Re: [linux-audio-user] Jack Rack & machine speach questions - (WAS:multiband compression with ecasound and ladspa)

On Thu, 20 Jan 2005 15:31:02 +0000, Steve Harris
<S.W.Harris@email-addr-hidden> wrote:
> On Thu, Jan 20, 2005 at 07:00:32 -0800, Mark Knecht wrote:
> > On Thu, 20 Jan 2005 07:22:45 +0000, Steve Harris
> > <S.W.Harris@email-addr-hidden> wrote:
> > > On Wed, Jan 19, 2005 at 09:58:40 -0800, Mark Knecht wrote:
> > > > My first idea was to use the info provided earlier that Jamin is
> > > > really using ladspa plugins and consider Jack Rack and a simple set of
> > >
> > > For clarity: its not just plugins, the EQ and crossover sections are not
> > > plugins, and probably shouldn't be.
> > >
> >
> > I didn't mean to imply that Jamin was only plugins, but rather that I
> > think you said that the compressors were so thee should be no
> > differences using them in Jack-Rack or Jamin.
> >
> > By crossover do you mean the section that divides the input signal
> > into the number of bands you want? Why should that not be a
> > specialized plugin or even a collection of stand alone linear phase
> > bandpass filters? What's lost in the audio by doing it that way?
>
> Yes. Its quite complicated, and the way its burnt into JAMin shares some
> of the CPU load with other processes, as an independent plugin the CPU
> load would be heavy.

OK, point taken. An optimized design will always win in that way.
However Moore's Law says that CPU performance will continue to up so
much that I'm guessing this won't matter to me. I can already run
about 10 C1's on my 3GHz XP machine so I'm guessing I have enough
processing power to do it either way.

>
> That said there are more appropriate designs for a crossover in a plugin,
> I just dont have the code around.

Maybe we can look into that sometime in the future. Again, I'm
presuming here that there is some value in being able to hook together
processing units in a user-desired way as opposed to accepting any
specific prebuilt collection and the limitation that come with it.

Thanks,
Mark
Received on Thu Jan 20 20:15:15 2005