match their file names, but this profile has an extended description.
There are probably a dozen or more different sRGB profiles on your computer right now, installed by various open source software packages. Most of them are so close to one another that which one you use will make no discernible difference in what gets displayed on your monitor. But some of the sRGB profiles that float around open source software have the wrong tone curve, which will make a difference in what gets displayed on your monitor. So to make sure we are all playing on the same page, I've included in the profiles pack the original, real, complete and unmodified sRGB profile (source: Argyllcms). That's the profile whose description is shown next to "Monitor profile" in the picture above.
- Make sure "Try to use the system monitor profile" is unchecked.
- Set Display rendering intent to Relative Colorimetric.
- Set Softproof rendering intent to Relative Colorimetric.
- Make sure "Mark out of gamut colors" is unchecked.
- Set File Open behavior to Ask what to do, and for the experiments on this page, always click "Keep" when you open a test image.
- Click "OK" to close the Color Management dialog.
Experiment 1: Monitor profiles and colors
The purpose of this experiment is to let you decide for yourself whether the following claims are true:
- Small changes in your monitor profile make visible and even grossly obvious changes in the colors you see on your monitor screen.
- Your eyes quickly adapt to the new colors — after a few seconds or a minute, colors that looked unbelievably ugly start to look normal.
Step 1: Open Gimp if it's not already open. Make sure the Color Management settings are still set up as above. Then open the following test images from the "Color Management Experiment Kit":
These three test images demonstrate that using the right monitor profile is especially important when judging the appeareance of highlights, bright, delicate colors; and neutral gray or white.
This experiment doesn't involve editing the test images, so you can minimize the Layers window and the Toolbox to free up some screen real estate. Also — and only for those of us who are a bit OCD — under View you might want to uncheck showing the Rulers and Layer Boundary, and then Shrink Wrap the image so there are minimal distracting elements surrounding the image.
Step 2: Move the test images to one side of your screen and (re-)open the Color Management dialog ("Edit/Preferences/Color Management"). Position all the windows so you can make changes to the Color Management settings while observing what happens to the test images (this might take a little hand-eye coordination, but you'll get the hang of it very quickly).
Why is it important to look at the image while changing the monitor profile? From your brain's point of view, changing the monitor profile is very much like changing the color of the light falling on the original scene, which means your eyes will quickly accomodate to the new "lighting". In fact, your eyes accomodate so quickly that you'll want to change back and forth between sRGB and the alternative monitor profiles several times, always keeping an eye on the image as you change the monitor profile, and focusing not just on different test images, but sometimes on specific colors in each test image.
Step 3: Watch what happens as you try different monitor profiles. Note: changing your monitor profile does not alter the actual image RGB values; it only alters the way those values are displayed on your monitor screen.
A neat thing about Gimp is that once you've selected a new monitor profile for the first time, Gimp remembers it, which makes it very easy to flick through from one monitor profile to another. When using Gimp, the experiment goes quickly because you don't have to click "OK" to see the resulting changes in the image colors. Just keep the Color Management settings dialog open and keep switching from one profile to the next as you examine the test images. (Caution: make sure you only have one instance of Gimp open and running, because changing the monitor profile in one open instance of Gimp doesn't change it in the others.)
The "Color Management Experiment Kit" includes the following monitor profiles (for now, ignore the remaining profiles):
- sRGB-IEC61966.icc — the original 1998 sRGB profile.
- sRGB-derived synthetic monitor profiles — each profile simulates one of the ways real monitor profiles can and do differ from sRGB:
- NEC-2190uxi-monitor.icc — custom-built profile for a good quality LCD monitor.
- low-end-laptop.icc — custom-built profile for a low-end laptop screen.
- If your internet search turned up a profile for your monitor, add it to the list.
- your own custom monitor profile, if you normally use one
Click "Edit/Preferences/Color Management" and confirm that the currently selected monitor profile is "sRGB IE61966-2.1 (Equivalent to www.srgb.com 1998 HP profile)". Leave the Preferences dialog open (don't click "OK") and look at the three test images. Now select "srgb-less-intense-red.icc" as the new monitor profile (again, leave the Preferences dialog open) and watch what happens to the test images. Then do the same with "srgb-more-intense-red.icc".
As you change monitor profiles, watch the baby's skin tones and the pool water behind the baby. Also watch the sky and water reflections behind the canoist and the water surrounding the swan. You might want to alternate back and forth between the standard sRGB profile and the "srgb-less-intense-red.icc" and "srgb-more-intense-red.icc" profiles several times, each time watching a different test image.
