The latency I'm describing isn't to do with syncing up with instruments.
I'm mentioning latency so that you don't get any latency induced delays on your voice from when you sing a note to when you hear it. I'm sure you understand from using Logic how sample buffers work with a DAW and a soundcard and latency. If not, look it up in the Logic manual before you go any further! This latency isn't a problem for any audience but it can throw a singer off if done incorrectly.
What you are describing is plugging your vox mics through the firestudio, have it convert it to digital, then apply fx & levels in the DAW, then back to the firestudio to convert it to an analog signal again before it then goes out to the P.A.
It is possible to run everything through Live before going to a P.A. with no noticeable latency. But there are a few things you will want to do or think about if going the 100% Live thang.
----(& I would choice no other DAW than Live for this. For obvious reasons but one thing that I love about Live is that it always puts its metronome & audio engine above all else, even if you max the cpu & it glitches for a second it will keep on going and sacrifice graphics processing for the sake of audio, no pop up windows will ever stop the music. Logic, (and as far as I know Mainstage) will kill audio & give you the annoying audio engine or RAM message. If you haven't already got Live then that would be the only time I would consider Mainstage. As although it looks cool and superficially can do SOME things slightly easier than Live. In my experience it is a bit of a toy and struggles like Logic in being powerful but VERY prone to crashing, ableton 8 with all the updates is fantastic interms of what you have to do during a set to crash it)----
Setting your buffer down to 32, 64, 128 or 256 samples isn't the only step for latency reduction! Be wary of plugins that will trigger extra Auto Delay Compensation. If you don't already understand delay compensation then look it up immediately. You will have to be cautious of plugins that induce their own form of latency. Some plugins are called "sample accurate" or "zero latency" etc. This basically means that the plugin can do all its calculations on your audio and send out the desired effect before the setting on your buffer runs out. Plugins that aren't sample accurate will take a few cycles of your buffer to do its thing. Plugin Delay works by delaying all audio processes so that they are in time with the slowest plugin and so all the audio comes out together (that is REALLY oversimplifying but look it up)
Most of Live's own plugs should be sample accurate but I do not know of that exactly. Live doesn't say in the DAW the level of delay its plugins report to the DAW. But I think there might be a utility plugin you could download that can help??? Somebody else might know this.
Plugs with just a little latency are o.k. but I would stay well away from convolution reverb etc (especially that brilliant Nebula EMT Plate verb)
With only 8 mics you probably will be able to get your buffer down nice and low whilst not running out of CPU power (lower buffer = higher CPU, also you won't need any higher than 44.1 Khz sample rate in a live situation, (or any situation many argue) but raising the sample rate reduces the latency at the cost of higher CPU so its also a judgement call depending on how intensive your sets are becoming.
Also find out if the presonus has any low latency monitoring facilities/cuemix via an onboard digital mixer (like Motu & RME units) That way you can do the bulk of your dry mix there with very low latency and then roll in some 100% wet verb etc from the DAW. You cant really change the settings in the cuemix between songs etc but basic compression, limiting and levels can happen there.
If I'm honest, I still prefer my spiltter approach where you just worry about the special FX and the FOH does the compression, & overall big picture because you want the FOH guy to fiddle with your levels not you. If you are going for exaggerated effects that are very integral to your performance and are going to be too many and change too frequently for a FOH guy (like with my act) then using a DAW is great (but still best done with a split for no latency woes). If you are after subtle and delicate compression and verb/delay then I think the FOH guy should handle it all and you should just focus on performing. Just focus your energy chatting with the FOH at sound check etc to let him know what you want. Most FOH guys love a band that wants them to play with certain FX in certain songs!!!
In this long post I've oversimplified a lot, & no doubt made generalisations & mistakes but these are some of the things you need to consider and educate yourself/test further. Having said all of the above, whether a delay is noticeable is subjective. If it isn't a noticeable problem for you then don't worry about it.
Also there are some great topics on things to change on your mac prior to performances so look them up. But basic things like turn off screen saver (obvious) and all other energy and screen saving option(turn off hard drives sleep etc) kill all auto update things like for mac update, microsoft office etc, no calendar pop ups, TURN OFF WIFI, kill bluetooth, don't launch the dashboard after restart etc. Also kill any start on launch programs (making ableton one though is a brilliant idea) as you need those precious seconds if something goes belly up & you need a restart)