Calibrating your printer, screen, and scanner

[Leif Ohlsson]

This little thread is for Gil Russell, who already years ago, way back in the days of Papermodelers.net, pointed out that I should do something like what I now finally have found it necessary to attempt - calibrate the whole chain from scanner, over screen, to printer - and do it by iteration until the result was acceptable.

Those of you who may have followed the build log of the DHC-1 Chipmunk know that I ran into serious trouble scanning the original printed kit (http://www.papermodelers.com/forum/aviation/16374-dhc-1-chipmunk-wak-model-enlarged-1-16-a-5.html#post241985). Scans which looked good in the scanning preview, and good enough in Photoshop, when printed turned out to be overlaid with a kind of greyish hue, making the colors look as if they were seen under an inch of dirty dishwater.

Not good.

I solved it temporarily by turning up the brightness in Photoshop, and was satisfied enough with the prints produced to build the model. But fundamentally this was a deeply unsatisfactory situation. There must be a way, I thought, to calibrate scanner and printer so that the prints looked like they did on the screen, and the scanner produced a scan which, when printed, looked like it did on the screen.

If you are very eager to know just the advice I could give after trying to solve this problem for myself, jump to the end and summary of this introductory tale (http://www.papermodelers.com/forum/tutorials/16489-calibrating-your-printer-screen-scanner.html#post242191). What follows now is a perhaps rather lengthy description of the crooked road i took to get there…

My starting point thus was that I was satisfied with what the screen looked like, and it seemed logical to start with how to make the printer print what the screen showed. (As it turned out, this was a faulty supposition, but more of that later, in the order I came to understand it.)

For testing, I made up a simple little test print sheet. (The size of it is 210 x 81 mm, which corresponds to the surplus paper strips I get when making my long sheets, "L3" 420 x 216 mm, from regular A3.)

http://www.papermodelers.com/forum/attachment.php?attachmentid=102958&d=1316951412

The test sheet contains standard color samples gracefully provided by Rubén Andres (http://www.papermodelers.com/forum/recolors/12647-re-colouring-ford-trimotor-peter-zorn-32.html#post238983) in his recent magnificent redrawing of the Ford Trimotor, and a couple of typical model parts from a downloaded model. Since all details are downloaded, the goal most be that the print-out of them should look as close as possible to what they look like on the screen.

[Leif Ohlsson]

I now started to print this little sheet, juggling the printer controls for color calibration, brightness, etc. In my setup they look like this, if you bring out the "Advanced" features. You should be able to get something similar in your setup (but don't go running there just yet; it might not be the ultimate solution…):

http://www.papermodelers.com/forum/attachment.php?attachmentid=102959&d=1316951676

At this point, you must understand, I was acting on the supposition that my screen was alright (since I was satisfied with what documents and photos looked like on it), and that the printer had to be adjusted to produce an output similar to what the screen showed. As it were, I ran up a sizeable amount of test prints:

http://www.papermodelers.com/forum/attachment.php?attachmentid=102960&d=1316951676

Then something strange happened. By sheer accident (fortuitous, indeed), a number of ordinary paper slips had sneaked their way into my heap of surplus strips. Must have come from me cutting down some ordinary paper A3 sheets at some time or other. The print-out on this kind of regular offset paper looked much better than on my thick 200g supposedly inkjet papers!

So far I had worked on the assumption that my inkjet paper would best be described as "Epson Matte" in the print out menu, and therefore set the printer to that. The difference when printing by mistake on ordinary paper was amazing - much clearer, much more yellow than I could achieve by adjusting the printer.

Even if it isn't totally logical, I therefore tried the same setting with but changed the paper quality setting to "Plain papers" - and the result was much better, also on the inkjet paper. Another three attempts at changing variables actually went over the top, making the print too light, too yellow. So I reverted to the settings arrived at by accident.

