GRAPHICS> Scanning : File Format : Linework : Photos : Image Manipulation : Filesize : Transparency : Animation : Banners : Buttons : Backgrounds : Protecting Your Artwork

Linework - quality optimisation

If you're starting with source material that isn't very good quality, there are things you can do to improve it.

Take the graphic below, which has been sourced from a childrens' book that was over 50 years old. The paper is often very foxed or yellowed and the ink faded. This is what it looks like when I start. To get from the top image to the bottom one could take five or ten minutes of digital manipulation, but the end result is infinitely better than if I'd scanned it as line art, as I've had control over the loss of detail all the way through the process.

dhakshvkshcv

I won't go through the process in fine detail, there are whole sites devoted to Paint Shop Pro and Photoshop tips - I'll just give you a basic run through the processes:

Firstly, I scanned the image at 144 dpi and full colour so that I'd have "room to move" (so it was actually twice the size that it appears here).

Secondly, I increased the contrast and brightness until the background was much lighter. I do this slowly, 10 or 20% at a time to make sure I'm not losing the linework. The idea is to increase the difference between the linework and the background as much as possible.

Once the background is much lighter, I use Magic Wand to select it. The tolerance I use on the Magic Wand depends on the variation of tones in the area I want to select - this example varied from quite pink to quite yellow, so I had to use a relatively wide tolerance - this means that you lose a bit of detail, as the linework will get eaten away a little - hence increasing the contrast earlier. With my background-colour set to white, I then delete the selection. This gets rid of the mucky bits caused by the texture of the paper, foxing etc. Experiment with the tolerance - I usually start with about 10% variance from the chosen pixel's RGB colour.

Magic Wand only selects one area, so you'll need to repeat this process for all the main areas that need to be cleaned up. You can do it all at once by holding "shift" to add to the existing selection or "control" to subtract from the existing selection (that natty trick works with most tools in most graphics packages, and is one of the most useful things I've ever learned - it means you can even combine tools to make a selection, such as magic brush to select a large area and then "control" and lasso to remove a certain area from the selection before deleting it).

Once I've got rid of all the mucky bits and have a nice clean white background, I resize it to 72 dpi and the width or height I need for the page. Having the graphic as full-colour when you resize will give you a much better result than doing so at 256 or less colours. This is worth remembering when you're working with other people's gifs - always increase a gif to full-colour while you're doing any manipulation, then drop it back down to 256 or fewer colours when you save it.

Resizing the picture will make it go a bit fuzzy (and remember, I mean making it smaller as you'll never enlarge it for reasons discussed earlier) - this is because the software has had to average pixels to calculate the smaller version of the image. This can be improved by using "Sharpen Image" - this gives you back nice sharp linework, and is one of the many functions that will only work at full-colour.