[LAU] LADSPA processing time for audio zeroes

Arnold Krille arnold at arnoldarts.de
Sat Mar 28 12:26:23 EDT 2009


On Saturday 28 March 2009 16:39:35 Ken Restivo wrote:
> When a LADSPA plugin is being sent zeroes-- i.e. when the volume of the
> input data is down to nothing-- is it still sucking up CPU cycles?
> I understand from the JACK API that anytime a client gets a callback with
> data, it has to drop down and deal with the data and then return, even if
> the data is zero. But that could be lightweight, if the data are zero, or
> is that very compute-intensive for the plugin?
> I'm trying to build an outboard effects chain in various LADSPA hosts (JACK
> Rack, ecasound, AMS, others... haven't settled on one yet), and when I've
> got that plugin's volume MIDI'ed down to zero, I'd like it to not be
> dominating CPU cycles at that time.

Actually with double- (or any floating point-) resolution it is rather hard to 
detect exact 0.0 values. And because of the variable exponent, the cpu usage 
for numbers near zero is higher then for numbers near 1 (or any other higher 
value). The keywords to search for are "denormal problems".

In practice this means that while setting a volume to zero "should" give you 
the desired effect of less cpu-usage, in fact the opposite is the case in the 
worst case. When you are already controlling your effects via midi, why not 
bind the "mute" or "active" (don't remember the current naming in jack-rack) 
to a midi controller. An in-active effect shouldn't eat cpu cycles when the 
host-author did his job right...

Have fun,

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: This is a digitally signed message part.
Url : http://lists.linuxaudio.org/pipermail/linux-audio-user/attachments/20090328/455f247b/attachment.pgp 

More information about the Linux-audio-user mailing list