Re: [linux-audio-dev] LADSPA and run_adding()

New Message Reply About this list Date view Thread view Subject view Author view Other groups

Subject: Re: [linux-audio-dev] LADSPA and run_adding()
From: Roger Larsson (roger.larsson_AT_norran.net)
Date: Tue Nov 21 2000 - 02:07:18 EET


On Tuesday 21 November 2000 00:31, Paul Sladen wrote:
> On Mon, 20 Nov 2000, Roger Larsson wrote:
> > On Sunday 19 November 2000 12:49, David Olofson wrote:
> > > On Sat, 18 Nov 2000, Kai Vehmanen wrote:
> > > > On Fri, 17 Nov 2000, Paul Barton-Davis wrote:
> > > > > I have to laugh - here's the exact opposite point of view from
> > > > > vst-plugins today, about processReplacing, which is the equivalent
> > > > > of
> > > >
> > > > [...]
> > > >
> > > > > Can both points of view be right ?
> > > > >
> > > > :) Yup, been following the same vst-plugins thread. But seriously,
> > > >
> > > > this guy is just comparing run_adding() and run_replacing() using one
> > > > use case.
> > >
> > > Yep. And we actually have three situations to take into consideration
> > > here;
> > >
> > > 1) plugin taking data from one buffer, mixing into another
> > >
> > > 2) plugin taking data from one buffer, overwriting another
> > >
> > > 3) plugin taking data from a buffer, while overwriting it
> > >
> > >
> > > The third case is what breaks some of the reasoning behind switching
> > > to run_adding() as the default. run_adding() requires extra buffer +
> > > zero operation in chains (3), whereas run() requires extra buffer +
> > > add in mixers (1). run_adding() also requires an extra zero operation
> > > to overwrite a buffer (2).
> >
> > With run_replacing you may do 3) in chains with only one buffer !
> > If the engine that allocs buffers know that a plugin is never looking
> > back - impossible with run_adding !!!
>
> The difference is, that the audio-server/host can *patch up* for the
> plugin only suppling run_adding()... not so, when you only provide the
> facility to overwrite the current buffer. The server knows that all it
> has to do is to add a call to "zero" before execution the plug.

I do not follow you here.

Why shouldn't a server be able to insert mixers by itself if you connect
outputs? The scaling? But then it would be the user that needs to scale
before adding - possibly only scaling a few of many; like one of four.

/RogerL

-- 
Home page:
  http://www.norran.net/nra02596/


New Message Reply About this list Date view Thread view Subject view Author view Other groups

This archive was generated by hypermail 2b28 : Tue Nov 21 2000 - 02:49:56 EET