I want to set up a template in Cubase 5 that is sort of like a virtual S*L console. I will have the UAD SSL4K plugin on each track, starting with about 12 tracks, and then have an instances of AlexB's CLC and Panner on each track and the CLC program on the master buss. I have already tried this with a couple tracks and I have to say it sounds great, pretty much it has that glue and polish of my favorite albums from the late 1980's and early to mid 1990's.
The thing is that with 4 tracks and 4 busses I am already hitting the peak processing power on my Intel Q9550 overlocked to 4.0GHz.
So what are some ways to gain more processing power? Are there any adjustments I can make in Nebula3 to increase performance without affecting sound quality?
I was thinking about trying out CUDA but I need to have an ATI graphics card as my primary graphics card because of programs I use for school. Has anyone tried the network computer feature in Cubase 5? I haven't researched it yet but supposedly you can network a computer to the main DAW machine and use the other computer's processing power.
I too would like to know ... as I have a stock Q9450 on an XP-32bit w/4gigs ram.
I primarily do mastering/restoration with WaveLAB. But I also use PiduleVST as an insert so that I can process a stereo track in a vitual split mode. I can then use AlexB consoles, panners and busses, or anything else on each channel of a stereo track.
The effect of this is like a true input to a console. Running the tracks in this manner creates quite the 3D imaging [rather than running in series].
The drawback ... double the number of NEBula instances. So I have to render out quite often.
But I've heard 'stories' of users running a hundred instances of Nebula I want machine build specs of this !!! PLEASE.
It gets said all the time and has already been mentioned in this thread but I'll say it again: render/freeze/apply FX. The biggest downside to Nebula right now is that it will bend your workflow around the limitations CPUs. I don't personally have a new fancy 23 core processor so I'm not completely sure what modern chips will handle but I feel pretty safe in thinking you're not going to load up a full AlexB/tape emulation (20+ input tracks -> 20+ tape programs -> however many tapebooster+ you're using -> 20+ inputs back into the console emu -> buss group programs -> master channel program) on any equipment in the next year or two. Maybe when the server product is ready you'll be able to setup server farms that will make this a reality, can't speak to that.
What Nebula offers is an alternative to having access to classic consoles. Just like working on the real consoles, it has an immediate and real affect on your workflow. For some people this is a deal breaker and I can understand that. Personally, it's worth the time and effort for me, I vastly prefer the results I get with it versus without Nebula. But surely it's not for everyone; I'm sure there are plenty of situations where quality has to be overlooked in favor of speed.
Well I had my buffer set at 256 but now I set it at 1024 and I have about 70% CPU usage with 20 instances of Nebula (non-reverb). I don't like running a huge buffer because I usually track at the same time and I don't like a big delay.