Poor man's calibration process

[Ndukes]

Having struggled many times with finding a reliable method of maintaining correct calibration from camera through monitor to printing I recently developed my own ‘cheapo’ method which really seems to work.
I started by purchasing a ‘Camera Trax 24 Color Card’ (£11.95 from Amazon UK). I then photographed the 24 Color Card using ambient daylight in the same room where I process my images. The camera selected 6,000k equivalent on auto white balance and I processed the image ‘straight out of camera’.
The following image shows the photographed card on the monitor with the actual card resting at bottom of the screen. It is essential that the ambient light intensity and colour hitting the card matches the brightness setting of the monitor. To my absolute delight the colour balance on my Samsung SyncMaster T240 was ‘right’ with all three colours level.

I also re-created my own version of the card in Photoshop CS6 using the correct colour references for each colour (https://en.wikipedia.org/wiki/ColorChecker) and checked this against the original card and the photographed version. At this stage I was able to calibrate my monitor reasonably satisfactorily but was not satisfied with the subtle pastel and darker shades and wanted to move on to calibration of my printer.
So I developed my own re-created card into a version with nine levels of luminosity for each colour and proceeded with the next step, checking and calibrating my printer to match my monitor with the colours in the original 24 Color Card.

(I am using an Epson Stylus Photo R1800 and have never succeeded in getting satisfactory results printing through Photoshop using available prescribed profiles. I have always used printer colour controls for better results.)
The following image shows the re-created version displayed on the monitor together with two prints of the file, on Epson Archival Matte and on Premium Glossy paper and using Epson inks, resting on the bottom of the monitor. To achieve such close matching I had to run a number of test prints at various settings. It is important to note that the actual 'white' colour of the paper varies with paper type as can be seen here.

I now keep both of the printed cards beside my monitor and always check the cards against the displayed card image file on my monitor before any processing session.
The main lesson for me throughtout this exercise is the importance of constant ambient lighting when processing or printing images. I have adjustable venetian blinds on the window beside my desk and, although my study does not get direct sunlight, I try to avoid working on sunny days and in artificial light.
Sadly, Ireland is good for overcast grey skies but such depressingly grey and neutral sky does after all have its silver lining.
Hope some find this helpful.

[ST1]

Interesting write up of your process Neville. Colour management and Calibration especially end to end (camera, screen to printer) as you've described is undoubtably the way to go for those who print at home-office. Well done Neville

[Ndukes]

Thanks Peter. It's a method that works well for me.

[Manfred M]

Neville - your approach is not particularly different than a number of other ones that have been tried over the years by many players including Microsoft and Apple. They all do not work particularly well because they rely on human vision to do the calibration.

There are a couple of things working against you:

1. Not all humans see colours equally well. While colour blindness is a well known phenomenon, it is just a more extreme example of the variability we have. In general, there is a gender difference; females generally have better colour vision than males. In the "old" days, photo labs preferred females over males for this reason. Colour vision is also age dependent; our ability so detect tonal variations decreases as we get older.

2. The second factor is human physiology. Our eye / brain visual system has evolved to see things as looking "normal" over a wide range of colour temperatures. This adjustment starts within as little as a tenth of a second and stabilizes within a second or so. This means the longer we look at something, the less likely we will be able to see colour casts.

So, I'm glad this approach seems to work for you. On the other hand, I suspect that repeatability and consistency might be more variable than you think.

[wobert]

Test your own color vision here-
http://www.xrite.com/online-color-test-challenge

[Simon Garrett]

In addition to the potential difficulty of human assessment of colour, I think there might be one or two more potential issues here.

If I understand your method correctly, you are getting the digital version of the colour chart by photographing a colour card with your camera.

How do you know that your camera is making accurate colorimetric measurements of the colour patches? In general, cameras aren't colorimetrically accurate. You can improve the accuracy by calibrating your camera. You can do this with the ColorChecker Passport, or (assuming the Camera Trax 24 card is colorimetrically accurate) use Adobe's free DNG Profile Editor (https://www.adobe.com/support/downlo...atform=Windows). Instructions are at http://wwwimages.adobe.com/content/d...umentation.pdf. That way you get a profile for your camera.

Then shoot the colour card in raw, and use the profile you've created in ACR or Lightroom.

