I’m still slowly building up my kit and continuing to agonize over what ought to lie at the end of the roadmap.
I’d love to know… who out there is using Gig Performer and computer hardware to its utmost, spending big to max out the hardware possibilities, and monitoring post-GP? Where are the sane, sobering limits of a single PC for hosting such a setup?
My setup currently involves three instances of GP on separate computers, and I am trying to decide whether I need a fourth, or whether I might consolidate the system to use as few as two.
A ‘Show Control’ mac, which receives song selection commands, calls up presets on the other computers/instances, manages midi signal flow, lighting automation, and provides clicks, tracks, and an automated looper for all band members. (This will always remain separate, as it can be set up with lots of buffering and latency without affecting performance… (no “real-time” processing goes on here.)
A ‘Synths’ PC which handles synth patches for me (2ch Stereo ‘Synth’ plus 2ch stereo “Bass Synth Buss” for the FOH guy’s control if present) as well as for the other band members (4 additional channels as two stereo pairs.) This is an old tower PC I was gifted. (i7-2600 w/ 32GB DDR3 RAM at 1600MHz) I’m pushing the limits of this machine, even though I use physical-modelling plugins wherever possible. However, there is an additional major limiting factor in that I’m using Dante Virtual Soundcard as my audio interface. This requires me to keep the latency buffers extremely low, as DVS already adds 6ms-10ms or more of latency to my pre-monitor feed.
A ‘Drums’ PC running MODO Drums and a Sampler plugin. (i7-3770 w/ 32GB DDR3 RAM at 1660) This also uses DVS as an audio interface, but processor/RAM demands are much more reasonable… has not been a big problem thus far
In the works:
- I want to add vocal and guitar(s) processing strips. These would include
-4x full-featured vocal strips (gate> mild pre-compression> subtractive EQ> de-ess> Transducer/Exciter> multiband compressor> tuner/harmonizer> ambience) the output of which would feed the board (pre-monitor), as well as supply the formant feed for guitar/keyboard part vocoders.
-4x Guitar strips: One for post-pedalboard compression>impulse response/cabinet emulation/mono>stereo for guitar, One similar track for a bass, and another pair for full pedalboard processing of Bass or Guitar or (2-channel) Chapman Stick, as the song demands. Typically only two of these channels would be used simultaneously. (One bass and one guitar) unless I’m playing stick (One guitar and Stereo stick) The non-bass tracks would also have vocoding available via the vocal feeds. Again, this is all pre-monitoring.
There are some slight efficiencies to be gained by consolidating all of this. For example, the formant feeds would exhibit less latency if they were to avoid a trip via audio interface. I will only be playing one instrument during a given song (Synths, Bass, Guitar, or Stick,) so a monolithic approach would never need to process those paths simultaneously. The same goes for my lead guitarist, who would play guitar, bass, or synth; never more than one at a time. (Although a potential added member would require doing more parallel work.) Drums and Vocals are typically required at all times.
Given the above, should I hold out any hope that instances #2, #3, and #4 might be possible on a single machine? What about a maxed-out Mac Mini or (to take my ludicrous plans even further beyond reason) a bottom-of-the-line Mac Pro? Are there rules of thumb for figuring out how much additional function I’d gain by purchasing current hardware?