Am 17.11.21 um 17:24 schrieb Will Godfrey:
Somehow this appeared on the ordinary users list - might cause a bit of
confusion!
Sorry for that, seemingly I picked the wrong address from autocompletion.
> I got quite a lot of 'missing reference' errors. Is that simply that you
> haven't generated them yet, or did I miss something?
do you mean "Warn: Missing calibration. Can not judge runtime"
If this is the case, then please run the testsuite once with the --calibration
switch, which adapts the timing measurements to your hardware/plattform.
Anyway, it now seems to run correctly. Adding a'..' before 'apply' in padsynth
got that one right too (won't be necessary in the latest master). I notice there
is more 'leakage' on padsysnth than addsynth, and even more on subsynth,
although still more than 100dB down.
Just for sake of completeness: you used a yoshimi build from current master, or
from my "power"-branch? and with -O3 ?
And what numbers do you get for the sound delta on the individual tests?
And, of course, which gcc (or clang) did you use for building?
Sorry for asking so much questions (doh) Basically we still have to find out,
how much minor differences we get on various setup, and how to what degree we
can stabilise these results.
-- Hermann
_______________________________________________
Yoshimi-devel mailing list
Yoshimi-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/yoshimi-devel
