Explore related tutorials & articles
53
Replies
Both the devices you mentioned have their own peculiarities - the i1 from GretagMagbeth is a good, but rather old spectrophotometer, it has poor sensitivity in shadows. It was good on older CRT displays, but on modern monitors can be issues in shadows. Modern software such as DisplayCAL (free software) allows you to take these features into account (there is both compensation for drift in the shadows due to repeated measurement of black patches, and a whole block of settings for black color correction).
The SpyderX is a colorimeter with good sensitivity, but it cannot directly measure the color. The principle of operation of the colorimeter is closer to the camera matrix - it measures the amount of light that has passed through the light filters. At the same time, he does not know anything about the spectral characteristics of the light source itself. That is why at the beginning the software prompts you to choose a monitor model (in order to use the spectrum of its backlight as a base for measurements). The slightest deviations of the reference values from the actual state of the backlight lead to an incorrect color assessment and, as a result, to an incorrect profile.
You can download DisplayCAL and try to build a SpyderX correction table for your specific monitor (first patches are measured with i1, then the same colors are measured with SpyderX) - this will solve both problems. You will be able to use SpyderX to build a profile and still have accurate measurements.
Also I would like to ask you again - upload the monitor profile file.
After viewing the profiles, I noticed that your device is not an i1 but a eye-one display, that is, it is also a colorimeter. But if you are satisfied with the result of his work, you should not focus on this. SpyderX built almost the same profile as the eye-one display, but with a hue of about 10 degrees relative to the white point (that is, at the setup stage before profiling, the monitor type was specified incorrectly, or it is correct, but the reference data does not correspond to the actual state of the display). You may be able to resolve this issue if you select a different display type in the program settings.
Regarding viewing - it is obvious that you see it correctly in Photoshop. I see the same (by the key points in your description) and your screenshots. Here, apparently, the problem is in browsers:
On the left is a browser without any CMS support, on the right is Opera. We can notice that a browser without CMS support shows the image a little differently (more saturated and contrasting) - this is just the native space of my monitor. But Opera (like any browser on the Chrome engine) shows something in between - no longer sRGB, but not yet the colors of my monitor.
We can also conduct the opposite experiment: convert the image to the color profile of my monitor (it is close to AdobeRGB, I have no doubt about its accuracy), then assign the resulting image to sRGB, simulating browser behavior.
On the one hand, we get an image close in lightness to the original file, and on the other hand, we again see the difference in saturation (but in the opposite direction).
It looks like the browser CMS is using the monitor profile as its working color space. I suggest you do the following:
- open Edge, Chrome or Opera
- in the address bar, enter chrome:flags
- find the Force color profile setting, select sRGB there, restart the browser
- open your image in Photoshop, convert its colors to the color space of the monitor (color profile) with the command edit-> convert to profile
- Assign sRGB to this image, save a copy and open it in a browser.
- Compare with the original sRGB file opened in Photoshop.
The images should look the same: (I also measured the colors in the color checker - there are differences, but they are minimal).
Since the work of CMS browsers is based on the color profile of the monitor (and we all perfectly understand that there are no two identical devices), then most likely you will not be able to achieve the same display of photos on different devices and in different browsers. Neither embedding profiles nor changing system settings will help. I would be glad to be wrong, but it looks like it is.
I've tested all of this on a single monitor system. The monitor is used as a color proof device and how accurately it shows color in Photoshop is beyond any doubt. So I'm also leaning towards a bug in the CMS system of Сhrome-based browsers.
Firefox gives a more adequate result, but I still see differences in shades.
Would you mind taking the same screenshot on your system? (open the image in firefox and the same in Photoshop). I use this image during tests
I said above that I sensed something right under my nose that I missed.
Now it's getting clearer. Everybody seems to assume that web browsers don't do full color management, at least not in the way that Photoshop does. They just "use sRGB", which in fact doesn't make any sense.
This is incorrect either way. All major web browsers today do this exactly as Photoshop. The underlying mechanism is exactly the same. A source profile is converted into a destination profile (monitor), and the result of that conversion sent to screen.
And the result is exactly the same - the browsers even have the equivalent of Photoshop's working space. If a profile is missing, a default profile kicks in. For the browsers, that's sRGB. Either way, it's converted into the monitor profile.
So whenever a web browser doesn't match Photoshop, something's wrong. That's buggy behavior by definition. That's color management failing.
Now, I'm not saying anything about the quality of browsers' CMS. You may not get a perfect numerical match in all situations everywhere. But they will be very, very close, and for all practical purposes identical.
I seem to begin to understand what the matter is. Firefox correctly (perfectly correctly) displays files with an embedded profile (left picture). When there is no color space tag in the file, it displays it in much the same way as Chrome-based browsers (right picture).
It's just that your words made me doubt the correct color settings on my system. But now I understand for sure that the matter is exclusively in the browser.
I seem to begin to understand what the matter is. Firefox correctly (perfectly correctly) displays files with an embedded profile (left picture). When there is only a color space tag in the file, it displays it in much the same way as Chrome-based browsers (right picture).
Is seems you are saying that embedding means attaching the entire profile, and tagging means only including the name of the profile.
Where is this difference specified in Photoshop when saving a file?
Isn't embedding the same as tagging?
Isn't embedding the same as tagging?
Sorry, I was in a hurry and wrote it wrong. Of course, we were talking about untagged files and embedding is the same as tagging. I'll correct the message now.
jazz-y, did you set Firefox to mode 1?
In the mode 2 setting, images without an embedded profile are not color managed. All tagged material will always be correctly color managed regardless.
By @D Fosse
No, I used color_management.mode 1. Switch to 2 and now the untagged image looks the same as with the embedded profile. Thanks for the advice!
All browsers now assign sRGB to untagged material (no profile).
So all browsers should display tagged or untagged images the same, and they should look like the preview image in Photoshop with Proof Colors not turned on?
This is not what I'm seeing in my latest tests. Unless I seriously messed something up, which, of course, is entirely possible...
My tests show that only Firefox displays an sRGB image correctly and only when tagged. In this case, the Firefox display matches Photoshop with Proof Colors off. In every other case, the browser display matches Proof Colors > Monitor.
AdChoices