what the f@#!!

wankee · Post by **wankee** » Fri Oct 13, 2006 8:30 pm

in live 6 i can't make any larsen by self feeding the returns tracks anymore!!!!
this is very bad news

Machinate · Post by **Machinate** » Fri Oct 13, 2006 8:31 pm

try right-clicking on the sends and hit enable.

kabuki · Post by **kabuki** » Fri Oct 13, 2006 8:32 pm

hmmm. that is a problem.

i think.

wankee · Post by **wankee** » Fri Oct 13, 2006 8:35 pm

YEEEEHAAAWWWW machinate you make my evening :]
appologizzez for my ignorance

Robert Henke · Post by **Robert Henke** » Sat Oct 14, 2006 7:56 am

technical background:

disabling sends in return tracks reduces the amount of internal feedback connections. this has two good effects:

a) it reduces CPU load
b) it makes latency compensation easier. under certain conditions the crossfeedback introduces a significant und inevitable increase in overall latency.

This is why we turned them off by default in Live 6.
( if you want them all on per default just create a template set )

Robert

Johnisfaster · Post by **Johnisfaster** » Sat Oct 14, 2006 8:00 am

Robert Henke wrote:technical background:

disabling sends in return tracks reduces the amount of internal feedback connections. this has two good effects:

a) it reduces CPU load
b) it makes latency compensation easier. under certain conditions the crossfeedback introduces a significant und inevitable increase in overall latency.

This is why we turned them off by default in Live 6.
( if you want them all on per default just create a template set )

Robert

so robert, does this cpu reduction apply even when compared to a send that is set to negative inf? as in, disabling a send is more cpu efficient than turning it all the way down?

hambone1 · Post by **hambone1** » Sat Oct 14, 2006 8:05 am

Johnisfaster wrote:[so robert, does this cpu reduction apply even when compared to a send that is set to negative inf? as in, disabling a send is more cpu efficient than turning it all the way down?

Try it and see!

Johnisfaster · Post by **Johnisfaster** » Sat Oct 14, 2006 8:10 am

I would have, but I am kinda assuming that the cpu increase is so minimal that I may not even be able to see it on the cpu meter, I was thinking that if it was an increase in cpu it would be a more behind the scenes increase

Johnisfaster · Post by **Johnisfaster** » Sat Oct 14, 2006 8:16 am

just tried it with a whole bunch of sends containing reverbs and a breakbeat, though all of the sends were set to inf. disabling and enabling made absolutely no change to cpu that I could detect. so I was hoping robert could clarify.

Robert Henke · Post by **Robert Henke** » Sat Oct 14, 2006 10:02 am

Disabling saves CPU because we might not need to calculate delays for latency compensation. turning the knob to -inf might save CPU because some effects do not need to calculate. I assume it does not matter so much in practise.

Robert

Pitch Black · Post by **Pitch Black** » Sun Oct 15, 2006 4:43 am

Robert Henke wrote: disabling sends in return tracks reduces the amount of internal feedback connections. this has two good effects:

a) it reduces CPU load
b) it makes latency compensation easier. under certain conditions the crossfeedback introduces a significant und inevitable increase in overall latency.
Robert

This is a new revelation for me, as I've been using live sets of 110 tracks, bussing through six subgroups that I originally put in there to speed up hardware output assinging. I'm going to investigate if routing all tracks direct to the hardware outs gives a significant performance boost...

I'm guessing that it should? I never alter anything on the busses themselves, so I don't actually need them.

*dashes off, pulling on lab coat*