Re: [Audacity-quality] Spectral edit parametric EQ effect does not complain when zero bandwidth

[Steve t. F. <ste...@gm...>]

On 20 December 2014 at 16:12, Paul Licameli <pau...@gm...> wrote:
> And which Lisp file and version are you quoting there?  I do not see exactly
> that in my version.

SVN source code:
https://code.google.com/p/audacity/source/browse/audacity-src/trunk/plug-ins/SpectralEditParametricEQ.ny

Steve


>
>
> On Sat, Dec 20, 2014 at 10:37 AM, Steve the Fiddle
> <ste...@gm...> wrote:
>>
>> On 20 December 2014 at 15:01, Paul Licameli <pau...@gm...>
>> wrote:
>> > It was harmless that it "merrily did nothing."  It makes sense to have a
>> > warning instead.  That could easily be done in the Lisp instead, which
>> > is
>> > how it was already done for the multi-tool.
>>
>> The error message is done in Lisp:
>>
>> (cond
>> .....
>> ((= bw 0) (throw 'debug-message (format nil "~aBandwidth is
>> zero.\nSelect a frequency range." p-err)))
>>
>> Steve
>>
>> >
>> >
>> > On Sat, Dec 20, 2014 at 8:05 AM, Steve the Fiddle
>> > <ste...@gm...>
>> > wrote:
>> >>
>> >> On 20 December 2014 at 11:11, Paul Licameli <pau...@gm...>
>> >> wrote:
>> >> > Wait, I don't see how Steve's 13786 addresses this.  I thought we
>> >> > were
>> >> > talking about something in the Lisp.  Avoidance of a division by zero
>> >> > in
>> >> > Lisp code was strictly needed only for the multi tool.
>> >>
>> >> Currently, if bandwidth is zero, the spectral parametric EQ raises an
>> >> error message that says that the bandwidth is zero.
>> >> Are you saying that should not happen?
>> >>
>> >>
>> >> Steve
>> >>
>> >> >
>> >> > Equal high and low frequencies might be meaningful for the "shelves"
>> >> > which
>> >> > just make a composition of a high shelf and a low shelf when both are
>> >> > defined, and the threshold frequency is in the transitional part of
>> >> > the
>> >> > spectrum.  So the effect may be nontrivial then.
>> >> >
>> >> > For the parametric eq, I could pass the center and the width to the
>> >> > Nyquist
>> >> > library function eq-band, which did not require me to calculate it as
>> >> > Q
>> >> > which would be infinite.  eq-band seems to be written so as to take
>> >> > zero
>> >> > width as meaning no effect.
>> >> >
>> >
>> >
>
>