Sub-pixel font rendering: it ain't just about the looks

June 12, 2007

Apple released Safari for Windows yesterday, and there’s been some discussion on the difference between how Apple and Microsoft render anti-aliased fonts.

I personally prefer a good font designed for on-screen reading (e.g. Verdana, Georgia) rather than anti-aliasing. On my ThinkPad (running Windows XP), I have ClearType turned off, because I find it distracting — the letters are blurry, and some of them seem to have a pinkish tint. On the other hand, I also own a Mac Mini running Tiger, and I’ve never really had a problem with the anti-aliasing there. There’s a good comparison between Windows 2003, Vista, and OSX here. In that photo, the OSX font looks too heavy, and a little jagged. The Windows 2003 font looks too wispy, and slightly colourful. Vista looks much better than both.

Anyways, Joel Spolsky has weighed in on the argument, and he seems to prefer Microsoft’s implementation. According to him:

The difference originates from Apple’s legacy in desktop publishing and graphic design. The nice thing about the Apple algorithm is that you can lay out a page of text for print, and on screen, you get a nice approximation of the finished product. This is especially significant when you consider how dark a block of text looks. Microsoft’s mechanism of hammering fonts into pixels means that they don’t really mind using thinner lines to eliminate blurry edges, even though this makes the entire paragraph lighter than it would be in print.

Seems like a pretty reasonable explanation to me. But then Joel says:

Microsoft pragmatically decided that the design of the typeface is not so holy, and that sharp on-screen text that’s comfortable to read is more important than the typeface designer’s idea of how light or dark an entire block of text should feel.

The thing is, the design of the typeface has a lot to do with readability. The height of the letters, the angle and width of the strokes, the serifs…these all play a part in the readability of a font. You mess with them, and you are messing with the readability. Microsoft’s rendering isn’t just affecting “how light or dark an entire block of text should feel”; it actually changes the placement and weight of the strokes. It might appear “nicer” at first glance, but that doesn’t mean it’s better or more readable. You wouldn’t take a bicycle and say, “Well, the exact placement of the parts is not so holy, so I’m just gonna move them around a little so that it looks a bit better.” They are exactly where they are for a reason.

I’m not actually picking sides here — I don’t think either Microsoft or Apple is right. They are two different approaches, with their own advantages. But I think it’s wrong to call Microsoft pragmatic, and make Apple out to be all hoity-toity and unwilling to compromise their artistic integrity.

For the uninitiated, here’s a good page on font readability, and another on sub-pixel font rendering.

Update: Dave Shea, who is actually knowledgeable in typography, agrees with me. Jeff Atwood points to some studies that were done that show ClearType to be slightly more readable than standard (non anti-aliased) text, as well as a white paper describing FontFocus, which is similar to ClearType, but doesn’t use subpixel rendering, so your black and white fonts aren’t rainbow tinged.