Re: [yoshimi-user] [graggrag.com] Yoshimi keyboard
- From: cal <cal@xxxxxxxxxxxx>
- To: tom <tompotts@xxxxxxxxxxxxxxxxxxxx>
- Date: Thu, 17 Mar 2011 20:50:16 +1100
On 17/03/11 02:57, tom wrote:
I cant guarantee anything (time pressures etc) but if you can point me
to the code you think may be at fault I'll have a look at some unit (and
or other) testing for it if I can get my head round it.
Tom
Thanks Tom, much appreciated. If I could point to any specific code I'm pretty
sure
I could fix it :-). That said, the problem is ADnote specific, and that's where
all
the unison changes lie, so Synth/ADnote.cpp has to be the main area of
interest. In
getting to this point we've tried logging the calculated note frequencies at
every
point I can identify, and they've all seemed correct. I simply can't identify
where a
frequency calc is coming out wrong.
I keep coming back to this point:- on a system where a local build fails, we
can drop
in a binary built from the same code on my system and it pitches perfectly.
We've
established that more than once. The only theory I can come up with on that
point is
that there's some subtle difference in the way maths functions are being linked
during
the builds. As best I can figure, there's something strange going on with the
linkage
of math lib functions on 32 bit. If the app code was indeed defective, then
surely at
least one 64 bit build would be affected.
I figure I'll tidy up the code I was using for the AV Linux testing over the
weekend
and put that up sometime soon. Health status is utterly crap at the moment,
hence the
evidence of grumpiness (and frustration). Sorry 'bout that.
cheers, Cal
Other related posts:
