solenum / msdf-c Goto Github PK

I've noticed that msdf-c generates some glyphs, mostly lower-case glyphs, with more space around the main character data than msdfgen provides.

Here are some examples. Glyphs generated from Roboto Regular at 32x32, a range of 1, and with autoframe.

Chars: a, @, h, c, -

Chlumsky/msdfgen:

exezin/msdf-c:

256x256 Example:

I noticed the SDF images people were posting in issues seemed really broken, so I tried it out myself and sure enough got the same result. I was wondering how come the non-SDF output image was still OK, and I noticed that when you transfer msdf to bitmap_sdf, you convert negative values to unsigned integers (uint8_t).

Bad	Good

Hey,
I'm using an almost identical version of main.c but with the following changes:

  char fontdefault[] = "hiragino-kaku-gothic-pron-w3.otf";
  int size_sdf = 50;
  int size_bitmap = 14;

  char *c     = "あ";
  char *out   = "out.png";
  char *fontf = fontdefault;

The result is pretty blurry:

I tried setting size_sdf to smaller values like 14 but the quality is still not too great:

Perhaps I'm doing something wrong. How can I improve the result?

Shouldn't this allocate contour_count+1 elements?

https://github.com/solenum/msdf-c/blob/87857619bf8f40e82d1a63c3ec46bc54ea17277d/msdf.c#L766C67-L766C67

in msdf.c, line 910:
edge_segment_t *parts[3] = {};
edge_split(&contour_data[i].edges[0], parts[0], parts[1], parts[2]);

parts[3] are wild pointers.

fix:
edge_segment_t parts[3] = {0};
edge_split(&contour_data[i].edges[0], &parts[0], &parts[1], &parts[2]);