Comment #73 on issue 2658 by [email protected]: wrong stem rendering
(inconsistent thickness)
http://code.google.com/p/lilypond/issues/detail?id=2658
Regarding comment #69: I am oscillating in my opinion. Pixels sticking out
is bad. But a +20% variation in thickness in print is also bad. I've
tried devising a scheme where beams can be strokeadjusted just at the right
and left edge with stem stickness in mind, but making this work with the
various imaging models is just not feasible.
Now we had stroke adjustment on unconditionally before, and it made staff
line thickness consistent. Why is the output now worse elsewhere for
print? Because of the way things were drawn, stems turned out somewhat too
thick on average, still with some variation, but at least covering their
outline fully and then some. That prevented things like beams sticking
through but lead to the ungainly variations in thickness of the stems and
much too thick stems in previewers with the Cairo rendering approach.
So currently I am leaning backwards to let stroke adjustment revert to the
default, leaving us with a variation in print line thickness. Possibly
with an option for overriding.
How can we improve over that state? By discretizing our placements and
thicknesses. If line thicknesses are (say) a multiple of 1/150 dpi and are
positioned at multiples of 1/150 dpi, we don't need no stinking stroke
adjustment. Every item will render looking identical and things will line
up properly.
If we do all of our positioning _ignoring_ an underlying grid, we will
always have to compromise between accurate positioning and accurate
thickness, and if various items are positioned independently on a
not-supported-by-reality grid, the compromises will show.
Now how to arrive at reality? 1/300 inch is 0.24pt in PostScript
parlance. We are currently working with a 0.4 pt base which are 1/180 in.
That does not map fabulously to current printers (well, Epson ink printers
maybe). So perhaps we need to fatten up to 0.48pt as our metric. It's not
like we don't already get to see this thickness in rasterization...
--
You received this message because this project is configured to send all
issue notifications to this address.
You may adjust your notification preferences at:
https://code.google.com/hosting/settings
