Steve Harris <[EMAIL PROTECTED]> writes:
> I don't think this is a good idea myself:
>
> 1) there are dozens of up/downsampling techniques with different
> tradeoffs, so picking the right one automatically is not really
> possible (unlike mixing).
Hi!
I'm sorry I didn't say it aloud, but I really meant that the resampler would
be just another client. I agree that embedding automatic resampling logic to
the engine is not wise.
Now when the resampler is just another client, you can choose which technique
to use.
>
> 2) its too complicated
How come? As I already said, it's basically as complicated as making a dozen
audio type strings called "audio@44100 Hz" etc. Doing it right is just moving
the "44100 Hz" part from the string into a variable.
Another use case: remember LADSPA? Remember how it has an audio signal rate
and a control signal rate, where the control rate is the audio rate divided by
an integer? Now writing a laaga client to do ladspa hosting requires that the
audio input/output ports have a different sample rate from the control
input/output ports, no?
--
-Jarno
