Desktop Publishing Explained

This is the first of a series of articles to explain the theory and methods used by the printing industry as it pertains to computer generated artwork and text.

The best place to start in understanding the principals involved is to look firstly to what is required at the point your file is passed to the printer then going back to the beginning to work out how we arrive at that point.

With modern offset printing machines, the image we have prepared is printed on to paper is transfered by way of printing plate, generally made of aluminium but now commonly made from a plastic polymer. The image is got onto this plate by one of three usual methods. The first and more common is for the image to be put onto film with an imagesetter then contacted with the plate under vacuum then exposed to ultraviolet light. The second and most recent system utilises an imagesetter that puts an image diredtly on to a plate from a computer file bypassing the need to produce film. The third method utilises a machine that will transfer positive artwork, normally produced on a laser printer, directly to a plastic polymer plate.

The main disadvantage with the third system is that laser printers cannot produce a sufficiently fine dot to produce high quality photographic images and for any printing that embodies photos, either the first or second methods need to be employed. As the second method is still fairly new, the machinery involved is expensive and is at this point in time is generally only used by larger printing companies so I will cover the requirements of the first method, imagesetting to film, in detail.

Halftones

The reason that we need to use film for photographic images is that when printing either colour or black in white photos, the range of colours or shades of grey is far great for us to use a separate ink for each colour or shade. The way to get around this is to break a photo up into small different sized dots and to start with I will explain how this is done with black and white photos as this will be easier to understand. These different sized dots which generally are too small to be seen clearly without the aid of a magnifying glass merge together with the white of the paper to give the impression of different shades of grey, this depending on the relative size of each dot. This process is called halftoning where darker areas are needed, the cells are larger and overlap, for the very lightest areas, the cells are so small they don't print.

The number of 'halftone cells' used is called the halftone screen frequency, and is measured in lines per inch. The lower the frequency, the more visible the cells become, as an example an average newspaper uses a halftone screen frequency of around 90 to 100 lines per inch. The key to understanding this, is that the halftone screen value is different form the actual resolution of the image, but they are linked, and the screen frequency should really determine the image resolution. The halftone value only comes into effect when printed, the image resolution applies from whenever it was scanned, edited or created.

Lines per Inch

If you look closely at a newspaper photograph, you can easily see the 'dots' or halftone cells, if you compare that to a high quality magazine, the halftone screen frequency is around 150 to 175 lines per inch and is much less visible. So the higher the halftone frequency, the less 'dotty' the images.

So why doesn't everyone use a high screen frequency like those in the high quality magazines? Well, it depends upon the quality of the paper used, newsprint is rather low quality and porous, so smaller and fine halftone cells would simply run or bleed into one another and merge into larger cells. Larger cells are used in the first place, to avoid uncontrollable smudging. With high quality coated or glossy paper, the dots run or bleed less, so a very small cell stays a very small cell, and higher halftone screen frequencies can be used, achieving a finer quality print.

There is also a link between the halftone value and the maximum resolution the printer can output, which is covered later below. When the image is spilt into halftone cells at the time of printing, the size of each cell, and therefore the darkness, is calculated by using the average brightness of the pixels or dots in the image within each halftone cell. For this to work smoothly, the ideal is to have 2 pixels or dots in the image to each halftone cell, any fewer and the image may look jagged, any more and they are ignored, but they still add to the file size.

So for best results, image resolution should be twice the value of the halftone screen frequency used in the printing. For a newspaper, using a 100 lines per inch halftone screen frequency, any images would be scanned ideally at 200 dots per inch, and for a super glossy Vogue type magazine, with a 150 lines per inch halftone screen frequency, the images would be scanned ideally at around 300 dots per inch. Now I used the word ideally, that's because often they are scanned at less than twice, in practice you can get away with 1.3 to 2 times the halftone screen frequency, and the lower resolutions keep the file size down.

The whole thing ties in? Yes. If you measure image quality by the number of 'greys' or levels of shading it can contain, you can work out the scanning resolution and halftone screen. The greater the number of greys (otherwise known as levels of shading) available, result in a smoother more detailed image. Imagine an image created with only 2 levels, i.e.. black or white, it would end up almost as a silhouette, with 4 levels a little shading could be introduced, with 8 even more. Currently, the maximum available with PostScript is 255 levels, but that's only via the highest quality black and white printers known as an imagesetter, used to produce master artwork ready for conventional printing press output, most desktop printers offer far fewer levels.

To calculate the maximum number of shading levels that can be output: maximum number of greys = (printer resolution / halftone screen) 2 (squared) and that: ideal image resolution = halftone screen x 2. With a typical desktop printer, with a resolution of 300 dots per inch, to obtain a useful number of shading levels, the screen frequency has to be relatively low. Working this out as, 300 dots per inch printer resolution / 60 lines per inch halftone screen frequency, then squared, gives us 25 shading levels. If the screen frequency was higher, say 100 lines per inch, 300 dpi / 100 lpi, then squared, would give us 9 shading levels, not really enough for a good image.

If a printer is 'PostScript' compatible, you can set your own halftone screen frequency, but setting it too high simply reduces the number of shading levels, and possibly picture quality. But it can be very useful for images that require a smooth tint of a single grey or colour.

For commercial printing press output, the 'master' artwork is output on a very high resolution printer, know as an imagesetter, which can output at resolutions up to 5000dpi, though typically 2540. This means that the halftone screen value can be increased too. Working this out, say for Vogue, 2540 dots per inch imagesetter resolution / 150 halftone screen frequency, then squared, results in 286 shading levels, more than the 255 limit, and giving a high quality image.

So with a 300dpi printer, the halftone screen frequency will have a default setting of around 60 lines per inch, and as the ideal resolution is twice that, you don't need to scan at anything higher than twice that, so the magic number for scanning resolution is 120 dots per inch!

The next artlicle will take a look at line art.