[ardour-dev] issues with plugin in/out counts

John Anderson ardour at semiosix.com
Tue Nov 23 22:46:30 PST 2004

On Wed, 2004-11-24 at 06:36, Paul Davis wrote:
> >On Tue, 2004-11-23 at 15:38, Paul Davis wrote:
> >> in the last few days, i've been staring down some of the problems
> >> people have reported when using plugins that change the number of
> >> active streams in a track (e.g. a 1in/2out reverb).
> >
> >My personal biggest frustration with the way plugins work is related to
> >panning. I have a mono track, and I pan it. Then I add a 1-in/2-out
> >reverb. Now I have to redo the panning. Say I then remove the plugin and
> >add a second 1-in/2-out reverb. I have to redo the panning again.
> but you added a plugin with 2 outputs! how is a panner configuration
> that was built to manipulate a single stream supposed to work when
> there is more than one stream?

Of course it's not, which is why you (who understands stereo panning
*much* better than I do) decided to revamp the panning code ;-) What I'm
saying is that when something like adding and removing plugins causes
something like panning to be reset each time, it says to me that the
design isn't coping adequately with the possibilities it offers.

With the input/pre-plugin/fader/post-plugin/panner/output mixer strip
setup, adding a 1-in/2-out pluing *has to* force the panning to be
reset. Which is why I'm thinking that perhaps panners are the perfect
places to separate the enforcement of 1-in/1-out vs 2-in/2-out plugin
ordering. Before panner, only 1-output plugins allowed. After panner,
2-output plugins allowed, or 2 x ganged 1-output plugins. Following
through on that logic though implies that a stereo width control is also


for mono -> stereo tracks anyway.

It certainly solves the problem of rearranging plugins with different
numbers of outputs. Although to do so it creates a constraint on which
plugins can be used at which stages. But IMVHO such a constraint makes
sense. It doesn't make sense to use a 2-in/2-out plugin on a mono

But now the 1-in/2-out is interesting. In a sense a panner is a
1-in/2-out plugin. So generalising the above:

1-in/1-out pre-fader-plugin
1in/1-out post-fader-plugin
1-in/2-out (panner or plugin)
2-in/2-out post-panner-plugin

AFAIK mixer strips on consoles are essentially 1-in/2-out (in my mostly
theoretical understanding), excluding sends. (And excluding 5.1
configurations, which I know nearly nothing about) So maybe using a
1-in/2-out mixer strip design in an architecture that handles things
other than 1-in/2-out (as ardour does) is creating, er, cognitive

> >The other thing that seems to me doesn't work here is that stereo
> >panners operate as stereo-width as well as panners. That sounds vague
> >because I haven't really managed to get my head around stereo panning
> >yet.
> Thats the core of your problem, I think. Most people's experience of
> panning (including most other DAWs, PT excepted) do not do stereo
> panning correctly.

I'll not disagree with you, having not worked with any other DAWs. This
however can be an advantage - I can think about the design without prior
prejudice and ask the naive (stupid?) questions ;-)

> >So maybe it's worth having a post-panner plugin stage where 2-in/2-out
> This is an idea that has been mentioned before. We might do that at
> some point. You can "fake" it now by routing to an intermediate bus
> with the plugin in its own redirect lists.

That's what I'm doing, although it's a wee bit clunky. I spose one could
see having a post-panner plugin stage as a convenient form of routing to
a stereo bus.


More information about the Ardour-Dev mailing list