bug-lilypond
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Proposal: Take action to make PDF viewers render Lilypond-produced f


From: David Kastrup
Subject: Re: Proposal: Take action to make PDF viewers render Lilypond-produced files correctly
Date: Thu, 15 Mar 2012 16:29:03 +0100
User-agent: Gnus/5.13 (Gnus v5.13) Emacs/24.0.94 (gnu/linux)

Jonas Olson <address@hidden> writes:

> Bar lines, brackets and perhaps other elements produced by Lilypond
> appear too thick in most PDF viewers. As I have understood the issue,
> this is an error on part of the PDF viewers.

Most of it, as far as it appears to me, is a rounding problem.  It is
popular to blame the PDF viewers, but the truth is that if they are told
to draw left and right border independently, and the overall thickness
is 1.2 at position 0.4, then position 0 gets rounded down to 0, and
position 1.6 gets rounded up to 2, and we get a thickness of 2 instead
of 1.2.  Where of course a thickness of 1 would be more suitable.

The solution is to draw things of line nature with setstrokeadjust set
which does not rasterize left and right edge independently but adjusts
the rounding in order to get results that are less wrong.

Of course, if they join up with things drawn without that setting, or
having a different width but a common edge, then you might get jags at
the boundaries.  So setstrokeadjust can open a can of worms of things
that almost but not quite line up.

The solution LilyPond adapts is to avoid thinking about it and blame the
PDF viewers for closest rasterizations corresponding to what LilyPond
demands.

> Giving a Lilypond PDF to others is problematic because of this as the
> file will most likely not look good on the recipients screen and
> paper.

It is just the screen.  The paper tends to have enough resolution, even
though it would still help at 300 dpi.

> Having these bugs fixed would be a great practical gain for Lilypond
> users.

It's not a bug but a design choice.

> I propose that a test case is created that makes it easy to determine
> if a PDF viewer suffers from this problem or not.

They all do.  It gets ameliorated if the PDF viewer renders at higher
resolution and then uses antialiasing to boil this down.

> Preferably, this test case should exaggerate the effect of the
> rendering bug greatly. We could then file bugs with the PDF viewers in
> question and push for having the situation corrected.

There is nothing to be corrected in the PDF viewers.  The only thing
that would help there is a higher rendering resolution in connection
with antialiasing.  That is not specific to LilyPond.

-- 
David Kastrup




reply via email to

[Prev in Thread] Current Thread [Next in Thread]