Elements of Color

When defining the elements of color for images one of the trickiest things to get used to is the thinking of colors as "adding" of light to blackness (the absence of light).
     This is a complete reversal of what we are used to. From earliest of times, as a child with a box of crayons, we have grown to see color as something that we "add" to white (such as to a blank piece of paper). This is actually called "subtractive color" due to the fact of rather than adding color (with a crayon for example), we are decreasing the amount of light that is reflecting back to us from the paper's surface. We are subtracting light. If we were to subtract light evenly for all colors we would be yielding darkening shades of gray working our way towards black. If we were to subtract light for only certain frequencies of the visual spectrum, we would yield different opposing colors reflecting back to us from the paper's surface.
 

  

For cameras (film and digital), for scanners, for computers, etc. the colors we are capturing, defining, and reviewing are colors of light. These images are defined by using different shades of three primary "additive colors" - RGB (Red, Green, Blue). Through varying the values of these three colors of light, we can build a broad range of combinations yielding a seemingly endless selection of colors. 

The negative view of this image (above) is another way of looking at this. Areas that are the darkest in this reversed image (such as the clouds) is where light is the strongest. The reversed image also shows the image using opposing colors (such as orange being the opposite of the blue, as found in the sky).

It is tricky to think of colors in reverse. It is tricky to see something that is white (such as the clouds) as being full of color (100% red + 100% green + 100% blue). It is difficult to think of the blue sky as the absence of red and green.
~~~
     When adjusting colors for images, prior to preparing for print production, it is good to keep your images as RGB color values. At some point though this will need to be converted to a different "subtractive color" palette. The image file will need to be converted from RGB colors to CMYK colors. Printing equipment, printing with inks or toners onto paper, will need color values defined for the colors of inks or toners being used. "Four color printing" is with cyan, magenta, yellow and black colors ("process colors"). The images (below) show the different CMYK colors and once these are printed and layered one on the other, they will yield a printed photo with the colors blended.

The process for making this conversion may be as simple as making a specific selection within the photo/paint program being used for the image. Usually TIF files are used for images in print production, though there are also other image file types that can recognize colors defined as CMYK values.
     For making adjustments to images, since this should be done prior to the conversion to CMYK, there are many different programs that could be used for making refinements (while it is in its original RGB state). For making the final conversion though, some programs are limited and may not provide the option of then switching to CMYK.
~~~
     There are other considerations for refining images for production, and color images in particular, can be extra challenging. Other posts to this blog site have related to adjusting grayscale images though in time, would like to offer more information relating to working with color.
~~~
Photo source: One taken while cycling the roads of Cumberland County, PA, just northwest of Carlisle.
~~~

Moire Patterns & Scanning

Regarding problems relating to moire patterns, lets consider the challenges related to scanning previously printed items which consists of halftone dots. As is seen (above) in the foreground road and lawn areas there is a  problem with this scanned image of a color printed postcard. Below is another example from a different scan of the same item.

When scanning printed items there are two different patterns that mostly likely will conflict with each other. One pattern is the rows of halftone dots on the item being scanned. The extreme closeup (below, left) of the postcard reveals the halftone dots that are being used in the printing process. The second pattern is with the rows of pixels that the scanner creates (below, right) as it scans the item and assigns colors to each pixel. Chances are great that the two patterns will conflict with each other creating an unwanted moire pattern. (For further description, see earlier post on Moire Patterns.)

When scanning, these small halftone dots are actually the subject matter to be most concerned about (not the building, the sky, the lawn, etc.). The goal is to try and scan the dots accurately. To do this, higher settings are needed for resolution due to the smallness of the dots. The example shown (above, left) was scanned at 600 pixels per inch.

    Aside from scanning at higher resolutions, another trick is to try scanning the item by laying it on the scanner at different angles. Since the unwanted moire is coming from a conflicts with two different patterns of rows (of dots, and of pixels), sometimes the difference between the angles will make matters worse, other times it will make things better. This generally will work (trying different angles) more successfully with items that were printed originally in only one color. Multi colored items will consist of different halftone angles for each of the different colors being printed making it a bit more challenging to find an angle for scanning that does not conflict with any of the different colors. 

For judging quality of the scans, here is another example (above). The scan (left) captured the halftone dots reasonably well. It was scanned at a higher resolution and at an angle. The scan (on the right) captured none of the halftone dots, and consists of a pronounced and unwanted moire. 

~~~

A note of caution: When judging scanned images for quality, this really needs to be done only by zooming in close enough for seeing the actual halftone dots. This is to avoid yet another instance where moires can occur. The computer being used for scanning and viewing your images, has a pattern all of its own built into the display. Even with a good scan of an item, with good capturing of the printed dots, when viewed at different zoom sizes, you will see different odd looking moire patterns appear on your screen. The odd (and cool looking) pattern seen in the above example is only due to the size I was viewing it on screen. It was due to a conflict of patterns with my monitor. If I were to zoom in close though, the quality of the halftone dots would be evident. 

