It's easier to to use the console emulation post recording rather than during.
First, render every recorded audio track with an input program. (There is a very useful Nebula batch render program & GUI available for free)
If your song uses vsti so there is no audio file, you can place the input program on the channel during mix (but you may quickly max out your processor or memory so maybe better to finalise sounds & arrangement then make them into audio or just leave the consoles off the instrument tracks).
Insert the Master Bus program on your main output.
Stert mixing & as you create groups for your guitars, vocals etc insert a bus group program.
That's pretty much it..Alex recommends trying to feed around -18db to the Nebula input, but I don't think you need to get to hung up on this at first...just try it with the recorded levels as they are
What about the Nebula buffer ......I think it's called dspbuffer in master page. You can reduce this for less latency but more cpu hit.....If you're running Nebulareverb at default the latency is massive but for eq, consoles, tapes etc I drop it to 1024 & have read others get away with 512.
You can make a copy of Nebula with this setting saved so you don't need to set it every time.