Test printing a full set of the model convinced me that everything was still over the top - too much yellow, too much red, everything too bright and saturated. After that, it was a matter of working myself backwards, to a neutral setting for everything, except the new choice of "Normal paper", even if it really was inkjet, or at least coated paper, I was printing on. That now became my standard setting.

At this stage, I was thus back at a perfectly neutral printer setting, with the difference from my earlier practice that paper quality was set to "Plain papers", even for my thick 200g coated paper.

[Leif Ohlsson]

At this point I recalled some very short advice by Wilfried (http://www.papermodelers.com/forum/aviation/16374-dhc-1-chipmunk-wak-model-enlarged-1-16-a-3.html#post240996), stating that trusting Colorsync always had produced acceptable results for him.

Now, this for me represents a murky area. So far, Colorsync had mostly appeared as a nuisance when opening documents in Photoshop or Illustrator, the computer screaming incessantly about an embedded color profile this, and another one that did not match the color profile of the document, and what did I want to do? Mostly I just OK'd what the programme suggested by showing one button of two or three as being checked.

Obviously, this was not the way to treat Colorsync. Following Wilfrieds hint, I looked through the computer's tool programmes, and certainly, there was something called "Color Sync tool". That, however, seemed far above my competence and furthermore did not seem to contain any serious adjustment options at my level.

So I searched, and read up on what Color Sync really is. Turns out it's a very useful thing. By allocating what's called a color profile to each document, the idea is that the color management will be appropriate throughout a chain of applications, such as a camera, a computer, an application, and a printer.

To do this, however, it is essential that Color Sync is allowed to do so uninterrupted. Which means it should be turned on, of course. Which I hadn't for my printer, since I believed that I could do a better job manually than the computer. Which I couldn't.

So the next and final step for the printer was to turn Colorsync on. Which automatically disabled all those "Advanced" controls I had been fiddling with in vain. From now on Colorsyn would do the job. The only option left for me was to choose between three different profiles for the printer - one "Epson IJ printer", one "sRGB IEC something", and one "Automatic". I printed test copies of all three:

http://www.papermodelers.com/forum/attachment.php?attachmentid=102962&d=1316951922

Since I could not discern any difference between them, I set the setting to "Automatic" for now. What matters is that the prints were good, better in fact than anything I had achieved so far. And the means to get there was to leave Colorsync on and do nothing about color correction more than that!

Thank you, Wilfried!

[Leif Ohlsson]

Getting the printer to produce an output I could trust, the next discovery was that my screen may not have been the absolute norm of correct color rendering, as I had thought until that moment. The prints displayed a distinct yellow-red, light tan hue on the lower area of the stab, simulating clear-doped linen. This was nowhere to be seen if I viewed the test sheet on the screen.

So now I had to try to make the screen render an impression of the test-print sheet which looked like the printed part. Not the way or the order I had thought things should have progressed.

Using the screen adjustment tool with "Expert mode" checked - as was recommended on several websites - was no child's play. I worked back and fort with the levers for brightness and color, which had to be manipulated repeatedly through a number of iterated steps. But every attempt proved to be worse than the original setting.

http://www.papermodelers.com/forum/attachment.php?attachmentid=102963&d=1316952080
What really made a difference that I could accept, and be happy about, were three separate adjustments:

1) Reducing the brightness of the screen to some 75 percent. One website recommended making all colour adjustments at only 60 percent prightness, but this I cannot live with (old eyes, I suppose)

2) Changing the gamma value from the pre-set 2.2 to something which was called the "original" value of 1.73. Why the preset value differed from some "original" one I cannot explain, and neither do I know what a gamma value is, other than that it has something to do with contrast.

3) Changing the contrast of the screen until the light tan clear-doped linen part of the test print sheet showed up most clearly. The contrast adjustment was made in a separate panel for help to visually impaired (I imagine). This is what helped the most, and I think it ought to be standard in the screen adjustment panel. In fact, no manual mentions it, and I only discovered it by total chance.

Making these three changes my screen now displays colors much like the printed test sheet, while still retaining an overall impression when displaying images that is equal to, or better than before.

