[PD] dac~ workings

[vanDongen/Gilcher]

On Thursday 30 September 2004 17:30, Ian Smith-Heisters wrote:
> One of my students was wondering how dac~ works under the hood, and
> since I couldn't figure it out, thought I'd ask y'all.
>
> Is there a difference between these three patches (the patches are below
> but I'll describe them for your ease):
>
> 1. Two osc~, each connected to both channels of the same dac~

for each channel the dac~ sums its inputs

>
> 2. Two osc~, each connected to both channels of two separate dac~
>
each dac~ mixes its output into the outputbuffer I think, but I would have to 
check the source to see how it is done.

> 3. Two osc~, each connected to the same +~, which is connected to both
> channels of a single dac~
>

here the mixing/adding is done by the +~ object.

So the differences under the hood are the place the adding is done.
With simple patches like this there is little difference.
With complex patches there might be block-length delays between different dac~ 
objects. But I am not sure when that happens, its in the docs somewhere. But 
that only happens when the different signal objects are in 
subpatches/abstractions anway. 
If there is a performance penalty?
I think that connecting everything to a single dac~ is the most efficient, but 
that the performance hit from doing it differently is very minor.


Gerard