A little color test

gordo

Well-known member
I got inspired by the test that Magnus conducted so...
I assume that when you do a screen capture/grab that image reflects what you see on the screen.
If that's true I would like you folks to download the attached image and open it in Preview on the MAC or Preview's equivalent on the PC.
I'd like an indication as to what you see on your screen
Take a screen capture/grab and email it to me at pritchardgordon @ gmail (dot) com
Indicate whether you are on a PC or a MAC, whether your display is calibrated or not, and what application you used to view the image.
I'll collect the images and post them for comparison purposes (they will be anonymous)

The more samples - the better! (I'd like at least nine.) There's no processing involved so this is not a test in any way.
 

Attachments

  • Test.jpg
    Test.jpg
    323.3 KB · Views: 203
OK folks...thanks to those that participated in this exercise. I hope you all have a prediction as to the result.

Please see attached screen grab - it shows the result in a small enough file format for this test.

Only one image actually looked different than the others (2nd row 1st image) it was visibly slightlyless saturated (which suggests that the test was valid). But the rest were effectively the same despite being from different displays, calibrated or not calibrated, and different operating systems.

IMHO this basically shows that unattended color management is more than good enough to prepare images with some degree of confidence that what I see you will see. And if your proofs look like what you see and your presswork looks like your proofs then life is good.

If I continue to see what I want to see as the image goes through page layout and PDF creation then isn't that the right way?

Is color management a mess or has it finally disappeared as some would say that it rightly should?

Or is this test totally off-base? If so in what way?
 

Attachments

  • Test results.jpg
    Test results.jpg
    805.3 KB · Views: 217
OK folks...thanks to those that participated in this exercise. I hope you all have a prediction as to the result.

Please see attached screen grab - it shows the result in a small enough file format for this test.

Only one image actually looked different than the others (2nd row 1st image) it was visibly slightlyless saturated (which suggests that the test was valid). But the rest were effectively the same despite being from different displays, calibrated or not calibrated, and different operating systems.

IMHO this basically shows that unattended color management is more than good enough to prepare images with some degree of confidence that what I see you will see. And if your proofs look like what you see and your presswork looks like your proofs then life is good.

If I continue to see what I want to see as the image goes through page layout and PDF creation then isn't that the right way?

Is color management a mess or has it finally disappeared as some would say that it rightly should?

Or is this test totally off-base? If so in what way?

I was thinking about this and at first I wasn't sure what you were trying to do. I think you were trying to show that for some conditions, colour works well without too much colour management fuss.

Thinking more about this I was wondering if indeed you were off-base. You will have to tell me if my following comments make any sense.

OK. You post an image. We see an image. Some people copy the image with the screen grab function and send it to you and you look at the images. They tend to look the same to you.

What are images? We see something but the images are basically data. You send data, some people copy that data and send that data back to you and your system sees the images look basically the same. Why not if all the images are the same data.

We see images but there is nothing to say what we see is the same as what anyone else sees because our display devices might show the same data differently. You would need to see all the display devices side by side to determine if they look the same but I don't see how that can be done with this test.

That may be the off-base situation. I don't know for sure but it is suspicious.

I do think that now with more modern computers and displays etc., there is probably a lot less to worry about with colour management due to more consistent devices and the greater ease to specify a colour with the rgb output.

I think most people do not remember the early colour TVs that had problems obtaining a predictable colour and required colour adjustments. Kind of like an offset press operator. :)

Offset printing is quite different due to the non linearity and not independent nature of the CMYK channels. Added to that, the inconsistent print conditions around the plate cylinder which adds to the problem. This complexity which is partly related to poor press design has probably resulted in inadequate and complicated methods being developed in an attempt to manage colour, which now seems to be applied to digital printers.

Moral of the story. Improve the capability of the print device and colour management becomes easier.
 
I was thinking about this and at first I wasn't sure what you were trying to do. I think you were trying to show that for some conditions, colour works well without too much colour management fuss.

Thinking more about this I was wondering if indeed you were off-base. You will have to tell me if my following comments make any sense.

OK. You post an image. We see an image. Some people copy the image with the screen grab function and send it to you and you look at the images. They tend to look the same to you.

It's not that they look the same to me - it's that it appears that the image looks the same to them.

