Re: Imaging

Albert Yan (pointnshoot@earthlink.net)
Mon, 23 Sep 1996 21:48:05 -0700

Terry King wrote:
>
> Dave
>
> You said:
>
> >One has to truly understand the concept of densitometry, how RGB
> >corresponds to densitometry, and how dye/pigment imperfection affects this
> mapping. Otherwise we can only talk about cook-book type discussion like
> >setting this value to this or that, etc. If we talk too much into details,
> >then it might make the whole group feel like this is just another computer
> >newsgroup. Sorry if I am wrong to think this way. If you do have specific
> >question, I will try my best to answer if I do know the answer.
>
> To those of us who are acquiring our first Photoshop and printers, although we
> may have years of experience of alternative processes, it is very frustrating to
> be given the recipes without understanding the principles involved. Many of us
> do alternative processes partly because we do not like doing things acording to
> the book. As you say understanding the concepts is important. Friends happily
> babble on about CMYK and RGB and what various things do but unless I understand
> the concepts, the skeleton, I find it difficult to put flesh on the bones. I am
> reading and listening at the moment hoping to acquire the knowledge to ask
> sensiblequestions. Please keep talking as you do.
>
> This digital thread is very useful to those of us without education or
> experience in this area. If those who are are contributing can benefit from any
> expertise those of us listening may have in thealternative processes themselves
> then that is one of the great strengths of this list.
>
> Terry King

Let me add a bit to this - what Dave talks about above refers to how the color gamuts (the
range of colors that can be reproduced by each of these devices) of the source image, the
scanning (digital input) mechanism, the display mechanism and the output mechanism
(printer) map one against the other. Color differences add up as you go from step to step.

One can map these changes by a variety of means. A calibrated (measured and characterized)
input image that is then followed through each of the stages and measured at each point is
commonly the way that the system is calibrated. Usually this means that the measurements
of hard copy are made using densitometers, if the monitor is measured then a
spectraphotometer is used. A colorimeter is a densitometer that measures in color
coordinates rather than reflection of the various primaries.

The complication here is that most of the input images are generated using CMY (cyan,
magenta, yellow) dyes. Photographic images are composed of CMY dyes, commercially printed
input is generated using CMY and K (black is added to the process to improve highlight
contrast). The monitor is RGB. Most of us think in RGB because it's intuitive, whereas
CMY(K) is not. PhotoShop is generally used in the RGB mode for the same reason. To print,
therefore, the RGB image must be converted to CMY(K). Photoshop has a way to contol this
transformation, but it's beyond most of us to understand. I must confess that after all my
work building digital cameras, color scanners and color printers, I still find it useful
to think in RGB and let software do the hard work for me.

There is a bunch of SW out that is supposed to help with this whole process and as a start
I would encourage you to look into that SW. Apple on its Macintosh systems has SW called
ColorSync. Properly used (and it's not easy to do it right!) it's supposed to aid in the
calibration and, after the calibration, perform all the transofrm for you so you don't
have to do them. Trust me, ColorSync or some equivalent SW is the way to go. Scanners and
monitors are linear - 3x3 matrices are as complex as you'll have to go - but printers are
non-linear, and that makes life REAL interesting. Stick with the commercially available
SW.

Hope this helps, not confuses.

Albert