On Mon, 2012-08-20 at 19:26 +0000, Fons Adriaensen wrote:
On Mon, Aug 20, 2012 at 12:34:38PM -0400, David
Robillard wrote:
So far I have done all of MCP and all of VCO.
You'll find a call to exp2ap() with a complicated sum as
the argument. One of the terms of that sum, usually near
the end, is a constant. Adding 0.75f to that should
transpose verything from middle C to A (440) - 3/4 octave
up.
Makes sense. Great, thanks.
Also please check that all FP constants are single
precision as well as all math.h functions. I was
rather careless with that when I wrote those plugins.
It matters, in particular if you compile for 32-bit
SSE.
There is a GCC flag that warns about this (dubious float promotions), I
will run them through that.
Concerns about fragmentation?
Cheers,
-dr