What are images? We see something but the images are basically data. You send data, some people copy that data and send that data back to you and your system sees the images look basically the same. Why not if all the images are the same data.

The data that I sent is the same. It's the representation of that data that I expected to differ. RGB values (the data I sent) do not represent color. They represent tone values. The hues that are modulated by those tone values generate the color. So, if the RGB hues of your display are different than mine the color you see created by the same RGB values will/should be different.

We see images but there is nothing to say what we see is the same as what anyone else sees because our display devices might show the same data differently. You would need to see all the display devices side by side to determine if they look the same but I don't see how that can be done with this test.

Correct, by assumption was that a screen capture would allow the comparison because the screen capture represents the data as displayed. If the screen captures all look the same then doesn't that mean that the displays show the same result of the data?


That may be the off-base situation. I don't know for sure but it is suspicious.

I do think that now with more modern computers and displays etc., there is probably a lot less to worry about with colour management due to more consistent devices and the greater ease to specify a colour with the rgb output.

Not just more consistent and with greater ease but with a greater homogeneous representation of the image across different displays despite their apparent core differences (e.g. calibrated vs uncalibrated). That's the important bit because if true then color management gymnastics (cost and effort) provide very little added benefit.

I think most people do not remember the early colour TVs that had problems obtaining a predictable colour and required colour adjustments. Kind of like an offset press operator. :)

That's true for way back computer displays - color TVs (broadcast signal and displays) still have issues.

Offset printing is quite different due to the non linearity and not independent nature of the CMYK channels. Added to that, the inconsistent print conditions around the plate cylinder which adds to the problem. This complexity which is partly related to poor press design has probably resulted in inadequate and complicated methods being developed in an attempt to manage colour, which now seems to be applied to digital printers.

I just wanted to see if print suppliers would see on their softproofing displays basically what I see on my display. If my test is valid - it appears they do.

Moral of the story. Improve the capability of the print device and colour management becomes easier.

Methinks that's a different issue.
 
The screen capture doesn't capture what we see.
I did a small test and changed the settings on my monitor.
And of course that was not captured. How could it?
Comparison of a photo of my monitor and a screen capture
Photo vs Screen Capture.jpg
 
Last edited:
The screen capture doesn't capture what we see.
I did a small test and changed the settings on my monitor.
And of course that was not captured. How could it?
Comparison of a photo of my monitor and a screen capture
View attachment 4476

Dang!

So what is a screen capture capturing if it's not capturing what is displayed on the screen?
 
A screen capture captures whats displayed on the screen but it doesn't capture what I see on on the screen.
Only a photo can do that. The example I provided is the exact same picture.
The one on the left was captured with my phone, the one on the right is a screen capture of the same picture..
 
A screen capture captures whats displayed on the screen but it doesn't capture what I see on on the screen.
Only a photo can do that. The example I provided is the exact same picture.
The one on the left was captured with my phone, the one on the right is a screen capture of the same picture..

Unfortunately a camera can not capture accurate colour. So there are some problems there too.
 
Last edited by a moderator:
As I understand it, and I'd welcome corrections, the files you got back will show what RGB values are being sent to the monitor, not the effect a monitor profile has on how they're displayed.

I think that taking a tagged image, opening it in Photoshop, converting to and assigning the monitor profile, saving it tagged with that monitor profile, then displaying all the results in Photoshop on a calibrated monitor, would be closer to the test you're trying to perform.
 
As I understand it, and I'd welcome corrections, the files you got back will show what RGB values are being sent to the monitor, not the effect a monitor profile has on how they're displayed.

I think that taking a tagged image, opening it in Photoshop, converting to and assigning the monitor profile, saving it tagged with that monitor profile, then displaying all the results in Photoshop on a calibrated monitor, would be closer to the test you're trying to perform.

More complicated but if you, Magnus, and others would like to do that I'd be happy to rewind the tape and try again.
 
Hi Gordo, interesting test!

I think this test is impossible to execute without a spectrophotometer or colorimeter. You need to measure this to get useful data. A screenshot wouldn't tell you anything about how the image look on the specific display, not even give you an indication. If the displays are correctly calibrated you could get some data by doing it the way Danny suggests.

