Hi Tim,
Thanks, that's very informative feedback! How does oversampling work,
practically speaking, in audio - I understand the concept from my years
of computer graphics, but it is a little different when you don't
necessarily control the fidelity of the input.
Would the approach to use a sample-rate converter to essentially
interpolate samples, then do the processing, and then sample back down?
How does that work for live streams of data?
Thanks,
Guy.
On 07/09/15 15:44, Tim Goetze wrote:
[Guy Sherman]
And the code is at: https://github.com/guysherman/si-plugins
For high-bandwidth input or high-gain clipping, you'll need to run the
nonlinear operator at substantially elevated sample rates unless you
want synthesised harmonic content to alias audibly. I'm fine with 4x
oversampling for electric guitar at 48k but going higher can still
improve things depending on circumstances.
Also, it's not a bad idea to put a (preferably configurable) pair of
high- and lowpass filters both before and after a strong nonlinearity.
Asymmetric clipping gives rise to a DC component which most sane
people agree needs to be filtered out. Too much bass going in makes a
clipper sound farty, and lowpass filters taking some edge off pre- and
post-distortion are highly useful for musical purposes.
And welcome to the list,
Tim
--
*Guy Sherman*
e: g...@guysherman.com
w: http://guysherman.com
_______________________________________________
Linux-audio-dev mailing list
Linux-audio-dev@lists.linuxaudio.org
http://lists.linuxaudio.org/listinfo/linux-audio-dev
