Re: Small pitch bends correct and tested. (issue4654063)

[Graham Breed]

Colin Campbell <address@hidden> wrote:

> Thanks for this, Graham!  I've added this to the tracker
> as issue 1711. Can you also provide a test file which
> would demonstrate the problem? I recognise this may not
> be all that easy to do, but it would help in verifying
> the fix.

Here's an example based on one Richard Sabey sent me.

In 2.14.1, the sixth tone sharp g comes out as a flattened
g sharp.  The pitch bend message is 0xe0 0x71 0x2a and the
note-on 0x90 0x38 0x5a.  Without pitch bends, g the
note-on is 0x90 0x37 0x5a, so g is 0x37 not 0x38. The third
byte of the pitch bend should be between 0x30 and 0x50 if
they're minimized.

A one semitone pitch bend is 0x1000.  The f is flat by a
f is flat by a sixth tone or a third of a semitone, so the
LSB of the bend should be (-0x1000/-3)&0x7f or (0x4000 -
0x2000/6)&0x7f, which is either 0x2a or 0x2b depending on
how you round it.  With 2.1.14, it's 0x39 (0xe0 0x39 0x35)
and with the patch it's 0xe0 0x2b 0x35.


                          Graham

pitch_bend_test.ly
Description: Text Data