[Leif Ohlsson]

Next, I took the printed version I was happy with, and stuck that into the scanner. The idea now was to make test-scans of the already printed part, and print those scans again. Adjustments to be made until the printed scans where as similar as possible to the original printed part.

http://www.papermodelers.com/forum/attachment.php?attachmentid=102964&d=1316952258

Above you see: left - the scanner adjustment panel, top right - the scanned image as displayed afterwards on the screen, and bottom right - the original test print document, the printed version of which is laying in the scanner.

I'll make a long story shorter here by saying that I started with making a number of scans, and compared the scanned (but not yet printed) output onscreen, much like you see it above. Adjustments were made between each trial scan in "brightness", "color", "temperature" and "saturation", until I was satisfied with how things looked on the screen. All three versions were reasonably similar - on the screen.

This scanned version I now printed. Since the final scan adjustments had produced something which looked like the original document on the screen, and I had already calibrated the printer and screen to replicate each other, as close to "what you see is what you get" as I could come, expectations were for a good print result.

This, however, was not to be. The printed result was similar to what had started this whole investigation - there was a greyish hue over both background and colors. So had everything then been in vain?

[Leif Ohlsson]

Everything was definitely not lost - I had calibrated my printer, as well as my screen, and the control over the scanned output was within reach, but for the adjustment of the background grey. My final ace in this deck of calibrating cards was drawing up a number of strips in Illustrator (use any graphic programme, vector or bitmap, which can designate adjustable shades of black, grey, and white) consisting of one strip of 100 percent black, one strip of 50 percent grey, and one of white.

http://www.papermodelers.com/forum/attachment.php?attachmentid=102965&d=1316952521

These strips were then printed, and cut out. The idea is to cut out a small strip containing all three shades and glue it onto each original to be printed in the future. There is always some clear spot which will accomodate such a small strip, cut right across the black, grey and white.

After scanning, I will run each and every scan through Photoshop and correct the image by sampling the white point, black point, and intermediate (50 percent grey) point. This is done in a panel for adjusting brightness. Henceforth, this will be my new regime for scanning.

Printing two copies of the test sheet scanned with my previously arrived at settings, I subjected one copy to adjustment in Photoshop before printing, while the other was printed uncorrected. The result was most satisfying:

http://www.papermodelers.com/forum/attachment.php?attachmentid=102966&d=1316952521
Above, you will see the uncorrected scan in the left background, while the corrected version is in the foreground. In the background right is the printed original. This is the copy that has been scanned, and then printed again for both the uncorrected and corrected versions. With the pen I am pointing at the litte black-grey-white strip glued to the original. The scanned and re-printed version of this strip is pointed at with the red pincers in the lower left foreground.

Note the following:

The uncorrected version in the background left retains the greyish hue which was part of the problem to start with.

The corrected re-print of the original print in the foreround has no such greyish hue, and a good similarity in colors to the original print in the right background. It is definitely less sharp at the edges, which I believe mostly is due to applying the anti-moirée setting in the scan. This I deem to be unavoidable in scanning.

This completes my attempt to calibrate scanner-screen-printer for the moment. As a note to myself, these are the settings I applied in the scanner (as a hint to what kind of settings you might want to look for):

Brightness: -20%
Color +15%
Temperature +20%
Saturation ±0

Unsharpen mask - High
Descreening - Fine 175 lpi
Background lighting correction - none

[Leif Ohlsson]

This is the short version of the lengthy notes above:

Printer:
• Set to "Color Sync".
• Check which setting for paper quality renders the truest prints in your own opinion.

Screen:
• Reduce brightness to some 60-75 percent.
• Check if a lower gamma-point setting than 2.2 will give more true rendering.
• Check if contrast adjustment will give a truer rendering.

Scanner:
• Include a black-white-grey strip on every original to be scanned.
• Adjust for correct white point in a graphic programme by sampling the scanned test print in the appropriate panel of your graphic programme.
• Make adjustments in the scanner panel of brightness, color, temperature and saturation to render the truest result when finally printed.

