Behdad Esfahbod

All distance-field based text rendering approaches use the "signed" distance field. That part is not the question. Whether you use a sampling of the SDF or a less destructive representation...

I think we should merge freetype-gl, GLyphy, and msdfgen into a unified library that exposes a sane API while allowing for different algorithms, as well as mix-and-matching them based on...

One major difference in how GLyphy is implemented vs the other two, is that GLyphy does it's font processing online, on a per-glyph basis, whereas freetype-gl and msdfgen preprocess the...

> msdf might confuse the atlas algorithm, I would think, but it's probably the most desired format. Why? If anything, glyphy will confuse atlases as it's storing one-dimensional binary data,...

> Freetype-gl also processes on a per-glyph basis and on demand. For the atlas, I use the Skyline Bottom-Left algorithm (from Jylänki, J., 2010. A Thousand Ways to Pack the...

> For starters the projects could be moved into a common organization. Any ideas for a new project name? That can always be done later ;)

Myanmar Text from Microsoft renders the original test correctly.

This code: ``` float dist = texture2D(u_texture, gl_TexCoord[0].st).r; float width = fwidth(dist); float alpha = smoothstep(0.5-width, 0.5+width, dist); ``` looks wrong to me. width should be the size of a...

Ok I see how fwidth(dist) is supposed to work. Still, I think my solution is more correct, exactly when dist is small.

> This is the result I obtain for the texture gradient: > texture_gradient > and for the supersampling: > supersampling How is the e rendered? So, e looks fine, d...