It looks as if you have found a bug, indeed. If I try the following
modified version
of your example, it produces its output on channel 10 with LilyPond
version 2.8
but on channel 2 with LilyPond version 2.10. The bug is in
lily/performance.cc,
where a test
if (s->channel_ < 0)
was removed, which means that the explicit setting of channel_ = 9 in
lily/staff-performer.cc is lost.
\version "2.10.0"
\score {
\new DrumStaff <<
\drummode {
bd4 sn4 bd4 sn4
<<
{\voiceOne \repeat unfold 16 hh16 } \new
DrumVoice { \voiceTwo bd4 sn4 bd4 sn4 }
>> \oneVoice
}
>>
\layout {}
\midi {}
} %end lilypond
However, Simon, there was a problem with your input file as well.
The <<{...} // {...} >> feature is hard coded to provide ordinary Voice
contexts, so it will not give you the desired DrumVoice contexts. I'm
actually
surprised that you get any reasonable printed output. Clearly, there is no
corresponding MIDI output. By explicitly instantiating the DrumVoice
contexts above, I solved that problem.
Regards
/Mats
Simon Dahlbacka wrote:
On 11/13/06, *Mats Bengtsson* <address@hidden
<mailto:address@hidden>> wrote:
As long as you use a DrumVoice context to typeset the percussion, the
MIDI output should automatically contain the corresponding drum
sounds
in channel 10.
/Mats
In that case, something else is broken. I
ncluded is a snippet from 7.4.3 Percussion staves in the 2.10 manual
(wrapped into a \score block)
%begin lilypond
\version "2.10.0"
\score {
\new DrumStaff <<
\new DrumVoice = "1" { s1 *2 }
\new DrumVoice = "2" { s1 *2 }
\drummode {
bd4 sn4 bd4 sn4
<<
{ \repeat unfold 16 hh16 }
\\
{ bd4 sn4 bd4 sn4 }
>>
}
>>
\layout {}
\midi {}
} %end lilypond
The produced midi does *not* sound like drums... (apart from only the
first four notes audible...)
/Simon