VST Players - Single or Multiple?

I read somewhere about using a different instance of a VST player (I use UVI and Kontakt players) for each synth/patch loaded. So, which is better for machine performance:
One player per synth.
One player with all the synths.

Or does it even matter?

Thanks for your thoughts/insight…


1 Like

Hi Bob,

That is a loooong discussion going on here…

For me it comes down to this:
I use rack instruments (multiple sounds from one instance) for GrooveAgent, as this allows me to split each drum sound to it’s own channel.
I also use Rack instruments for HALion and the KORG M1 plugin as these are multi-timbral… But this is not required, with the new instrument tracks, you can also split up the in and out put over multiple channels… I just keep it up as this is my workflow…

Bottom line is that you can do almost all with Instrument tracks, that you would do previously only with Rack instruments. It all depends on your own preference, and if your hardware can handle multiple instances of the same plugin at once.

There is no real audio advantage when using a plugin as single instance or as multi-timbral instrument.

I hope this helps and maybe other forum dwellers can chime in with their tips etc.

Good luck!

1 Like


At this point in my development, I’m using Rack Instruments for all my VST instruments be they single voice instruments or multi-timbral. I just find Racks more clear to work with. I’ve noticed no real advantage to using Instrument Tracks. but others seem to prefer working with them. As far as CPU overhead, I see no difference between a Track and Rack instrument. I just find it clearer to use everything as a Rack Instrument. I like how the Mixer sets up when Rack Instruments are used. In my experience, CPU usage is more dependent on the complexity of the voice streams (patches, programs) in use.

Thanks, gents… The rack makes things simpler so I guess I’ll continue to do that. I remembered reading something about performance somewhere… I should make a note of things…


For my work flow, MIDI Tracks talking to instruments in ‘Rack Mode’ tends to work out best. It keeps the audio buses of the VSTi on a separate fader on the Mixing Desk, where the Instrument Tracks more or less merge it all into one channel. I also find that MIDI tracks are bit more flexible for my needs (the ability to redirect their output to any instrument I like…MIDI aux sends for layering, etc.).

For me, the exception to this rule is if I want to make “MIDI Loops”. In that case I’ll move everything to an instrument track. (I just make the new instrument track, copy and paste my settings from the 'Rack" instance to my new “Track Instance” in the VST Rack, then use “MIDI/Merge MIDI in loop…” to freeze and merge my regular MIDI tracks down to a single Instrument Track part. Now I can export it as a “MIDI Loop”. If I want more than one instrument track in a MIDI Loop…I just select them all first and then export.

The benefit to 'MIDI Loops" as opposed to just saving regular MIDI files, is that the associated VSTi and its preset(s) is attached to the track(s) you export as a MIDI Loop. This allows you to ‘preview’ them directly from the Media Browser exactly as you made them without having to pull them into a project first. You can later import MIDI Loops exactly as you made them on a new track, or you can pull them onto existing MIDI or Instrument tracks (in which case it’ll try to use whatever instrument you’ve already got set up for the track).

Hi all

I think the performance issue Bobupstairs was talking about was to do with multiple CPU cores. If one uses several instrument tracks Cubase can spread them over several cores whereas a rack instrument loaded up with quite a few “instruments” will only use one core. Remember reading about this on the forum, is this relevant to your concerns Brian? I personally always use quite few instances of Kontakt instrument tracks and don’t have any problems
However I suppose it comes down to how you prefer to work :slight_smile:

Best Regards


I misread the thread Dave…I’d been on a different thread discussing Instrument Tracks VS MIDI to Rack Instruments, and it was about ‘workflow’ differences. Sorry…I got my thread wires crossed :slight_smile:

For my home workstation needs CPU performance hasn’t been an issue thus far, and I usually choose to work with MIDI Tracks and “Rack Mode” instruments.

I have a really old PC based on Hex Core Phenom II with DDR2 Memory (Windows 10 64x). I’ve got it overclocked to about 3.8ghz on all 6 cores (as fast as I could get it without over-volting the chip), and am running the memory bus at stock speed with really tight timings.

A typical session for me is 30 to 40 instruments/voices spread across ARIA, Halion 5, and Groove Agent 4 loaded as Rack instruments. While some of these plugins have ‘manual multi cpu’ options, I’ve just left them at their defaults, which I assume lets the DAW and OS decide how to do it. For VST effects I rarely am using more than 4 or 5 at once, with a heavy convolution type reverb being one of them. My CPU has yet to even break a sweat with these particular plugins. No core has ever topped 20% that I’ve noticed. If I watch Task Manager, all 6 cores usually have pretty even usage with the first a tiny bit higher than the others.

I had assumed the OS was automatically distributing the load across all 6 cores? I.E. Send next instruction to least busy core…wash, rinse, repeat…

For my particular system I have noticed that when I use the PCI based Delta 1010, the CPU use is about 5% lower across the board than when I use the USB based Tascam 1200. No problems with either one though. Since I’ve no way to ‘lock the clocks’ between the Delta 1010 and Tascam, I don’t aggregate the outputs, however, I have aggregated the inputs from both units and CPU usage doesn’t change a noticeable amount as compared to just using the Tascam by itself.

Thanks, Dave. That was the info I was looking for. I don’t have a problem but wondered if someone who understood the under-the-hood stuff had any insight.