My best results for SampleModelling Brass are by turning wet up to 100% for those instruments as the samples used in SM were recorded in an anechoic chamber. MIR really breathes life into my SM Brass in a way that I haven't yet experienced. That said other solutions come close but still feel a bit like a nice reverb veneer over a extremely dry recording.
For BWW I've loaded the close samples and so far I'm liking the results. BWW was recorded at Teldex for which a MIR roompack is available so theoretically if you had the room pack you could probably skip MIR for BWW(loading the mix or room samples) and but put the rest of your (relatively) dry instruments in MIR.
System resources are a limitation for me right now so I don't have a MIR instance for each SM instrument, LASS divisi section or individual BWW woodwinds. i'm basically using MIR on sections VLN1, Horns etc. and increasing the width of the instance to taste.
Also curious here about best practices for power panning to balance baked in positions in LASS. What I've done is essentially move the power panner for the dominant side (left for violins, for example) towards the center until the stereo output meters in VEPRO looks a centered as possible.
I'm pretty blown away by MIR right now. I can't believe I'd not tried it sooner.
-DJ