~~~

Now that quality dots have been scanned, the next step will relate to what we now do with them to avoid a problem that can occur during print production. Subsequent posts will related to how to refine these types of images for production. 

~~~

In spite of the problems and challenges that come from these patterns, I do find them to be fascinating and fun when found in the "real world." This comes to mind almost every time I travel the PA Turnpike and drive under one of the many bridges with large screened fences on both sides of the road above. Between the motion of my travel, and the conflicting patterns of the two screens above, I get to watch a constantly changing and moving moire. 

~~~

Photo information: The color postcard is of Lamberton High School, Carlisle PA. The second photo is the interior of the "A Street Church" (home of Carlisle Brethren in Christ Church from 1920-1952) scanned from a dedication booklet published by the Planning and Development Committee in 1972. 

~~~

Halftone Grays & Dot Gain

A bit more about halftone patterns and their dots....
     Many of the adjustments that are made for image tones relate directly to how they will print as halftone dots. As seen below, for lighter tones we are speaking of increasingly smaller black dots, and for darker tones increasingly smaller white dots. In print production there are certain limitations and tolerances that will effect how these dots will appear.

For conventional printing, the ink that is being used (such as black) will absorb into the paper which can result in having the image swell in size. The printed black dots become larger ("dot gain"), and the printed areas behind the "white dots" become wider. This too can happen with digital printing where toner (rather than ink) lays on, and is baked onto, the paper's surface. This layer of toner has a thickness to it which too can effect the sizes of the dots. 

      There are many considerations for anticipating how much "gain" to expect in production. Much of this goes beyond what is needed at this point for this discussion. One consideration though that should be mentioned is noting the stock selection. The paper being used can have a huge effect on what will happen once ink or toner is applied. Generally speaking, coated stocks are better for holding printed detail than uncoated papers. 

For refining tones with taking dot gain into consideration, would like to first consider the darkest areas of the image. In the example above, at some point the white dots will close up and no longer be visible. Much of what is to the right will print as 100% black (without the small white dots). In addition to this, many of the other darker tones, and the mid tones, may print darker than anticipated. 

  

 

For the lighter tones of an image, the opposite can happen. At some point the dots become too small to print. Rather than things printing darker (as mentioned previously for the mid and dark tones), some of the very light areas may print lighter. If tonal detail is found in these lightest areas of an image, this may be lost or diminished once printed.  

~~~

What to do?

    Though the description of the problem may seem confusing, the solution really is not that bad. One way of looking at it is that rather than having a color palette (in this instance of grays) that run from pure white (0%) to full black (100%), you would have a slightly adjusted palette (with different values). 

     For example, if printing on coated paper, it could be that your adjusted palette of grays goes from 4% to 96%. When adjusting tones for your image for highlights, all tonal detail should be at least 4% or higher (anything less may not print and appear as pure white). For darkest tones, these can be adjusted so that they are at most 96% (anything greater will print as full black; the small "white" dots not being visible). Keeping all the tonal detail within 4-96% should do the trick. If printing on uncoated stock, different values may be needed - such as 7-92%. (The print shop may have some suggestions for what values would be best for the equipment and stock being used for the job.)  

     Anther consideration is with all those tones between the extreme ones. If an image is mostly dark tones, rather than running everything up close to the edge (on the borderline of being black), maybe they can be lightened with the darkest areas of the image being set much lower, such as in the lower 80%s. For all images, it may be good to anticipate that all mid to dark toned areas will appear a bit darker once printed. 

~~~

Once you get past all of the technical details, when it comes to adjusting tones, it really is not that bad. It becomes less of a value-specific process and becomes more intuitive. You soon develop "an eye" for seeing tones in a different way - expecting to see mid to dark toned areas as lighter shades gray; highlights as darker shades of gray; white as a very light gray (of a certain value); and black as a dark gray (of a certain value). This is not to say that some details cannot go beyond these limits. Much of this is a judgment call. In some instances highlights may be small and in the middle of the shot, and full white may be fine. At other times, some of the darkest points may be details that are linear in nature and full percentage of black may be fine. Decisions on all of this can become natural and will become less about the numbers and more about the aesthetics of what you are seeing - with adjusted eyes. 

~~~