Map voices to channels in MIDI output.

[lilypond.git] / lily / performance.cc

diff --git a/lily/performance.cc b/lily/performance.cc
index 33f6fb971ee13de6d803fa21d4ddcfcd5f555b20..b07c8604cfde4c445f1fc1259d64e3c181cab29f 100644 (file)
--- a/lily/performance.cc
+++ b/lily/performance.cc
@@ -53,44 +53,17 @@ Performance::output (Midi_stream &midi_stream) const
   if (be_verbose_global)
     progress_indication (_ ("Track...") + " ");
   
-  int channel = 0;
   for (vsize i = 0; i < audio_staffs_.size (); i++)
     {
       Audio_staff *s = audio_staffs_[i];
       if (be_verbose_global)
        progress_indication ("[" + to_string (i));
-
-      int midi_channel =  s->channel_;
-
-      if (midi_channel < 0)
-       {
-         midi_channel = channel;
-         channel ++;
-         /*
-           MIDI players tend to ignore instrument settings on
-           channel 10, the percussion channel.
-         */
-         if (channel % 16 == 9)
-           channel ++;
-       }
-
-      /*
-       Huh? Why does each staff also have a separate channel? We
-       should map channels to voices, not staves. --hwn.
-      */
-      if (midi_channel > 15)
-       {
-         warning (_ ("MIDI channel wrapped around"));
-         warning (_ ("remapping modulo 16"));
-
-         midi_channel = midi_channel % 16; 
-       }
-
-      s->output (midi_stream, midi_channel);
+      s->output (midi_stream, i);
       if (be_verbose_global)
        progress_indication ("]");
     }
 }
+
 void
 Performance::add_element (Audio_element *p)
 {