Saturday, July 5, 2014

Calibrate Your Screen!

Yep, we had that topic before, i.e. using the tools provided by operating system, see post here.
Of course, this is just a start. Despite those methods work pretty well concerning the gamma calibration of your screen, getting the gamut (color space) and white balance right is near impossible.
However, if your are serious about your photography, I actually want to work with a screen that displays colors as good as possible. Right, when doing B&W only, this is probably not of your concern.

Available gadgets
There are a couple of different gadgets on the market, which help you calibrating your screen.  For convenience, i.e. my local photo store, to which I can cycle in just about 10min, had a Datacolor Spyder4Express, which I obtained for a little less than €90. In fact, the Spyder4Express actually does the calibration for me, fully automatically.

Problems may occur!
Sounds all brilliant and easy, but, you will be disappointed at first. You may even consider returning the device claiming warranty for a defective unit. To prevent you from arriving at such considerations, I would like to share a few experiences I had with the device. I will also try to explain a little bit why things happen, when they happen. Although those experiences are based on the Datacolor Spyder4, other devices of similar nature may show similar effects.

Position on the screen
The Spyder4 is supposed to sit snug on you screens surface, so that no light can fall onto its dedicated color sensors. To allow for this position, a counterweight is fit on the USB cord. This weight can be repositioned along the cable.
  • Try to slowly track it into the position which fits your screen best. Apply minimum force possible, after all, you don't want to rip the USB cable apart.
Angle of the screen
In the good old times, monitors were big chunks of roundish glass, making in easy for whatever one dangles in front of it, to make good contact with the bulgy glass surface. However, those days are gone, and so are the rounded surfaces of computer monitors. Today, one owns a very flat bit of (mostly) plastic surface, which sits recessed in a frame of more plastic. Meaning, that in a normal operating position, a device, such as the Spyder4, will not be able to touch the surface in the intended manner.
  • Tilt your screen upwards, as far as needed (or a little more). This allow the color calibration gadget to actually reach the screen's surface.
Stray light
Datacolor, and other probably too, warns about light falling directly onto the screen during calibration. While this is a nice move, it seems to rather pointing back in time, when the display light was actually generated by a beam of electrons hitting a phosphorous layer at the inner surface of the goldfish bowl. We came to the conclusion that those days are over, right?
So, what is wrong with the advice? It is by far insufficient to not point a light source onto the screen during calibration.
Here is why: modern displays employ light guides, to guide white light (from the screen's light source) to individual cells containing liquid crystal materials. Although not intended to, some minimal amount of ambient light may be collected by the monitors surface, to be re-emitted through the screen, interfering with the calibration run. This actually happened to my. Different color if indirect ambient light resulted in different calibrations.
  • Darken the room in which you calibrate your monitor before doing so!
Green cast
This is reported on the interwebs, and it hit me too. I figure, the problem is caused by 2 factors: 1) ambient light interference and 2) angle of view when using the screen.
Concerning the first point, see the stray light problem. Now imagine that you calibrate your screen in your study, normal dim lights around you, e.g. incandescent bulbs... this is warm light, the calibration process pics up traces of that light an corrects for by shifting towards green.  Try the following: set your digital camera to a white balance for incandescent light and take a photo in bright sunlight... it will be green!
Concerning the second point, the calibrated monitor forces you to stare at it at the angle of calibration. Why is that? In the good old days, having your goldfish bowl on your desk, any individual pixel radiated the same color into any direction, nothing was polarized. Today's screens, however, are entirely based on polarized optics. Polarization is a tricky thing to deal with, hence, I will not go any deeper into the whys and what photons do anyway. Instead I wish to invite you to do a quick experiment: create a white surface on your screen, e.g. by starting a text processor, now look at the screen at various angles from above and below. I am using a relatively inexpensive PHILIPS LC-display, on which I can observe a shift from whit to green, when looking at the screen from below, and a shift from white towards blue/violet, when looking from above. BTW: right-left, on my screen at least, creates a red color cast. Back to the color calibration gadgets, those are designed to measure the emissions of the screen exactly perpendicular to the screen's surface thereby defining the angle in which the screen emits the correct colors.
  • Position your screen such that you look at it at an angle of 90˚.
Observation
It seems that a slight blue cast is more acceptable than a green cast, at least to me that is. Interestingly enough, all calibrations I have done so far seem to shift away from blue. I guess manufacturers want to avoid costumers to complain about greenish displays and therefore shift their color space towards blue.

All in all, I believe everyone working with visual content should calibrate their displays. Consumer level gadget are really inexpensive, seen what can be achieved... definitely worth their money.


No comments:

Post a Comment