I can't hear any distinct difference between 48Khz and 96Khz. There is supposedly a difference between 44.1 and 48 and there is theory to back it up when you look at the higher frequencies of your hearing range.

I don't think there is really any need to sample or record at anything above 48Khz/24-bit, no matter what audiophiles believe (Remember that these are the same guys who pay $100 for audio cables...)

Yes, I do agree with you there - I was just making a point with the OP. You can eliminate all the audio 'artifacts' with just good quality converters and a decent sample rate.

I should have phrased it a bit clearer, however.

Um... why not just get good quality ADCs/DACs, record at 192Khz sample rate, 24-bit (maybe even 32-bit...) resolution uncompressed WAVE (or FLAC)?

The quality won't deteriorate over time like your analog.

As a musician and a sound engineer I shun everyone who relies on Auto Tune to make themselves sing in pitch!

If you don't have the vocal ability to sing in tune then you shouldn't be singing.

I think it's disgraceful that AutoTune be used for anything other than correcting minor blemishes and should never be used live. In fact, I usually take the stance that if I can't reproduce the effect live then I won't put it into the song. The audience has paid to see a live performance. Not your studio album played through speakers.

Unfortunately this is becoming all too common nowadays, using digital tools to touch everything up because you can. I weep for music's future.

This will require a computer to manage but it is possible, especially with Jack and probably Ardour too or a custom grown Jack application.

Since you need to output to multiple amplifiers, you'll need something like RME's HDSPe RayDat which has 36 outputs. This card should be supported by ALSA (and in turn, Jack) but you should always check thoroughly beforehand. Either way, you'll still probably have to spend some time configuring :)

Now, the RME cards don't provide Digital-to-Analog conversion on their cards, so you'll be needing a DA converter. For the ADAT, you can use a couple of Behringer ADA8000 (these have mic preamps too, but you won't be needing those). For S/PDIF and AES you can use two Behringer Ultramatch SRC2496. To prevent possible clock sync issues you might need to get the RME TCO (An expansion board with Wordclock sync outs) which will connect to all your converters.

Now you have a computer connected to around 36 outputs that can be connected to separate speakers.

You probably won't find software specific to what you're trying to do but you could write your own if you're up to it. I think what would work is if you have an application that shows the line your speakers run across. At each physical speaker, there is a virtual speaker in the software. You can then position points in the space around the speakers. The software could then calculate the distance between the two speakers to work out what volume each should output that sound as. You could make this work with MIDI and Ardour so Ardour handles all the audio processing but your application handles where the sound clip is positioned (ie: it's acting as a sort of front end).

Hope this helps ya!