Which "modified red" profile makes the highlights look more red than the real sRGB? Which one makes the highlights look less red (which is to say more cyan, cyan being the opposite of red) than the real sRGB? Which profile makes the colors more saturated? Which profile makes the colors less saturated?
Extra credit: Change your monitor profile back to the real sRGB ("sRGB-IEC61966.icc"). Select "Swan-at-Montreat.jpg" and go to "Image/Mode". See where it says "Assign Color Profile" and "Convert to Color Profile"? Select "Assign Color Profile" and
Repeat with the remaining pairs of sRGB-derived synthetic monitor profiles: "srgb-less-intense-green.icc" and "srgb-more-intense-green.icc", and with srgb-less-intense-blue.icc and srgb-more-intense-blue.icc. For reference purposes, the opposite of green is magenta and the opposite of blue is yellow. Then try the NEC monitor profile, the low-end laptop profile and any other real monitor profile you might have found.
Sometimes the changes are subtle, so it helps to go back to "sRGB-IEC61966.icc" before switching from one sRGB-derived monitor profile to the next. You can see the colors change, and at first they look odd, or at least different. If you stare for a while, do the new colors start to look normal?
If there is any doubt in your mind as to the actual colors in these test images, check with the Gimp eye-dropper:
- The pool water behind the baby is a light, bright cyan (not very saturated and just slightly more blue than green), and in case you are wondering, the baby's delicate skin is not the least bit sun-burned.
- The sky behind the canoist shades from a very light, bright blue in the upper right corner, down to solid white at the left horizon; no part of the sky is magenta, red, green, cyan, or yellow; and neither are the reflections in the water behind the canoist.
- The water surrounding the swan is almost perfectly neutral gray; no part of the water is magenta, red, green, blue, cyan, or yellow.
Screenshots for Steps 3 and 4:
Step 4: Close the test images that are already open, open the next three test images, and repeat the first three steps. The first three images were chosen to emphasize highlights and bright, delicate colors. These three images emphasize mid-range tonality and more saturated colors:
The "Colorchecker brush strokes" and "Paint chip"s colors are moderately to very saturated. The colors in the "Painted wood balloon" image are very close to the Colorchecker colors, but now the colors have a context. Real "sunny day blue sky" colors are usually very slightly on the cyan (rather than magenta) side of blue; the blue sky in this picture is a reasonably believable "sky blue" shading to more pronounced cyan at the horizon. Your experience may vary, but to me some of the test monitor profiles make the sky in the Painted wood balloon picture almost painful to look at (at least until my eyes have adjusted).
Because your eyes adjust so quickly to changes in how colors are displayed on your monitor screen, it's especially important with the "Colorchecker brush strokes" and "Paint chips" image that you look at the image while changing the monitor profile. You might even want to flick your way through the monitor profiles while focusing on only one color at a time.
Extra credit: When the image profile matches the monitor profile
Background: sRGB isn't just a monitor profile. It's also the color space for most images that are displayed on the web or that come from a point-and-shoot camera. And it's a well-behaved working space, that is to say, a profile you might use when editing an image. It can also be an output profile, that is to say, the color space you might use for an image that you send to a printer.
All the test images in the "color management experiment kit" are actually in the sRGB color space and also have the sRGB profile embedded in their metadata. So when you use "sRGB-IEC61966.icc" as your monitor profile, then the test images have the same profile as your monitor. So the question is, when the image that you are viewing on your monitor has exactly the same profile as you've assigned to the monitor itself, what happens?
Procedure: Pick one of the test images, go to "Preferences" and select "srgb-less-intense-red.icc" as the monitor profile. Then assign "srgb-less-intense-red.icc" to the image itself. How do the resulting colors look, compared to when you've selected the real sRGB ("sRGB-IEC61966.icc")? Now go to "Preferences" and select "srgb-more-intense-red.icc" as the monitor profile, and then assign "srgb-more-intense-red.icc" to the image itself.How do the re Procedure: Change your monitor profile back to the real sRGB ("sRGB-IEC61966.icc"). Select one of the first three test images and go to "Image/Mode". See where it says "Assign Color Profile" and "Convert to Color Profile"? Select "Assign Color Profile" and assign "srgb-less-intense-red.icc". Click "OK", and watch what happens to the colors: do the colors get more red, or less red (which is to say, more cyan)? Repeat using "srgb-more-intense-red.icc". Now try the remaining pairs of sRGB-derived synthetic profiles. Note: assigning a profile doesn't change the image RGB numbers. Instead it changes the meaning of the numbers.
