Printing Colour Problems ...

Question

I know this is a common cry for help, but I am having a dreadful time trying to understand colour profiles, despite spending a lot of time reading up on the subject. Here's my basic dilemma:
I have calibrated my display (Apple 23" Cinema). I am using an old Epson 830U for basic home proofs. I have set up my proofing profile in Aperture to match the printer and paper I'm using. When I print from Aperture I select my printer and paper in the colour sync profile, and in addition I ensure that the paper type is correctly specified in the printer's settings (and I also deselect colour managment in the printer settings).
Does this sound right ? I've noticed that even just a small mistake, e.g. forgetting to select the paper type correctly, can lead to extreme colour differences.
At the end of the day I'm not despertate for perfect colour in my home proofs, but I am trying to get them to look approximately right. It's not uncommon for me to end up with extreme colour problems. I can't print anything from iPhoto without it coming out too dark and with a blue tint. Luckily, printing from Aperture seems to produce better results (I just wish I knew why).
Surely it shouldn't be this hard to set up a basic Epson photo printer on the Mac.
If anyone knows of an idiot proof guide to home photo printing using high end applications on Mac OS X, feel free to point me in that direction.
Paul

Answer 1

I suspect it really is intricate as opposed to hard. But here's my "Dummies/Idiot's Guide ...." for you.
The difficulties start by reminding ourselves that color is perceptual. Our retina sends a signal to our brain, and our brain translates that to a color. But, who knows what color you perceive as blue compared with myself if we both stand side by side and look up at a blue sky. we've just been told since a baby that it's blue, but our brain does it's magic and we "see" what it tells us. This translation issue between retina and perceived image is behind much color-blindness ... a difficulty in perceiving a color difference.
Since we can't "standardize" our brains (well not anytime soon), what can we standardize that can reduce variance in our actual and perceived perception?
Perception is 9/10's of the Law:
Ever taken your laptop outside? When you can actually see the image due to the relative brightness of the outside to indoors, did you notice that colors seemed different? Your eye is adjusting to the ambient light, and your brain gets in on the act. Your brain decides that the brightest light source is essentially "white" and so tricks us. That's why we can take a photo inside under tungsten light, and what looked white to us as we took the shot comes out as yellow-orange. Quantum Physics time: Tungsten releases "light" (electromagnetic radiation) at this frequency when electrons that had been bumped up to higher energy states (by us supplying power from the on/off switch) fall back to a lower energy state and give off or 'emit" a packet of energy comprising the difference between those two states.
Halogen headlights look brighter, but also whiter than tungsten headlights. That's because the radiation they emit is somewhat bluer. Xenon headlights are even more blue biased.
The sun gives off light from a whole collection of different elements, but by and large it is yellow/red. However, there is lashings and lashings of it compared to a tungsten bulb, and our poor, overworked brain is left to cope figuring out brightness vs. whiteness. Our brains do OK, but they do it by tricking us rather than science.
Home is where the heart of the problem is:
So, OK, we've taken some photos, and pressed this button to "auto white balance" or selected a "white balance" that tells the camera that although white isn't really white, pretend that it is, since our brain is also pretending. Of course, if we forget to tell our camera this, we have to tell Aperture or PS CS2.
But, sat there in front of our screen, what are we really seeing? Well, we are not seeing just the displayed image, no oh no. That would be too simple! No, our brains are passionate about their environment. What's the lighting? Fluorescent (I hope not), halogen, tungsten? This matters as our eyes/brain will calibrate to more than just the on screen image. What wall colors exist? Blue, red, green, off-white? Those refceted colors are going to fool our brain into believing that the image on screen has a tint. This is why Aperture's window is all gray, and photo professionals strongly suggest image editing is performed in an all gray room, with as "white" a light source as possible (halogen vs. tungsten, tungsten vs. fluorescent). BTW. I dislike fluorescents as all the bulbs seem to have hideously different colors, and age badly across time. That plus the rather unattractive greenish tints they give skin tones puts me over the edge.
Oh, and no desklights with their nice pool of incident light or other effect lighting. Save that for the living room. This room is to reduce the trickery that your brain can throw at you. So, nice uniform, diffused lighting.
Watching you watching me:
But if we're now living in our nice gray world, ready to edit those wonderful images that we used the right white balance for when shooting, what next?
Calibrating your monitor means you standardize against those who wrote a color profile for the printer, the paper or who will be printing images for you. Your image should look the same on someone elses monitor as it does on yours. That's step 1.