This is how many testprints it took me to arrive at the advice summarized above:

http://www.papermodelers.com/forum/attachment.php?attachmentid=102967&d=1316952640

I think there are 22 cards in this hand. Well worth it, since I now feel more confidence in the tools I work with so much. With the experiences recounted here, you should be able to get away with far less.

Leif

[Texman]

Wow, nice work Leif. Thank you for the effort and your
willingness to pass your knowledge on to the rest of us
mortals.

[mldixon]

Very nice tutorial on the subject. Well done and to the point. Thanks for sharing your hard work on this. :)

[Wilfried]

Hello Leif,
thank you for your deep diving into the world of calibration and ColourSync.
The secret behind ColorSync the profiles you can create for every appliance individually .
So you build a trustable chain in Color Management.
To calibrate the monitor there are now cheap versions of the professional on the market. I have and use the Spyder2 - please take a look here:

Spyder3Express - Datacolor - Global Leader in Color Management Solutions (http://spyder.datacolor.com/product-mc-s3express.php)

With lovely greetings
the Wilfried

[Bomarc]

Great topic Lief. Don't know if it's been linked before, and while geared towards photography, I found this page to be helpful for general DIY monitor calibration:

Monitor calibration and gamma (http://www.normankoren.com/makingfineprints1A.html)

Mike

[Gil]

Hi Leif,

Excellent work and narrative. Great way of getting to the heart of the problem and reaching a low cost, manageable solution. I like the grayscale solution idea. You might want to check out IT8.7 Color Targets (http://en.wikipedia.org/wiki/IT8) now.

Bien Cordialement, +Gil

[Zathros]

Wow, that was an excellent explanation. If I could offer a bit of advice for someone who wanted an incredibly diluted and simplified statement on where to start, start with a BLACK and WHITE Gray scale. The best way to adjust a color TV or Monitor is to start with Black and white, then move on the colors. If the shades are wrong in the gray scales, you will never get the color right. If you were so far off on your monitor, that you needed to get to a "starting point" before you started the serious calibration, the color Red is the best way, this goes for TV's too. Make sure the Red color never bleeds past the edge of the object that is colored red. By lowering the over all color settings like this, you will get close enough to start. If you don't start with Black and Gray scales, you will spend a lot of wasted time.

Ultimately, you have to print out something to check it. Someone told me once that color and your preference was subjective and that is why Monitors and TV's have these settings. They are not. each color has a frequency and you may have a preference, but that is all it is (which means much) but calibrating your monitor can "teach" you how to see colors. That's why this tutorial, the links that Gil gives, can be very important. Especially if you putting in the unmentionable amount of hours that some of thee models take to build.

Temperature controls on TV's were there to compensate on how the different Cable/Satellite companies were broadcasting, and to be honest, they have really made life difficult as it seems the printer manufacturers don't give a "Hoot" about matching what you see to what you get. There is a person on this forum who is probably the absolute authority on this kind of stuff, and if he rang it, a lot of information could be brought to light.

Thanks Leif, that took a lot of work. This should (with my posts excluded) be a "Sticky".:)

[codex34]

Now for a silly question I hope you find slightly amusing :)
Calibrated under natural or artificial lighting conditions?

Seriously though, how do you manage different papers?
The settings the printer software give you don't always concur with what paper your using, as you've seen, the colour absorbtion rates for red, cyan and yellow are vary greatly depending on the density, finish and coatings of the paper.
Sounds a bit daft, but is it possible to make a colour chart so that you can print on any given paper and select a colour setting to use from the printed chart?

[Jim Nunn]

Leif,

Thank you for all the effort putting this into a manageable form. This is a subject that has plagued me for years. Perhaps now I'll be able to get reasonably consistent printouts.

Jim Nunn

[Pat_craft]

Fondamuntal subject !!! I will read it conscientiously...

