Hi is this the the correct or optimum audio engine settings for Cantibi
le 3
Hi grossmano,
It looks like you have good settings, however, if you don’t require the 64 bit double precision audio it could be left off and give back some processing power to other needs. You can experiment with the buffer size as well but at 256 samples it is in the default safe setting.
Dave
Thanks Dave ,
I will try that as I need all the processing power that I can get. In fact I don’t even know what 64 bit double precision audio is , and in what situation I would need that . Any advice on that would be appreciated. Thanks again ,
Michael
Hey grossmano,
mostly, in VST environments, audio data is encoded in 32 bit floating point numerical format. Brutally simplifying, this gives you 24 bits of precision and 8 bits to scale it to adapt to very quiet or very loud passages. In theory, this gives you a dynamic range of 1680 dB, whilst human ears only have a dynamic range of about 150 dB until serious pain sets in…
So, 64 bits of resolution give you double this range - not something you’d practically need in terms of dynamic range. But an important point is that even while 32 bit floating point can cover such a broad dynamic range, its PRECISION remains at 24 bits - so anything after the 24th bit is cut off, so when things become very loud, some resolution is lost at the lower bits. This is the reason why doubling the number of bits may be useful in getting more precision, even if you don’t need the added dynamic range-
But given that 24 bits still give you a range of 16,777,216 steps for your waveforms, this is still pretty good resolution, so you would only only need 64 bits resolution in extreme exceptions.
For all practical purposes (especially live use), you should be good leaving 64 bit double precision audio turned off.
Cheers,
Torsten
1 Like
Yep, if you want to ease the load on your processor you can try making this larger but I find around 5-6 ms to be the butter zone before the latency makes playing feel spongy to me. And conversely, if the reaction time of your fingers to the sounds coming out feels natural don’t make your system work any harder by making it shorter. (Now if you play guitar through the rig that might be a different story…)
1 Like
Thanks for the tip. Will check it out . Have a great day, michael
Just to add to what @torsten said about the double precision…
one thing to remember is that whenever the plugins and host are at different precisions there needs to be a conversion between single and double precision which is an extra step in the processing pipeline.
-
When double precision is off, this practically never happens because all plugins support single precision (although some might internally be running double precision and doing their own single/double conversion but I can’t really comment on that).
-
When double precision is on, any plugins that don’t support it will have all input samples converted down to single precision and then the output converted back up to double precision.
So, if you’re using a lot of plugins that don’t support double precision I highly recommend not using double precision because of the extra load to do these conversions.
You can check if a plugin supports double precision via the insert plugin dialog “32-bit audio only” or “32 and 64-bit audio”
If all (or most) of your plugins support double precision then as far as Cantabile is concerned to performance different should be negligible, and possibly sound quality better (though I doubt you’ll notice it in any live situation).
ie: unless you have a specific reason to need it, leave it off.
Brad
1 Like
Interesting - I have a ton of effects and MIDI utility plugins that are both 32 and 64 bit audio, but not a single instrument.
So, I suppose this 64-bit precision might be useful if all I were doing was to run external synth boxes through Cantabile 3 for effects, or only run my MIDI controllers through Cantabile 3 to send out to such hardware boxes?
How serious a hit does a synth take being run at the 64-bit double-precision setting if it is 32-bit audio only?
Below are two pics of all the “32 and 64 bit audio” plugins I’ve collected (a partial list of the midi ones from pizmidi) as an example:
I suspected as much but it’s cool to have it be official…
It’s negligible but could add up if you’re using a lot of plugins. eg: suppose you’re running in double precision mode and you have a chain of single precision only plugins. Cantabile will do this:
Input signal → convert to single → plugin 1 → convert to double → convert to single → plugin 2 → convert to double → etc…
ie: there’s no smarts to eliminate the redundant conversions.
1 Like
I get it. But in the event you were running five instruments and a couple of them used 64-bit audio capable effects like the ones showed in my list, that would work out fine with only one conversion? (The single output of the plugin to the one or several double-precision capable effects?)
Just wondering if the “32 and 64-bit audio” effects would rather be running in the double-precision 64-bit environment, primarily.
Terry
That’s really a question for each plugin. Ideally they should have separate code paths for single and double processing but it’s possible some plugins might do all processing in double and provide their own internal conversion from single to double to single for the required single precision API.
How on earth would a typical earthling like myself ever be able to tell the difference??? 
Terry
That’s what I say.
post must be 20 characters ok
[Chapter One of Moby Dick deleted]
1 Like
You wouldn’t and typically you don’t need to worry about it. Just use single precision unless you have a good reason to want double processing (and for live work I question whether there’s ever a good reason).
3 Likes