Interesting Channel Strip behavior

[ggegan]

I've been warming up to channel strip lately. I don't like that it doesn't have 6dB/8va filters, but that isn't necessarily a deal breaker. The really interesting thing about the plug is how different settings affect its efficiency.

As an experiment on my HD Native system (6 core Nehalem) I set up a dummy FX mix session with 172 mono tracks 12 5.0 food group auxes with Massey L2007 limiters on each and 12 corresponding 5.0 record tracks. I put 5 mono to stereo reverb sends on each audio track with 5 stereo aux reverb returns with Altiverb on each. Then I populated each of the 172 audio source tracks with Channel Strip and opened up the System Usage window to see where I stood. This is where it got interesting.

If all the Channel Strip EQ bands were set to bell curves and all the individual processing modules were on, I was at 24% CPU usage. If I turned OFF the Expander module, the usage climbed to about 36%. If all the processes were on and I changed the high and low EQ bands to shelf, usage went up to about 54%. If I then turned off the expander the usage went to over 75% and was spiking wildly making it impossible to roll the system.

I can understand how shelf EQ could use up a bit more CPU power, but why would turning off the Expander use up so much juice?

[The Cargo Cult]

shelf behaviour is a bug.
try throwing an Eleven Free or an old waves 8 Q1 anywhere in the signal path of your session - the CPU usage will drop to half or less.

after that you can remove the Eleven and the CPU will be stable.
restart the session and you're back to square 1 with instance CPU usage.

I prefer the shelf as default so I'm on waves 8 with an unused Q1 till there's a fix.

justin

[ggegan]

Wow, that's just plain weird! But what about the Expander issue? That's even stranger than the shelf issue.

[The Cargo Cult]

strange indeed - so maybe also a bug.
if you can write up a recipe we might be able to get some traction on it.

[ggegan]

I'll try to distill the facts and relay them ASAP.