Re: Bar line shading in PNGs

[Jan Warchoł]

2011/7/31 Phil Holmes <address@hidden>:
> Don't think this is a problem, but it is a bit interesting.  I've been
> trying to run my pixel comparator to compare 15.7 to 15.5 and getting a
> difference in the bar lines of every image - presumably owing to the work to
> stop the PDF artefacts.  Looking carefully at the bar lines, you can see
> they're not actually black - they're various shades of grey, in both
> versions.  The difference is that the shading has changed.  I attach 2
> highly magnified images - one with normal contrast, and one with it adjusted
> to emphasise the shading.

Interesting indeed.  I guess it works like this: when rasterizing
postscript output Lily decides that barlines are not thick enough to
be represented by a perfectly black 2 pixel wide line.  Therefore she
chooses to draw it as a dark grey 2 pixel wide line (dark grey 2 pixel
wide line looks from the distance like something a bit narrower than
perfectly black 2 pixel wide line).  However, on the spots where
barline and stafflines intersect, the "blackness" of the barline is
added to the pixel shade, making it darker.
In other words, a pixel on the intersection of barline and staffline
has more reasons to be completely black (namely, two reasons: being on
a barline and being on a staffline) than pixels lying on a barline and
not on the staffline.

thanks for showing this!
Janek