Bitmaps, Channels, & the CMYK process

Firstly I will clarify an often misunderstood or underrated option that is available in all decent Painting programs (PhotoShop, Picture Publisher etc.). If our Desktop or GUI is set to a resolution of 800 x 600 (example), then the monitor screen is mapped at 800 pixels horizontally and 600 pixels vertically. A bitmapped image (i.e. scanned picture) is also mapped as an array where a resolution of 300 pixels per inch in an image that is 3 inches wide and 2 inches high will produce 900 pixels across the image horizontally and 600 pixels vertically - 540000 pixels in all..

Getting that out of the way at the beginning helps us to also understand just what a BITMAP is. A Bit is the basic and minimum value used by or referenced in a computer at both the electronic and software level. (Notice your file size references that may be 104Kb (example) or 104 thousand Bytes where a Byte = 8 Bits). Therefore a Bitmap is a map or organised array of pixels of Bit based information.

• Black and White images have only one CHANNEL.

• Greyscale images also have only one CHANNEL.

• RGB images have three CHANNELS.

• A Black and white Line bitmapped image can also be referred to as a 1Bit image. Each pixel can only be Black OR White.
• A Greyscale bitmapped image can also be referred to as an 8Bit image. The one channel contains 1 Byte of information per pixel, therefore 8 Bits.
• An RGB bitmapped image can also be referred to as a 24Bit image. Each channel contains 1 Byte of information per pixel or 8Bits. Therefore 8Bits (R) + 8Bits (G) + 8Bits (B) = 24 Bits.

1. Read the Filtering for Separations page now.
2. Read the Ideal Inks page now.

To look at specific problems associated with CMYK imagery, I will break them down into brief paragraphs on each topic.

Ink and Paper affects the result
As mentioned in the Intro to Colour start page, CMY pigments are far from perfect. The different bases that they are transferred to also differ greatly (even white paper is not all the same and can produce varying results). The best CMY pigments that man can produce contain elements of the other two colours - they do not absorb/reflect light perfectly.

Halftone Dots, the printers nightmare
Consider the lowly halftone dots. As small as they might be, they have to print with as little absorption and spread as possible (size and shape change). The pigment ideally will be printed with highly accurate density values and yet be translucent enough to correctly absorb or reflect a mix of light back to our eyes when the dots overlap each other. All this, plus a few other instances that alone would take lengthy tutorials, with only 3 base colours mixed by varying dot sizes trying to reproduce, literally, all the colours under the sun.

Conversion and Look-up Tables
When we scan a bitmapped image it is done so with light. Light from the original (Transmission for transparencies - slides, Reflection for photographs and artwork) is directed and focused on to a man-made array of minute light sensitive electronic devices, thousands of them. These devices respond by varying at their output, an electrical current applied to them, according to the strength of light hitting them - dark to bright. The light at this point has passed through, inturn, Red, Green and Blue filters. For each filter, analogue to digital converters then record the current value as one of 256 grey values for each pixel and bingo, we supposedly have our RGB information (instances of higher quoted colour depths will be mentioned in the Scanning Intro page, under construction!). If we think of CMY as being a product of the Additive colour system then the fact that software has to use lookup-tables (data bases) of information to convert RGB to CMY may be appreciated.

Black doesn't exist
The conversion continues by firstly converting the Additive light values to Subtractive values that will be represented by pigments, and secondly, includes error information related to the problems already discussed concerning the pigments, and finally produces a fourth level of information that will be used as a Key colour from the RGB information (also as a result of deficiencies). Remember, we cannot separate black, it has to be constructed once the RGB to CMY results are established.

The values recorded for each channel in each pixel do not therefore directly represent the original RGB light, but values that the video, monitor and output devices will use to either represent (the RGB video monitor), or create, the CMY strengths (halftone ink dots or perhaps blobs of hot wax in a Laser Printer).

There just aren't enough colours
Finally consider that we can only reproduce 10 or 11 million colours with CMYK processes (refer to the colour Gamuts image, Colour Intro start page). A 32Bit CMYK 4 channel bitmapped image therefore only maps the restricted number of reproducible colours.

So a multiplication of 256 x 4 cannot be applied to ascertain the number of possible colours.

When converting an existing RGB bitmapped image to CMYK, in a painting program like PhotoShop, the same processes without the scanner are applied.

The images above show a picture created in RGB and then converted to CMYK. If this image was created in a Drawing program and output as CMYK the results would be the same. In the Prepress industry, we proof our images using expensive "contract quality" proofing systems that come very close to imitating the press output. Many an artist goes beresk because 'what they saw (on the monitor) is not what they get.' Should the job get printed without a trade quality proof (ie by a cheaper inaccurate colour printer instead), the client will (and do) knock the job back and demand an expensive reprint.

The changes can be quite dramatic as displayed above and the Choosing correct DTP Colours page describes the steps that must be taken to avoid such a catastrophe with Bitmaps and Vectorised graphics.

Did you read the Filtering for Separations page and the Ideal Inks page.