[CM] LADSPA implementation

[e deleflie]

I think you're right, its not the LADPSA wrapper stuffing up ... but
its not the LADSPA plugin either....

it seems to be the re-sampling. When I dont attempt to resample the
4ch sound (before passing it through the LADSPA plugin) the results
are correct. When I resample, the results go funny.

I'm resampling from 48000 to 44100 like this:

(src-sound (/ (srate insound) 44100) 1.0 insound)

I'll keep debugging this .... but if anyone can see if my usage of
src-sound is incorrect (to resample a 4 ch file) please let me know
...

Etienne

On 6/20/07, Bill Schottstaedt <bil@ccrma.stanford.edu> wrote:
> > However, the content of the insound is in there with it ...
>
> This must be deliberate on the part of the plugin.  Snd doesn't add
> anywhere -- it sets the input buffer values from the reader, then
> calls the plugin, then writes whatever the plugin returns, again without
> adding anything.  I don't know which ambisonics plugin you're
> using (or perhaps they're all the same one under 1000 names),
> but the "AMB" plugins appear to me to be adding the input back
> into the output (I just glanced at the code, and have never paid
> much attention to this stuff, so maybe I'm confused) -- ask the
> plugin developer what is going on.
>
>