I know this is an older thread, but I thought this info might be useful.
 
I use Guitar Rig 5 pro for all my guitar work, and of course I like to run it as close to real time as possible.
 
With Double Precision on, I was getting crackles occasionally on my lowest latency settings, but when I turned it off, it was perfectly smooth.
 
My system is a not so great AMD quad core build with the Athlon ii X4 631 processor (2.6ghz). and guitar rig seems to not multi thread very well (when monitoring, one cpu would be 50%+ while the others were below 5% most of the time). so for this application, gaining that bit of processing by turning it off was extremely useful, as it may be for anyone who runs vst amp sims or other similar effects.

jasonringer
I know this is an older thread, but I thought this info might be useful.
 
I use Guitar Rig 5 pro for all my guitar work, and of course I like to run it as close to real time as possible.
 
With Double Precision on, I was getting crackles occasionally on my lowest latency settings, but when I turned it off, it was perfectly smooth.
 
My system is a not so great AMD quad core build with the Athlon ii X4 631 processor (2.6ghz). and guitar rig seems to not multi thread very well (when monitoring, one cpu would be 50%+ while the others were below 5% most of the time). so for this application, gaining that bit of processing by turning it off was extremely useful, as it may be for anyone who runs vst amp sims or other similar effects.

This is because your system could not handle it at that latency setting. I use Guitar Rig 5 without any crackles a latency of 1.2ms and with the 64nbit engine on.
You can leave the 64bit engine off and just enable it upon bouncing and exporting and you will achieve all the benefits of it. This is because when you record in sonar, its non destructive, so Sonar doesn't print anything. But its goosd to have it always on, so you can hear it in real time
 
CJ

It is clear there is not a consensus upon its need or benefits, if any.

There are some plug-ins which fall over badly with it enabled (one of the ReValvers that shipped with an old version of SONAR for one), so having the option is a good thing.

exactly CJ, my system couldn't handle it, because it's not so great. turning it off helped me with guitar rig 5, where low latency was important.
 
for those of us with lower end systems, it is helpful to turn it off when running high demand vst's at low latency. that's all I was saying.
 
I will be turning it on for mixing and mastering and what not, but as I haven't noticed an audible difference with it on/off, i'll leave it off for recording to free up some headroom on my crappy processor.

"recursive"
 
I read the definition. But still I no understand.

Recursive means there's a loop in the processing where the results of a calculation are fed back to the input and have further calculations done on them, which then loop around and are processed again, which then loop around and get processed again, then loop around and...
 
Thus after thousands and thousands of iterations any tiny errors that occur can accumulate enough to become significant.

Wow, flashbacks....
 
I only ever use the 64bit switch during exports, and only if I think the content might benefit.   Unless I'm freezing or bouncing a lot of tracks and still think that the content might benefit.  I'm using an older Q9300 so CPU is still a factor.
 
Are you recording a live band with distorted guitars, heavy compression, saturation, etc?  64bit will probably have little to no benefit.   Recording a number of acoustic instruments and using lots of EQ or other plugins? Or perhaps creating really clean electronic compositions with lots of soft-synths and FX?   Those two scenarios might be better candidates: the more ITB processing and mixing is required on projects that are intended to be clear, detailed, or even "lifelike" (particularly hi-res formats), the more that additional precision just might have an audible result.  No harm in trying, right?
 

drewfx1
Recursive means there's a loop in the processing where the results of a calculation are fed back to the input and have further calculations done on them, which then loop around and are processed again, which then loop around and get processed again, then loop around and...
 
Thus after thousands and thousands of iterations any tiny errors that occur can accumulate enough to become significant.

So IOW the more you mess with something the more accurate it will be?! Interesting. Kinda like multisampling maybe?
Compute it over and over again so much that the inaccuracies take to flight.
... or get bigger?

soens
So IOW the more you mess with something the more accurate it will be?! Interesting. Kinda like multisampling maybe?
Compute it over and over again so much that the inaccuracies take to flight.
... or get bigger?

Actually the best analogy I can come up with is analog or tape delay (or a simulation thereof): 1 or 2 delays are relatively clean, but if you crank up the feedback the repeats gradually get more and more distorted. That's essentially the idea.