3 ms·

Font Smoothing Detection in Browsers

9 points by latitude 15y ago

latitude 15y ago
That's my stuff. The page has been up for about 3 months now and the stats will go up in a week or two. Also, about a month ago I amended the test script to check which fonts are installed on visitor's machine. I think it could make a fairly interesting data point too. The script is currently checking for Arial Calibri Candara Lucida Sans Unicode Segoe UI Verdana Helvetica Lucida Grande Georgia Droid Sans It was also supposed to check for Trebuchet MS, but I misspelled it :) If anyone has any suggestions for additional fonts to test for, I am all ears.
- rorrr 15y ago
  Do you realize that there's RGB, BRG, vertial-RGB, vertical-BGR possible subpixel layouts in LCDs? http://www.lagom.nl/lcd-test/subpixel.php http://www.lagom.nl/lcd-test/subpixel.php Does your script take that into consideration?
  - latitude 15y ago
    It does not tell apart between these, no. It simply checks if there is any color fringing going on after a glyph is rasterized.
stevejohnson 15y ago
I think it's a mistake to classify all subpixel-level renderers as ClearType. There are different styles of subpixel rendering, some of which look better than others with large @font-face fonts. The system described puts Windows 7 and iOS in the same category, which is a really bad idea IMO.
- latitude 15y ago
  > There are different styles of subpixel rendering, some of which look better than others with large @font-face fonts. An example would be nice. Also, do keep in mind the context, which is an issue of a poor rendering of non-CT hinted fonts on Windows. > The system described puts Windows 7 and iOS in the same category, which is a really bad idea IMO. The JS detection result is meant to be used in conjunction with the UserAgent string of course, not just on its own.
CamperBob 15y ago
I'd prefer that browser developers give me more choices regarding how fonts should be rendered. Firefox 6 seems to have made an effort to look more like IE9, which is why I don't use IE9, and why I immediately Googled for hacks to disable forced Cleartype rendering as soon as I installed FF6. I know it's hard to believe, but not everybody likes Cleartype. And I'm tired of killing the first 4 hours every Firefox upgrade searching to figure out how to put things back the way they were.

9 points by latitude 15y ago

That's my stuff. The page has been up for about 3 months now and the stats will go up in a week or two. Also, about a month ago I amended the test script to check which fonts are installed on visitor's machine. I think it could make a fairly interesting data point too. The script is currently checking for Arial Calibri Candara Lucida Sans Unicode Segoe UI Verdana Helvetica Lucida Grande Georgia Droid Sans It was also supposed to check for Trebuchet MS, but I misspelled it :) If anyone has any suggestions for additional fonts to test for, I am all ears.

rorrr 15y ago
Do you realize that there's RGB, BRG, vertial-RGB, vertical-BGR possible subpixel layouts in LCDs? http://www.lagom.nl/lcd-test/subpixel.php http://www.lagom.nl/lcd-test/subpixel.php Does your script take that into consideration?
- latitude 15y ago
  It does not tell apart between these, no. It simply checks if there is any color fringing going on after a glyph is rasterized.

stevejohnson 15y ago

I think it's a mistake to classify all subpixel-level renderers as ClearType. There are different styles of subpixel rendering, some of which look better than others with large @font-face fonts. The system described puts Windows 7 and iOS in the same category, which is a really bad idea IMO.

latitude 15y ago
> There are different styles of subpixel rendering, some of which look better than others with large @font-face fonts. An example would be nice. Also, do keep in mind the context, which is an issue of a poor rendering of non-CT hinted fonts on Windows. > The system described puts Windows 7 and iOS in the same category, which is a really bad idea IMO. The JS detection result is meant to be used in conjunction with the UserAgent string of course, not just on its own.

CamperBob 15y ago

I'd prefer that browser developers give me more choices regarding how fonts should be rendered. Firefox 6 seems to have made an effort to look more like IE9, which is why I don't use IE9, and why I immediately Googled for hacks to disable forced Cleartype rendering as soon as I installed FF6. I know it's hard to believe, but not everybody likes Cleartype. And I'm tired of killing the first 4 hours every Firefox upgrade searching to figure out how to put things back the way they were.