getuhgrip...
What you're describing sounds more like ping-ponging. I wasn't the one who mentioned sub-mixing, but I think I might be able to shed some light on a few things.
I mentioned destructive editing. Destructive editing means that you apply an effect to a track, and it physically changes the file on disk. There's no going back unless you make a backup copy of the original track. Forgetting about the obvious drawback, destructive editing can be effective in saving CPU resources, because the effect is no longer realtime (however, n-Track still leaves the effect in the chain, in bypass mode, so that you can still see your settings). I know it seems like you'd never want to do this....but consider something like a bass line. You apply a compression effect until you acheive the bass sound that you're after. Once you get it right, it is no longer necessary to have that compressor running in realtime, so applying the effect destructively might be a wise way to free up some resources. Of course later you might decide that you need to play with the compression on the bass a bit, so you restore the backup copy you so wisely made, and uncheck the "bypass" option in the compressor effect on the bass track....whamo, you're back in business. Of course this isn't optimal, but with limited CPU resources, it can become necessary.
Aux channels in n-Track are like effect send-returns, but you can have as many of them as you want. To use an aux channel, you insert an effect into an aux channel via the n-Track mixer, and then you send a track to it with the track's "send to aux" volume slider(s). Finally, you adjust the aux channel's input & output volume sliders to taste. The benfits are:
a) You have more control over how much of your track you send to an effect or GROUP OF effects.
b) You can send multiple tracks through the same effect. Say you have 3 cymbal tracks, and you want them all on the same reverb, an aux channel with a reverb is just the ticket. Instead of three reverb effects sucking CPU power, you now have just one. Instead of adjusting three reverb units for your cymbals, you now adjust only one.
c) You can send tracks to an effect dynamically either using automation or drawing send to aux envelopes. Have you heard that weird song I did a long time ago called "Me vs. Me?" At the end there is some bluesy guitar and as it progresses, it seems to slow down and get all distorted and weird sounding. Basically what I did was to create an aux channel with a pitch shift plugin set to something like -1 octave. Then I drew send to aux envelopes on the bluesy guitars such that as they progressed, the more you hear the pitch shift plugin, until eventually it's all you hear. (
http://www.slackmaster2000.com )
d) You can control whether the aux channels are pre or post fader & EQ. This really isn't a benefit of using aux channels, but it is something to note when you are using them.
n-Track also has track grouping capabilities. I think this must be what is *often* meant by sub-mixing. You can basically hook a bunch of tracks together such that you can make changes to all of them together. I've never used this feature, because I've been doing this all long via the aux channels.
What you described when you mentioned sub-mixing is more similar to ping-ponging. This is something that I do on occasion, but it's not a good way to mix. For instance, I'll often create a mixdown of just my drum tracks when I'm tracking. That way I can still hear the drums while I record, but I don't have to have 8 extra tracks cluttering things up. I will also do this on occasion when I start maxing out my system. If I've created a guitar sound using 4 guitar tracks, and I'm real happy with it, I'll create a mixdown of those guitar tracks and add it to the project. Then I'll mute the original 4 tracks. I turn off "read data from tracks when muted" at this point. I usually don't like mixing that many tracks though.
Slackmaster 2000
)....