Also, make sure that the Realtek chip does native 44.1...
I feed my MOTU card from the SPDIF output of a Creative Audigy* and I had to change my workflow to 48k because the Audigy turned out to "fake" 44.1 by constant upsampling/downsampling around its fixed-48k DSP.

* because the FireWire connection is so fragile and the damn MOTU freezes all the time

Andras


On Fri, Nov 11, 2011 at 04:38, Ingo <ingo@miamiwave.com> wrote:
I think if you are using a mixed patch with osc and samples only the samples
would play correctly and any osc~ might get detuned. Not sure?

But anyway, as far as I know you can only up or downsample by the power of
two.

I did try to use a frequency multiplier of 0.91875 and it was playing in
tune but that still gave me a samplerate of 48k at the SPDIF out instead of
the 44.1k which was what I wanted.

Anyway, I suspect that some instabilities as well as certain losses of sound
quality may happen if the sample rate of Pd conflicts with the sample rate
of the sound card. This is because of the resampling that the sound card
needs to do. So getting Pd's sample rate matched up with the rate of the
hardware might have advantages even if you are using only analogue outs.

Ingo


2011/11/10 Roman Haefeli <reduzent@gmail.com>
On Thu, 2011-11-10 at 17:14 +0100, tim vets wrote:
> ..
> Lastly: I wonder if there isn't a way to downsample some subpatches to
> playback the 44.1kHz soundfiles in a 48kHz environment?

Why would you want to run an [osc~ 440] at a different samplerate, when
it plays a 440Hz anyway?

Regarding audio samples, you can use [tabread4~] fed by a [line~]
instead of [tabplay~] for up- or downsampling.
 
 
it was just a thought, I can imagine if you would have based some
sophisticated sample playback on a whole bunch of tabplay~'s or readsf~'s,
that maybe you wouldn't want or have time to change all that...
I checked [switch~] again and indeed you can enter 0.5 to downsample by
factor 2, does that mean you could enter 0.91875000000000007 to downsample
from 48 to 44.1?
Roman