Many VSTis which are purchased seperately (meaning not included with your sampler software) usually include a "player."

If you intend to play the samples live, it would be best to load this player and load the instrument sample that you intend to use. Most of these players will allow you to adjust all of the parameters of the sound (envelope, attack, release, sustain, amplitude, pedal purpose(s), etc.)

This is without a doubt the best way to go. The fewer resources used on the system, the less likely a crash or conflict will occur. Incidentally, a lot of smaller (meaning smaller budget) orchestras around here use my stage piano and samples of larger pianos or harpsichords or celeste and this is exactly how I do it. Garritan or Ivory include these players.

However, if you intend to record, you have a different situation altogether.

You may in fact load your sampler/recording software and load a midi track. In the midi track you want to identify the output as your VSTi that you intend to use and then identify an audio input track in which this is to record. (You can actually skip this last step in some sequencers and do it later. That way you can tweak the final sound after the performance - very helpful).

You'll want to be cautious of your latency settings. Too low and you all but assure yourself of a crash. Too high and the keyboardist will likely want to do you bodily harm since they'll play a note and a half second later, they'll hear it come out. (This will seriously freak out ANY pianist or keyboardist!)

Another option is to have 2 machines. The first machine acts as your sequencer and midi box. Your second machine acts as your recording machine. Run midi into your first box and load the VSTi using the included player. Feed the audio out of the first box (via digital preferred) into the second box as an audio track. This option has a couple advantages and disadvantages.

Advantage -
Lower system resource requirements mean more stable system and lower latencies.

Disadvantage -
You're stuck with the sound you get so you had better get it right. (However, you may also *record* the midi stream entering box 1 on either box 1 or box 2 and then load the VSTi later during the post-production phase to counter this problem. If I were to do this, I would simply run a redundant midi stream to both boxes. Box 1 would simply act as the midi sound module and Box 2 would record the midi stream for use later. You will need a midi patchbay for this such as the MOTU Timepiece or other similar device.)

Does this help?

J.

you should check this out
http://www.native-instruments.com/index.php?id=kore2_us
it's specifically for your purposes. you can use any vst plugin within it's software, create layers and splitzones and use vst effects as well.

never tried it though, but it seems to be a very nice tool to use vst's live.

I saw Herbie Hancock a couple of weeks ago, he was using Logic Pro's integrated instruments for his Rhoades and Clavinet sounds.

8)

There is only few software that may be the right for you.
My usual live-setup today is a Sony Vaio Notebook, a small and
a large masterkeyboard and optional a Fender Rhodes,
Clavinet, Korg organ or synth(style-dependent).

By now i used Brainspawn forte as my host for VSTi.
It´s kind of a virtual Rack you can fill with your VSTi s, adjust
MIDI-channels keyboardsplits and controllers, then save it as
a preset. There is a song-list you then can combine the presets
to a song and manage the songs in setlists.
While performing you can step through the setlist by hitting space
on the computer-keyboard or remotely via MIDI-program change.

I just purchased the NI-kore package, it includes a controler/audio-
interface combination and the Kore-Software.
Here you can do the same as forte does, though there are some
small limitations that you can workaround and will be fixed in the
upcoming kore2-software release.
But Kore is more than just keysplitting, it is a controller to
manage your live performance just from this cool device,
as it displayes the names of the controllers and features user-
pages for them so you always know each knobs function.
You step through your songs via the controllers up/down buttons
and immediately have access to the parameters you want
as long as they show up in the plugins VSTspecs.
User Pages can content Parameters of different plugins at once.
The software also features a sound-database function that
lets you choose sounds by attributes regardless of which plugin
creates it.
The audio interface has 2 line outs (analog/digital) and a headphone
out that i use as an additional 2track line out.

In comparison, forte always served me well, and as my songlists
easily extends 30 songs i didnt find time by now to reprogram
all my song-setups with kore, but all new shows i do with kore
allready.
Brainspawn forte costs $130 and is easily worth it, though, for
$170 more (NI kore1) you get a nice controller/audio-/midi-interface
plus the software and a free upgrade for kore2 software.
The Kore2 package got different hardware that doesnt have the
audio-interface any more and will be more expensive.
(funny what companies come up with some times)

There´s much more to tell but I hope this helps at first.

cheers
decode

http://www.brainspawn.com/products/forte/
http://www.native-instruments.com/index.php?id=kore1
http://www.native-instruments.com/index.php?id=kore2