A further issue - and I can't for the moment think what is the right answer - is how to get the right white balance. You used auto WB, but I'm not sure what the "right" WB would be to ensure colorimetrically accurate colour.

With colour calibrations issues, what you can't do is say "it matches, so it must be right". The danger here is that errors in one part of the system are being cancelled out by errors somewhere else. For example, if the camera is too green but the monitor is not green enough, the colour may match even though the monitor calibration is wrong.

The advantage of using a colorimiter (i1, ColorMunki, Spyder or whatever) is that you've not relying on every element end-to-end being accurate. You're making an absolute measurement on the end result on the screen.

I'm not pretending to be the world's expert, so if anyone thinks my logic is not right, please say.

[xpatUSA]

In addition to the potential difficulty of human assessment of colour, . . . .

I would mention that, when comparing two colors side-by-side (leaving illuminants to one side for now) the eye is quite good at judging similarity. So if I posted two colors and asked if they were "the same" even those with defective color vision would be able to answer quite easily. On the other hand, if I posted a single color and asked y'all to guess what it's hue angle is, in HSB space, answers would likely have quite a variance. Remember the "orange" Porsche?

Kind of like comparing bolt diameters either side-by-side, or first one then the other.

Not disagreeing, Simon, just adding a bit . .

[Ndukes]

........ So, I'm glad this approach seems to work for you. On the other hand, I suspect that repeatability and consistency might be more variable than you think.

Absolutely agree Manfred, and my wife is called into service many times as the critical reference on colour nuances, which bears out what you say about females preferred over males in colour labs.

Many thanks for your input.

[Ndukes]

........ So, I'm glad this approach seems to work for you. On the other hand, I suspect that repeatability and consistency might be more variable than you think.

Absolutely agree Manfred, and my wife is called into service many times as the critical reference on colour nuances, which bears out what you say about females preferred over males in colour labs.

Many thanks for your input.

[Ndukes]

..................

How do you know that your camera is making accurate colorimetric measurements of the colour patches? In general, cameras aren't colorimetrically accurate. You can improve the accuracy by calibrating your camera. You can do this with the ColorChecker Passport, or (assuming the Camera Trax 24 card is colorimetrically accurate) use Adobe's free DNG Profile Editor (https://www.adobe.com/support/downlo...atform=Windows). Instructions are at http://wwwimages.adobe.com/content/d...umentation.pdf. That way you get a profile for your camera.

Then shoot the colour card in raw, and use the profile you've created in ACR or Lightroom.

Thaty's pretty much what I did. My current camera is a Nikon Df and I exclusively use RAW, with Camera Standard 2 as the profile in ACR. The original photo of the color card was taken precisely where my monitor sits, processed immediately after exposure with no adjustments and displayed on my monitor under the same stable ambient lighting. So, I concede it's not perfect, but not bad either.

A further issue - and I can't for the moment think what is the right answer - is how to get the right white balance. You used auto WB, but I'm not sure what the "right" WB would be to ensure colorimetrically accurate colour.

Yes, agreed Simon. But decoration of my study is intentionally neutral and the ambient light which determined the auto white balance setting in the camera is the same as when I compared the photographed card image on screen to the actual card. I must confess that if there was any appreciable difference I would be less confident in the result. All colours are not fully accurately recreated but, as we all accept, that's not possible. But it is certainly close enough.

With colour calibrations issues, what you can't do is say "it matches, so it must be right". The danger here is that errors in one part of the system are being cancelled out by errors somewhere else. For example, if the camera is too green but the monitor is not green enough, the colour may match even though the monitor calibration is wrong.

.............. so if anyone thinks my logic is not right, please say.

True, and all your logic is right. But the result of developing this simple 'poor man's calibration process' is that I am now able to photograph a subject in my study or anywhere else, load up the file and display it on screen and print it with an excellent degree of consistency.

I would also say that the most important benefit is comparison of the printed version of the card with nine luminosity levels with it's version displayed on the monitor. This ensures that ambient light when processing or printing with not adversely affect how I 'see' the image on the monitor and makes for better consistency in processing and printing.

Many thanks Simon for your most helpful input.

[Ndukes]

Test your own color vision here-
http://www.xrite.com/online-color-test-challenge

Thanks for that Robert. Scored 7. (That's probably good enough)