Re: Fix trill timing in articulate.ly (issue 173750043 by address@hidden)

[Peter Chubb]

>>>>>   <address@hidden> writes:

> It's a bit presumptuous of me to comment on code I know little
> about, but it occurs to me that if it were that simple, it probably
> would have been done that way in the first place.

I wrote the original --- I hadn't thought that it would be this
simple!

The heuristics to choose the twiddletime in my code worked sometimes.
But often, (especially noticeable in slow passages) the trills sounded
too slow.  Try Messiaen's `Abime des Oiseaux' for instance, which has:

\relative c' { \tempo 4=92
  \acciaccatura a8\f-> g1(\trill\deprecateddim g16[ a g a] g8[)
  a16]\( g8[ a16(] g8[)^"rall" a16] g8.[\> a-- g--]\) r8\!\fermata
  \bar "|"
  }

near bar 31.


> Can you go into some detail about the nature of the problem?  What
> case was failing for you and how is your version better?

> Are the existing regression tests thorough enough to fail if you
> have oversimplified something?  Do you need to write a new
> regression test so that your case which was handled badly in the
> past does not break again the next time someone modifies this code?

There are no regression tests for articulated MIDI output at present.



-- 
Dr Peter Chubb                                  peter.chubb AT nicta.com.au
http://www.ssrg.nicta.com.au          Software Systems Research Group/NICTA