Just to give you a glimpse of the genius, I dug up a post that Thomas once wrote with regards to the issue of why fonts seem to get "bolder" when they are converted to outlines. Visit his blog, and enjoy this bit of useful information...
For those who care, I thought I should explain a bit more about why the text fattens when you convert to outlines.
TWO things happen when you outline fonts:
- loss of hinting
- change in fill algorithm
The loss of hinting makes certain features potentially inconsistent. For example, letter strokes that you expect to be the same width might turn out to be different widths depending on how they fall on the grid of the output device. Slight differences can get magnified unexpectedly, like rounded letters going below the baseline. This happens because the information that makes the outlines round consistently to the pixel grid has been lost.
The change in the fill algorithm combines with the lack of hinting to make the letters look fatter. Font rasterizing uses a fill algorithm that turns on a pixel only when the center of the pixel is within the glyph outline (center-scan). Graphics rasterizing uses a fill algorithm that turns on a pixel when any part of the pixel is within the graphic outline (overscan). Given that the outline is no longer being rounded to pixel boundaries at key points, the rendering will generally be at least one pixel thicker, occasionally two.
Of course, how much difference this makes depends on the size and style of the type, and especially on the resolution of the output device. At 2400 dpi with typical text sizes, the effect is pretty subtle. At 600 dpi with 6 point text, the effect is quite obvious.