Theory behind the first experiment
Back in the 1990s (the stone age of digital imaging), CRT monitor manufacturing collided with a growing consumer demand for seeing digital images with more accurate colors. Hewlett-Packard and Microsoft proposed sRGB as the solution.
Phosphors and primaries: To make colors, CRT monitors use red, green, and blue phosphors. Although technology keeps changing, most affordable LCD monitors (as of late 2012) also use phosphors. However, (i) LCD monitors don't use the same set of phosphors as were used in the old CRT monitors, and (ii) the phosphors used in LCD monitors vary from one make and model to the next. Truth be told, even CRT monitors didn't all use exactly the same sets of phosphors.
Why should you care about the phosphors used in monitors? sRGB is an example of a matrix monitor profile. Matrix monitor profiles have red, green, and blue channel primaries which represent the reddest, greenest, and bluest colors that the monitor can display. These channel primaries are determined in large part by the monitor's red, green, and blue phosphors.
Is sRGB the right profile for your monitor? The sRGB channel primaries are based on phosphors that were used in consumer-grade CRT monitors manufactured in the 1990s. When you use sRGB as your monitor profile, your image editing program assumes that sRGB accurately describes how your monitor displays colors. So if your monitor isn't calibrated to match sRGB (within very strict limits, calibration can make your monitor act as if it has different phosphors than the ones it really has), then your monitor isn't showing you the right colors.
Circles inside the horseshoe: The picture above is a representation of the sRGB color space profile. The brightly-colored horseshoe shape is a kind of footprint of all the colors that the average human can see. The red, green, and blue sRGB channel primaries are shown by the three small white circles with black dots in the middle. The primaries are connected by three lines which form a triangle. That larger circle in the middle of the triangle is the sRGB white point, but we won't be paying much attention it.
Phosphors and primaries: Remember I said that monitor primaries are determined by the type of phosphors used in the monitor? The numbers around the perimeter of the horseshoe represent wavelengths of light, and you can guess that the three phosphors used in the original CRT monitors on which the sRGB color space was based probably emitted light with wavelengths that peak in intensity near the corresponding primaries. Notice that I'm hedging a bit here: I'm not an expert in monitor technology, and no doubt many factors in addition to the actual monitor phosphors determine the resulting monitor profile primaries.
Phosphors and displayable colors: The sRGB color space holds — and a true sRGB monitor can display — all the colors that are inside the triangle determined by the three sRGB primaries, but it doesn't hold — and a true sRGB monitor can't display — any of the colors that are outside the triangle. To hold more colors, the primaries would need to be farther apart. Which means the phosphors would need to be different. (Lest there be any confusion, the footprint shown above is a graphical representation — it doesn't actually show you any of the real-world colors that fall outside the sRGB triangle.)
The sRGB-derived synthetic monitor profiles: I created the sRGB-derived synthetic monitor profiles by slightly modifying one primary per profile. In the following images, the location of each profile's modified primary is shown by the larger black circle with the white spot in the middle .
The three monitor profiles pictured above are identical to sRGB, except one primary is slightly less intense: the new primary is inside the sRGB triangle. Low-end LCD displays, and especially low-end laptop displays, usually have one or more primaries that are less intense (sometimes much less intense) than the corresponding sRGB primaries.
The three monitor profiles pictured above are identical to sRGB, except one primary is slightly more intense: the new primary is outside the sRGB triangle. Reasonably good and high-end LCD displays have primaries that are more intense than sRGB. Wide-gamut LCD displays have a green primary that is much more intense than the corresponding sRGB green primary.
sRGB compared to the NEC monitor profile and the low-end laptop monitor profile:
All chromaticity diagrams on this page are derived from and subject to the same licence as "CIE 1931 xy chromaticity diagram showing the sRGB gamut and the D65 white point". Licence: CC-BY-SA. Source: Wikimedia Commons.
Experiment 2: Monitor profiles and tonality
OK, this is where we separate the timid from the brave. This experiment is just as easy as the last experiment. The procedure and setup are identical and the results are just as obvious. But some of the vocabulary necessarily gets a bit technical, so my advice is ignore the technical stuff and instead focus on what happens on your monitor screen.
The purpose of this experiment is to let you decide for yourself whether the following claims are true:
- On an LCD monitor, using sRGB as your monitor profile makes it look like deep shadow detail has been crushed to black.
- On an LCD monitor, using sRGB as your monitor profile makes it look like an image has more "pop", more contrast and more saturated colors than it does in reality.
