Match VSFilter’s faux italics on PostScript-outline fonts, too #760
Conversation
libass/ass_font.c
Outdated
| .xy = 0x05700L, | ||
| .xy = ass_face_is_postscript(face) | ||
| ? 0x02d24L /* tan(10 deg) */ | ||
| : 0x05700L /* from GDI; roughly 0.34; tan(18.77 deg) */, |
There was a problem hiding this comment.
Technically, it is more sin(20 deg) instead of tan(18.77 deg)
There was a problem hiding this comment.
Thanks for the pointer. However, sin has no physical meaning in this context, whereas tan refers to the slant angle. I can add a comment that it may have apparently originated as sin(20°), but it doesn’t seem very useful—nor is it even a very good approximation of sin(20°), although certainly better than of tan(20°) for what it’s worth.
There was a problem hiding this comment.
It's sin(20°)*256, truncated to an integer, and then shifted left 8 to expand to 8.8 form.
There was a problem hiding this comment.
Well, yeah. “Truncated to 1/8th” being “not a very good approximation” (even rounding to 1/8th would have produced a different value). Anyway, is this relevant enough that you want it mentioned?
There was a problem hiding this comment.
Reworded to add a mention of sin(20°). Does this look good?
There was a problem hiding this comment.
Aaand removed again and reworded some more to avoid unofficial data.
astiob
force-pushed
the
faux-italic-postscript
branch
3 times, most recently
from
18864fa to
72c5bc8
Compare
Coefficient derived experimentally.
astiob
force-pushed
the
faux-italic-postscript
branch
from
72c5bc8 to
61992a1
Compare
#676 previously adjusted our faux italics to match GDI/VSFilter on TrueType fonts, but we later discovered that PostScript-outline fonts behave differently (font driver differences strike again) and our rendering became less similar to VSFilter’s after that change. This should fix it.
I derived the coefficient experimentally using Fansub Block. The first-approximation value I got was awfully close to 10° (9.998°), so I figure it’s probably exactly 10°. Thus, I’ve hardcoded the closest fixed-point representation of tan(10°) in this commit.
TrueType:
PostScript:
