[PD-dev] pd-double builds are actually double now

[katja]

On Mon, Oct 10, 2011 at 6:56 PM, Hans-Christoph Steiner <hans at at.or.at> wrote:

> I was thinking it would be nice to have examples of what double precision
> enables with audio.  I guess its pretty subtle for the most part, but your
> examples are good on your site.  I like the idea of Pd-double now becomes
> useful for general purpose math.  And the IR measurement example is a good
> one too.  Perhaps this doesn't really exist, but I was thinking it would be
> great to have a patch where you could hear the difference between single and
> double precision.

As the examples on the site suggest, you have to go to extremes before
single precision becomes insufficient. The attached patch
long-audio-buffer.pd illustrates what happens if a ~ 15 minutes
soundfile is played at fractional speed in Pd. Double precision can
handle that, single precision can't. At this length, the difference is
not subtle. Let me know if this example works for you as a demo.

Katja
-------------- next part --------------
A non-text attachment was scrubbed...
Name: long-audiobuffer-test.pd.zip
Type: application/zip
Size: 1283 bytes
Desc: not available
URL: <http://lists.puredata.info/pipermail/pd-dev/attachments/20111010/ae71a60c/attachment.zip>