[Art Deco]

Well done, Leif! Synching the color calibration across a chain of hardware and software is a notoriously daunting task, congratulations on your success! And thank you for all the work involved in documenting the process, it's another valuable contribution to the forum's collective knowledge. :)

[Leif Ohlsson]

Having followed links supplied at least a bit on the way, I now realize how little I know about this field. Let's use this thread as a repository for similar tips in the future. And if anybody has experiences of calibrating their equipment which could be useful for others, it would of course be wonderful to read about them.

Leif

[malachite]

Just like to add my voice to everyone else, a great tutorial and one to keep coming back to. well done Leif.

[Rubenandres77]

Leif, as it is usual with your threads, this is a very interesting and educative one :)

It is good to read the whole process you followed to solve the problem. The way you make things, with such detail and trying to get the best possible result, with all the trials, errors, and corrections you make to refine the results is very inspiring.

And this is not the first time you do such things. The Mustang resize and its numerous canopies, and most of your previous threads are something really delightful to read. In the end, the results are something important; but it is also important the methodic way you use to confirm them.

And of course an added value is the amount of information that other people share and contribute to enrich the topic.

Thank you Leif, and thanks to all the others who have posted here. It is great to learn from you :)

[Zakopious]

Excellent tutorial.

[rewalston]

Great tutorial, I've done what has been suggested here and my prints from PDF's are still too dark. The closest I can come is choosing "draft" in the printer settings. When I print photos though they look perfect. Getting really frustrated.

[Necroscope]

This baby makes my life much easier:
X-Rite ColorMunki Photo
http://dfqj586i2jk2a.cloudfront.net/UK/XRIT350/1/320x320

[Swampfox]

This baby makes my life much easier:
X-Rite ColorMunki Photo
http://dfqj586i2jk2a.cloudfront.net/UK/XRIT350/1/320x320

For $449.00(amazon)' it better make life easier.

How often do you have to recheck your system?


SFX

[rewalston]

I did something the other day that I don't quite understand. Whenever I have printed pictures on Photopaper and chose "Best" for the quality the pictures are perfect; meaning that they matched print and screen. So just for the hell of it, I chose Photo paper and Best, but printed it on my card stock that I used before. To my amazement the prints came out better than when I did the same PDF but chose normal quality or draft quality using for the same stock. I don't understand how putting more ink would cause better results. Any explanation?

Rusty

[SCEtoAUX]

I think it has something to do with the number of dots per inch the printer puts on the paper with the different settings. The Best setting would put more dpi than Normal setting. That is why Best prints slower than Normal and Draft prints fastest because fewer dpi are used in Draft.

[richkat]

Thanks Leif, this is a great help...........Rich

[rewalston]

I think it has something to do with the number of dots per inch the printer puts on the paper with the different settings. The Best setting would put more dpi than Normal setting. That is why Best prints slower than Normal and Draft prints fastest because fewer dpi are used in Draft.

Thanks SCEtoAUX, but what I don't understand, is that at normal setting everything looks almost black, but on best setting it's more blue like it should be.

[Necroscope]

For $449.00(amazon)' it better make life easier.
How often do you have to recheck your system?
SFX

It costs me about $750 in RU. :(
I've make home monitor calibration once and work monitors several times.
However I'm calibrating my printer at list once a month for two type of paper (white and silver(metallic)).
My main profit here is that I'm designing a model currently and have to have all colors on all pages to be exactly the same despite the time it's printed.

[Necroscope]

Thanks SCEtoAUX, but what I don't understand, is that at normal setting everything looks almost black, but on best setting it's more blue like it should be.

There are a lot of things involved on this way. F.e.- monitor color profile, PDF-file color profile, PDF reader color profile, printer color profile (different for different printing settings), color manipulation algorithm in PDF reader, in printer, paper color and it's reaction to your inks and so on.
So if you'd like to have exact color-matching - only way is to calibrate your equipment. If color-matching isn't a goal - try to play with printing settings until you'll got satisfactory results.

[hecfran]

Very, very useful!!!!