For both manual and automatic digitizing, ‘clean’ images, sometimes referred to as ‘cartoons’, work best. Such images have a limited number of solid colors which in turn have well-defined outlines. Ideally, they are:
- Well defined, where each shape is made up of pixels of the same color
- Clearly ‘blocked’, where each shape is a stitchable size, at least 1 sq mm
- Saved at a color depth of at least 256 colors (8 bit), or preferably millions of colors (16 bit).
|Clean picture with well-defined outlines||Clean picture with well-defined color blocks||Complex picture, needs editing to create clean color blocks|
Automatic digitizing techniques produce best results with images of the type found in clipart libraries or created from scratch in a graphics package. Automatic digitizing can work with images from other sources but they require some preparation. This is because most commonly available images are not made up of solid colors. Scanners introduce noise, while graphics packages perform ‘dithering’ and ‘anti-aliasing’. Automatic digitizing works least effectively with photographic images which may contain many dithered colors and complex forms. With photographs, however, you can pick out shapes that you want to embroider, leaving out unnecessary detail.
Images scanned from hardcopy drawings or existing embroidery typically contain a lot of introduced ‘noise’. While they can be used as input to automatic digitizing, once again, best results are achieved with relatively clean images consisting of solid color blocks. Typically, logos and simple drawings scanned from business cards, letterheads, books, magazines, cards fall into this category. Noisy images typically need to be prepared by reducing the color count and sharpening the outlines.
Dithering is a software technique which combines existing colors in a checkerboard arrangement of pixels. It is typically used to simulate colors that are missing from an image palette.
Like noisy images, dithered images need to be color-reduced before use. Be aware though, that while the software is good at processing dithered colors within a defined outline, it does not work so well with non-outlined images.
Anti-aliasing is a software technique similar to dithering which is used to soften hard outlines where color blocks intersect. It produces smoother outlines by ‘blurring’ the pixels where colors join.
Where anti-aliasing is deliberately used to blur outlines, these need to be ‘sharpened’ before use with automatic digitizing.