Caution: There are two completely opposite monitor conditions where this experiment will fail: You might not be able to see any patches at all. Or you might see all the patches, but the background might look gray instead of black, and all images (not just "Twenty-five gray patches") look washed out until you give them the equivalent of a hefty black point correction using some combination of Levels and Curves. In either case, you really shouldn't be using that monitor for image editing until it's been recalibrated or replaced if calibration can't correct the problem. (In our household, neither of our LCD monitors have either problem, but we have an old CRT monitor that has the former problem, and we used to have two rebranded Trinitron CRT monitors with the latter problem.)
Update: Time passes and technology marches on, so if you can see all the patches, and the background looks black, and most images on the web look sufficiently colorful with nicely dark shadow tonality, then maybe you are lucky enough to be using a new monitor with the type of technology that provides a true "zero" monitor black point. In any event, on with the experiment:
Step 1: Set-up. Open Gimp if it's not already open. Make sure the Color Management settings are still set up as above. Open "Solid-black-background.png" and maximize it so it fills your monitor screen. Then locate, but don't (yet) open the test image named "Twenty-five-gray-patches.png".
As you might expect from the file name, "Twenty-five gray patches" is composed of twenty-five gray patches against a solid black background. The patches are arranged in five rows that get brighter from top to bottom, and in each row the patches gets brighter from left to right. The first patch in the top row has RGB values of (1,1,1). The last patch in the last row has RGB values of (25,25,25).
The question is, which is the least bright gray patch that you can see when you open "Twenty-five gray patches"? It's very likely that at first you won't be able to see any of the patches in the first row, probably not in the second row, and maybe not even in the third row. However, as you continue looking at the image, your eyes will adapt and you'll start to see more patches.
Factors both intrinsic and extrinsic to your monitor determine how many patches of the twenty-five gray patches you'll be able to see. Non-monitor-related factors include how brightly lit your workspace is, whether your monitor screen is clean or dirty, and whether there are other bright or white colors displayed on your screen alongside "Twenty-five gray patches". Two things about your actual monitor affect how many patches you'll be able to see:
- Your monitor's black point, which is the "darkest dark" your monitor can display.
- Your monitor's (calibrated or native) tone response curve ("TRC"). Your monitor's TRC determines how fast the colors on the screen gets brighter, in response to larger RGB numbers in an image that's displayed on the screen.
Note: Your monitor profile doesn't change your monitor's actual black point or TRC. Rather your monitor profile is supposed to describe your monitor's black point and TRC. I was surprised to find out (I've been doing these experiments myself as I write them up) that the monitor profile's TRC is at least as important as the monitor profile's black point in determining how many patches I can see in the "Twenty-five gray patches" image.
Step 2: Make sure "Solid-black-background.png" is still filling your monitor screen, with the "Preferences/Color Management" dialog set to "sRGB-IEC61966.icc". Then open "Twenty-five gray patches" and locate the first gray patch that you can easily see under normal viewing conditions. Eye-dropper that patch to find the patch's RGB values.
Let's say the first patch that you can easily see eye-droppers as (9,9,9). Practically speaking, this means that any time you look at an image on your monitor screen, all the image pixels that are darker than (9,9,9) might as well be (0,0,0), because all the pixels with RGB values from (0,0,0) to (9,9,9) look exactly the same: they all look solid black.
Step 3: Try some monitor profiles with black points that are higher than zero and/or have TRCs (tone response curves) that are close to the gamma=2.2 TRC.
Watch the "Twenty-five gray patches" image as you switch between the following monitor profiles (all of the profiles are in the profiles pack, except of course any profiles you downloaded from the internet):
- sRGB-IEC61966.icc (the real sRGB profile: zero black point + the sRGB TRC)
- srgb-with-gamma-22-tone-response-curve.icc (derived from sRGB: zero black point + exactly gamma=2.2 TRC)
- sRGB-non-zero-black-point.icc (derived from sRGB: higher than zero black point + approx. gamma=2.2 TRC)
- NEC-2190uxi-monitor.icc (custom-built NEC monitor profile: higher than zero black point + approx. gamma=2.2 TRC)
- low-end-laptop.icc (custom-built profile for a low-end laptop: zero black point + approx. gamma=2.2 TRC)
- edid-NEC-2190uxi-profile.icc (NEC monitor profile magically pulled from the monitor itself: zero black point + exactly gamma=2.2 TRC)
- Any model-specific monitor profile that you downloaded from the internet will most likely will have a roughly or exactly gamma=2.2 TRC, and might have a higher than zero black point. Of the handful of profiles that I downloaded and examined, profiles for the higher-quality monitors tend to have higher than zero black points.