However, displayed color is also perceptual. CRT's render color differently IMO to LCD's. Both are transmissive, but brightness differs and perhaps more importantly, what is color? Outside, our eye saw all the different wavelengths it was capable of sensing and passed the relative intensity values per wavelength to our brainiac computer. Not so in our digital scanner or camera. In computer technology, color is rendered as a triangular space represented by a dot at each apex ... one red, one blue and one green. Those point "phosphors" in CRT vary from manufacturer to manufacturer and to the standard required. Close but no cigar. The electronics and guns also do not track perfectly linearly from pure black (no signal) to pure white (max. output of a phosphor of equal intensity to the others). This is called gray-scale tracking and unless it is near perfect, you can see green or blue shifts in color at the black end or white end etc.
Uh oh, that white question again. Come to that, how do I know it's the same white that I set my camera to when I shot? Well, we don't really wiothout setting to some standard. It's not that each phosphor isn't emitting, it's the balance across them at our determined highest brightness (white) that needs to be set. D65 is often quoted in TV's, representing a color temperature of 6500Kelvin (deepest reds are low temperature emmissions and brightest blues are the highest temperatures).
Now LCD's don't have phosphors illuminated by a stream of electrons, and they came long after we had developed pretty decent CRT technologies. So being lazy humans, we lifted the ideas and terms No phosphors, but they do have red, green and blue LCD's (each a sub-pixel, and the three combined are a pixel)triggered by an electronic signal. so, why aren't they called LED panels? Well, LED's are not very bright, and we need a mechanism to make them brighter. Those quaint watches came to the rescue ... LCD's or Liquid Crystal Displays. You know, those ugly greeny/gray, low contrast watches. Yes, that's the one. Sandwich those colorful but dim LED's between an LCD, and use the LCD to switch on/off a powerful light ... rather light turbo charging a 4 cylinder motor to make it perform like a V8! The posh name was TFT-LCD (for Thin Film Transistor LCD). Now we have a bright, flat screen. Yahoo!
But backlighting of the panel is not always very uniform (look at colors around your screen for different flat colors at different intensities and you may see patchiness). Nor do LCD's behave exactly like phosphors from deepest black (impossible when you have to have a back light which is always on and which illuminates the whole panel of glass) to white. But why sweat this, when the LCD rarely conforms perfectly to the red, green, blue phosphor standards.
BTW, it is the backlight that kills laptop battery life, not the LCD's themselves. Turing your brightness down as low as possible helps .... but I digress.
So we buy calibration hardware and software, and calibrate our monitor to a ruler flat grayscale from pure black to pure white.
And it looks bad.
It is rather idiosyncratic of us that a perfectly linear gray scale can leave us perceiving a lack of contrast. So, we distort the "gamma" value to provide more punchiness. Apple vs. Windoze computers tended to have different gamma's out of the box. Apple uses a gamma of 1.8, a rather flater, more pastel effect that, if my knowledge is correct, translates/translated better to the characteristics of the professional CMYK (Cyan, Magenta, Yellow, Black) four color printing processes used by offset litho printers. Windoze uses a gamma of 2.2, a punchier, gutsy, in-your-face effect. Marketing? I don't know. But it does appear a better match for RGB (our computer friendly Red, Green, Blue) inks used in inkjet printers.
So we calibrate our monitors, and then deliberately decalibrate them (if you will) to correspond to our output destinantion.
A quick digression here.
If our output is destined for print, what dos that mean to us? Well, part is covered above. But what about web, or tv viewing from DVD? Ahh hah! Gotcha!
We can control our calibration. We can even calibrate ourselves to external print shops, knowing that our image will appear 99% or more the same on their monitors as it does on ours. And they will have calibrated their monitors to their printers and paper (more on this below). But what about Joe Public?
Back to that color space.
A 10 year old color tv, used moderately over the years, performs no way like a new color tv. The phosphors "age." even the backlights in LCS panels age. Age is represented by diminished light output, and grayscale tracking. So, our 10 year old tv cannot represent the wide color space it could when it was brand new. Worse, even if we'd left our 10 year old tv in its box and just opened it, it would not be as capable as a brand new design. TV's are actually held down to 1950's standards rather than up for this very reasson, but the same is somewhat true of computer monitors. We don't know the age, quality or even the color calibration and room environment of our web browsers monitor. So, a much smaller "web safe" color space (sRGB ... standard - or small - RGB) was introduced. This restricted space should be renderable on most old monitors, though gross end consumer calibration issues will obviously remain uncontrollable.
Meanwhile, we want to keep as big a color space as possible ... difficult to put back in what's been taken out. (Set your camera to "Adobe RGB" please right now.) Working in the biggest color space allows us to target any lesser color space with relative ease. And some of these restricted color spaces belong to printers ... which of course, differ from printer to printer, (probably more associtated with inks to inks) and paper to paper ....
More on this next
White points