Problem with DSP "Not Enough DSP or Time Slots"

[hef]

getting this DSP error first: "Not Enough DSP or Time Slots"
and then "The plug-in could not be made active because a DAE error was encountered (error #-7224).

this is my first big surround project. i'm working on pretty good system(at least i thought) HD3 with 4 192's (d-control). i'm using 160 voices (5 DSPs). i could go to 6 DSPs- but i was there before and was having the same issues.

my hardware buffer is 1024 samples. 90% cpu usage limit. delay compensation: short
my dae playback buffer is set to Level 2(default). is this where my problem lies?

i'm on a dual g5 2.7 gig with 2.5 gig ram.

this is a 52 min project. with 5.0 music stems and a lot of sfx. but recently a similar lenght movie was done in this same studio and didn't encounter these issues. i think its the settings.

any ideas anyone? thanks.

[Richard Fairbanks]

Yeah, well, you're probably nearly out of DSP resources. I've hit the wall twice this year, both times the DSP window made it appear that I could fit in one or two more plugins, but I knew I was near the edge. In one case I was able to turn off premix tracks, printing them to new tracks without as much DSP, and deactivating them. The other time I was able to reduce the overall number of tracks by sifting through the effects and music stems to squash them onto fewer tracks. The advantage of reducing tracks is that you can get back the plug-in resources being used and sometimes also reduce the number of DPS chips required by Pro Tools to combine them all together. Surround has many more separate signal paths than stereo, especially when you include lots of internal bussing, and that takes more DSP.

Is there any chance you can run with 96 voices?

[hef]

thanks.

"run out of dsp resources" can i buy some more. does cingular cell rollover dsp voices? hahaha.

yeah i have been cleaning out unnecessary tracks. i probably need to do some more cleanup. its definitely better, but i'm not sure if it was from getting rid of some tracks and plugins and buses or if it was because i changed playback engine settings. i went to level 4 cache. does playback buffer have anything to do with this? i'm bad with this backend crap with protools, only been using it for like a year and a half.

i could try 96 voices and see if that helps. i've gotten it to the point where i can add plugins fine, but when i go to add a 5.1 bus to an aux it freaks out and reallocates resources, but doesn't crash. i'm going to need to add bus assignments to my stems in order to print M&E's and other splits before i'm done with this. don't want to be maxed out now.

maybe i'll have to bounce my sfx down before i make splits. that would free up a ton.

oh the joy....still beats working on the SSL Screen Sound system i used to complain about.

[Richard Fairbanks]

I don't know if playback buffer size is a factor in this game. 1) More plugins=more DSP. 2) More things that have to be combined together=more DSP. That second one is the one that is easy to overlook. As you add tracks you have to remember that every output signal (6 of them for a mono track through a 5.1 panner, multiplied by the number of tracks, each of which may be assigned to analog outputs so you can hear it, internal stem busses, those stem busses combine in different ways to submixes, and so on). Effects/cue sends are additional requirements. For each output or buss or effect send, Pro Tools can only combine a finite number of things together for it. Once that number is exceeded, more DSP is grabbed to be able to combine a bunch more. So adding just one more track or plug, if the numbers are high enough, cause a reshuffle and more DSP to be required just to keep the session running. That one thing can have a real consequence! Because surround mixing can require so many more output paths for each track, combined with the tendency to need more tracks because there are more speakers to use, you can see how your resources get used up quickly. Frankly, I think it is marvelous that we get any work done!

Oh yeah, ADC requires even more. I'd turn that off and shift tracks to do it the old fashioned way, then reduce the number of engine voices and see what happens. Also, eliminate panners where you don't need them. Little by little you can put your DSP to use where it does the most good.

[TheHenchman]

We always use 2 separate systems.
5.1 mixers and auxes eat up resources fast.

[hef]

ADC? you mean the delay compensation, right?

[hef]

2 separate system? wow! you mean 2 separate computers with 2 protools, one chasing the other?

[minister]

2 separate system? wow! you mean 2 separate computers with 2 protools, one chasing the other?

he does! lots of medium to big films are mixed on 2 or three systems.

i have an HD|ACCEL 3 and i am running 192 voices with 4 DSP's to mix this feature. 95 audio tracks and auxes for DX, PFX, FX, BKD, MX, LFE send 3 OUTBOARD REVERBS, 2 ALITVERBS, DOLBY SURROUND TOOLS for Ls/Rs generation, 2 hardware inserts... and a bunch of plugs...

my system usage is really full... i did print a few EQ's on some FX...

i was having issues for while, but it is all working now. so, it can be done. with out the DSP time slot warnings, but you may need to reduce your RTAS usage.

i might want an HD4, but, i would actually want a second system.

[dr sound]

We mix Features with 2 HD 6 Accel systems. Yesterday on the film we were mixing we had a total of 5 systems on the stage all chasing. I mix Dialog and Music on mine. My partner Chris Jacobson mixes FX and Foley. We had a dialog system for the Dialog Editor, a FX system for the FX Editor and a Music system for the Music Editor. It was quite a site in our machine room!

[TheHenchman]

2 separate system? wow! you mean 2 separate computers with 2 protools, one chasing the other?

Absolutely.
I think at the end of the day it's also cheaper than going the expansion chassis route.
I couldn't imagine doing any full 5.1 mixing on a single HD3 system.