- your own custom monitor profile, if you normally use one
What's the dimmest patch you can see using "srgb-with-gamma-22-tone-response-curve.icc" or "sRGB-non-zero-black-point.icc" as your monitor profile, compared to using the real sRGB profile? How about when you use "NEC-2190uxi-monitor.icc" vs "edid-NEC-2190uxi-profile.icc" as your monitor profile? On my own monitor, the dimmest patch I can see remains the same with "srgb-with-gamma-22-tone-response-curve.icc", "sRGB-non-zero-black-point.icc" and both NEC monitor profiles, but the brightness of the other patches changes from one profile to the next.
Step 4: Shadow detail, saturation and contrast in real images. Does any of this stuff about monitor black points and tone response curves have implications for viewing and editing real images?
Open the following test images (and if necessary reopen the Gimp Color Management settings) and then look at the images while flicking back and forth between the real sRGB and the profiles that have roughtly gamma 2.2 TRCs and higher than zero black points:
How much does the appearance of these images alter when you switch back and forth between "sRGB-IEC61966.icc" (the real sRGB), "srgb-with-gamma-22-tone-response-curve.icc", and "srgb-non-zero-black-point.icc" as your monitor profile? And how about when you switch between the two NEC monitor profiles? What happens to the deep shadow detail in "Branch over lake" and "Lakeside path"? What about the not-so-deep shadows on the baby's face? Does "Painted wood balloon" seem to have more "pop" — that is, more contrast and more saturated colors when using the real sRGB as compared to the other profiles?
All this is well and good, but the important question is, would you edit your images differently if you knew that your monitor is more accurately described by a monitor profile with a higher-than-zero black point and a roughly gamma 2.2 TRC, than it is by sRGB?
The "effective black point difference" between sRGB and the test profiles on this page is only around 10 points on a scale from 0 to 255. In actual practice the difference between sRGB and any given real monitor's black point/TRC can vary considerably more than the relatively small difference in the test profiles in the test profile pack. I've seen real monitors where the effective black point difference is 30 or more points for an 8-bit sRGB image.
To give yourself an idea of how much difference 30 points is between what you see and what you really have, open "Branch over lake" if it's not already open, open "Levels", and move the "Input Levels" slider from 0 to 30. Then to simulate the difference in the opposite direction, hit "undo" and move the "Output Levels" slider from 0 to 30:
Now imagine exchanging images with a friend whose monitor deviates from sRGB in the opposite manner as your own: the two of you could be looking at images with 60 points difference in the effective image black point as displayed on the monitor screen (been there, done that, took a long time before we figured out that our monitors weren't showing us the same colors and tonality).
Theory behind the second experiment
Black points: The sRGB color space is based on the display characteristics of CRT monitors. Back in the day, any decent and properly calibrated CRT could display true black (zero light emitted from the monitor screen), which is why sRGB has a "zero light" black point. If you were using the sRGB profile with a true sRGB monitor, you'd be able to see the first patch on the first row (or at least the second patch, if we allow for unavoidable flare and glare off the face of the monitor).
Unlike the old CRTs, an LCD monitor's "darkest dark" is a certain percentage of its "brightest bright", and there is nothing you can do to make that percentage lower than it is. Which means that monitor profiles with a zero black point don't accurately describe LCD monitors.
Tone Response Curves: To the right is a graph showing a gamma=2.2 TRC (the blue curve) superimposed over the 1024-point sRGB TRC (the red curve). Both curves start at (0,0) (lower left corner). The blue gamma 2.2 curve is below the red sRGB curve until they cross at around 26214 on the x-axis. Then the blue gamma 2.2 curve is above the red sRGB curve until the two curves meet again at 65535. It might look like the two curves are essentially identical, but as you have already seen in the experiments, the seemingly small differences lead to a very noticeable difference in what gets displayed on your monitor screen.
The premise behind these experiments is that without calibrating and/or profiling your monitor, you don't know what your images really look like. The experiments were designed to show how very small differences in monitor profiles produces very visible alterations in the colors displayed on your monitor screen.
So if seeing is believing, does your monitor profile matter? Are you sure that sRGB is a "good enough" profile for your monitor?
If you can't make or find a custom monitor profile that's more accurate than sRGB, all is not lost: rely more on your eye-dropper and less on your eyes when you edit images. Use your eye-dropper to check color balance in the highlights. And when it comes to setting tonality (Curves and Levels), think in terms of "Zones" and corresponding RGB values rather than "what looks good" on a monitor that you can't trust.
Part 2 of this Color Management Experiment Kit (not yet written) explores the remaining entries found in the Color Management settings dialog: black point compensation, relative and perceptual intents, and softproofing mismatched color gamuts. The same format will be followed: do the experiment and watch what happens; any requisite theory follows the experiment.