What you're trying to accomplish is a cool idea but this will not give you a flat response speaker cabinet, if I understand this correctly inverting the response would just give you noise wouldn't it? It would not produce the effect of a flat response cabinet. I don't know much about this stuff but that was my understanding.
You are correct, it would not be a perfectly 'flat' response... but it would be a baseline to add a replacement response.
And no, it would not simply add noise.
You can think of Impulse Responses in their simplest form as just steep EQ curves. (I know that's not EXACTLY what they are, but you get the idea)
So if your real cabinet is a Marshall 1960a but you want to have a EQ curve of a 4X10 Bassman, you would first have to take the 'Marshall 1960a' response back to some sort of baseline (i.e. as flat of a response as possible) and then add the 4X10 Bassman to replace it.
Of course, it would not sound EXACTLY like a 4X10 bassman because it is an open back cabinet and the Marshall 1960a is a closed back cabinet, but you could get the EQ curve - which is what I'm wanting to do (among other things).
Convolution (which is what Impulse Responses are) is the way of the future.
And I'm so glad that Cliff took this approach in doing the AXE-FX II. With convolution, you can get some incredibly great sounding emulations.
Of course, a single impulse response taken at one 'level' will not work for all 'levels' because as you crank up the gain to the speaker cabinet, there is some non-linearity going on.
But as I understand it, the new AXE-FX II has multi-sampled IR's (which is WAY COOL!)
If Cliff was able to do this without 'zipper' noise (i.e. switching from one IR to another IR dynamically) this is a major accomplishment.
And if this is true, then I don't see anything coming close to what this unit does anytime soon (not counting plugins on fast computers - but who wants to gig with that?)
All the best to you!
Larry