This is how I would do it:

1. Locate the current monitor ICC-profile (Profile A)
2. Make a new monitor profile using a spectrophotometer and i1Profiler or similar (Profile B)
3. Convert the test image to profile A
4. Assign profile B
5. Convert to Adobe RGB or similar (Absolute conversion)
6. View the different results on a calibrated reference monitor

You could also Install a monitor certification software like UDACT or similar and collect the results.

Best regards,
 
Last edited:
Hi Gordo, interesting test!

I think this test is impossible to execute without a spectrophotometer or colorimeter. You need to measure this to get useful data. A screenshot wouldn't tell you anything about how the image look on the specific display, not even give you an indication. If the displays are correctly calibrated you could get some data by doing it the way Danny suggests.

I think you're missing the point a bit.
There is a tyranny in numbers. I wasn't trying to get useful data, I was trying to get, in as simple way as possible, an insight into the relative visual differences between display - something like what's done with TVs in shop displays.

TVS_zps12c080b5.jpeg


Of course they can be optimized and/or calibrated and/or brought to a common appearance - but that's not what I was looking for.

I have a feeling that most computer displays are similar enough in their display appearance - and given the way the eye/brain works, that color management gymnastics are not needed. I have only my experience as an indicator so I was looking for a simple way to demonstrate what I thought was going on. Like I said, I've been doing color for reproduction, for thousands of images, on desktop computers for over 30 years and have only had a problem once (and that was my fault for not checking the highlight numbers). I take photos, like many people, as a hobby ( Gordon Pritchard's Photography ) and am a member on a photo enthusiast forum. People ask for critiques, enter contests, discuss image quality - all with an underlying unspoken assumption that what they are seeing on their displays (PC/MAC/calibrated/uncalibrated/color managed or not) is basically what the original photographer sees on their display when they post the image. Indeed, the whole process only works if the images look as intended despite the vagaries of computer displays.
 
I think you're missing the point a bit.
There is a tyranny in numbers. I wasn't trying to get useful data, I was trying to get, in as simple way as possible, an insight into the relative visual differences between display - something like what's done with TVs in shop displays.

If you follow steps 1-6 in my previous post the result will be what you are looking for, a visual comparison. The word simple is relative. :)

I do not know any other way of doing it, unless you gather all monitors in the same room as in your example photo.

I have a feeling that most computer displays are similar enough in their display appearance

That depends what you are doing. In my world they are not similar enough.

Best regards,
 
Last edited:
I’m not too experienced with Colour Management but for the purposes of this test could we have everyone (those that have them) take measurements on the actual monitor with a spectro of patches or areas and compare them to each other? Profile or no profile? The Spectro is basically taking the place of the human eye and would record the variations in “visual” appearance.
 
I’m not too experienced with Colour Management but for the purposes of this test could we have everyone (those that have them) take measurements on the actual monitor with a spectro of patches or areas and compare them to each other? Profile or no profile? The Spectro is basically taking the place of the human eye and would record the variations in “visual” appearance.

Yes but I think Gordo wants a visual comparison.
 
That depends what you are doing. In my world they are not similar enough.

I like the photo you posted. So I tried something similar in my den (not as sophisticated as your environment). Here is a snap of a 42" LG TV vs my 15" MacBook Pro - both are uncalibrated/color managed:

AsseenonTV_zps580a41e3.jpg


The artefacts on the TV are camera sensor moiré - not visible in real life - due to the lower resolution of the TV relative to the MAC display at that distance.

The biggest differences that I saw was that on the TV the image looked oversharpened - that's probably due to the setting on the TV.

Virtually no visible difference in grey balance or hues. I think most people who saw this in my den would say that the images were identical.

Keep in mind that this is the default condition of the displays - no manipulation or adjustments.

If this is true between these two very different displays/devices I would think that there would be even less difference between devices within a product range. Now there may be differences that an instrument might indicate, but from a subjective appearance point of view, as far as I can see, there is no important difference.
 
I hope I was clear in my previous post that both displays are uncalibrated anD not color managed.
 
I hope I was clear in my previous post that both displays are uncalibrated anD not color managed.

Of course the sample size for this test is only two. Could be just luck. :)
 

PressWise

A 30-day Fix for Managed Chaos

As any print professional knows, printing can be managed chaos. Software that solves multiple problems and provides measurable and monetizable value has a direct impact on the bottom-line.

“We reduced order entry costs by about 40%.” Significant savings in a shop that turns about 500 jobs a month.


Learn how…….

   
Back
Top