At work, I was tasked with improving the text display of our application in order to better support text kerning. Our application creates images that appear on television, so image quality is of utmost importance. Therefore, even small improvements in the appearance of any result we generate are very useful.
Our current text engine is implemented using Uniscribe, which seems to be the perfect solution. As already mentioned here , it supports ligature substitution in context with complex scenarios. It also processes languages โโfrom right to left and BiDi. This is important as we need to draw Arabic / italic languages โโperfectly.
Therefore, it seems very peculiar that Uniscribe does not seem to display information on glyph caching. I have shown a screenshot to demonstrate this problem.
alt text http://www.aliparr.net/kerning.jpg
My application does the same thing as a notepad, in which each character appears "monospaced." Notice how in Photoshop CS2 the bridge at the top of the โTโ hangs beautifully over the โeโ. I want to recreate this.
I know other APIs like Pango / Freetype, but it seems like a pretty tough decision to include all this just to make the final 1% of this task if Uniscribe is so good at everything else.
Am I missing a step using Uniscribe? What is the best solution? Can Freetype export kerning information in an easy way so I can integrate it with my existing Uniscribe solution?
Nb We just need to run on Windows - platform portability, fortunately, is not a problem that I need to worry about right now.
Welcome in advance!
source share