On Wed, Jan 12, 2011 at 3:32 AM, Richard Mason <whitne...@xtra.co.nz> wrote: > On Wed, 12 Jan 2011, Philippe Wittenbergh wrote > >> If you want/need to know the aspect ratio of a font, this service can help >> (for fonts installed on your local drives): >> http://fontdeck.com/font-size-adjust.html >> or this page: >> http://brunildo.org/test/fontlist3.html > > Both wrong because the aspect ratio of a font is not fixed but varies with > font size. > At http://www.emdpi.com/cssfontsizeadjust.html I have a download with graphs > of aspect ratio v font size (10px to 50px) for a number of fonts.
Yes, the computation of the aspect ratio gives results depending on the size. But this does not mean that such computations are wrong. Most of the oscillations observed starting from small sizes are caused by the inevitable fact that fonts are rendered on a discrete grid with very few pixels. I think that what is usually called 'aspect ratio' (without size specification) is the "ideal" ratio, when the font is rendered on a sufficiently fine grid. Indeed if you extend the computation of your graphs to higher sizes the oscillations reduce. In addition to the page mentioned by Philippe, I also have this ones [1], [2], where you can produce graphs similar to yours (they require a browser with correct support of em, Gecko 1.9 and also IE 9 seems working fine. The oscillations smooth out if you go till something like 1000px). The oscillations observed at small font-sizes probably reduce the usefulness of font-size-adjust at that sizes, but does not necessarily mean that is useless or that it must be fed values depending on the size (Moreover as technology improve and physical pixels density increase, we will hopefully have to work in a range where the oscillations of the above graphs are smaller). Regards, Bruno [1] http://brunildo.org/test/x-height-compare.html [2] http://brunildo.org/test/normal-lh-plot.html -- Bruno Fassino http://www.brunildo.org/test ______________________________________________________________________ css-discuss [cs...@lists.css-discuss.org] http://www.css-discuss.org/mailman/listinfo/css-d List wiki/FAQ -- http://css-discuss.incutio.com/ List policies -- http://css-discuss.org/policies.html Supported by evolt.org -- http://www.evolt.org/help_support_evolt/
