The Logic Off Topic list

> From: Kool Musick <koolmusick@...>

> Also ... does the modus operandi of Microphone Modeller truly have
> something to do with sampling? I thought the mathematical basis of all
> filtering, which as far as I know is basically what the MM does, was
> convolution? Or am I mistaken here? I thought, but I could be wrong, that
> the output of any digital filter, say a function y(k), was related to its
> input, say a different function x(k), through the relevant impulse
> response, another function, h(k)? Not that this is my area of expertise, or
> anything of course. I am happy to bow down to superior wisdom ... but I
> always thought that what related the two was convolution?

You're right, for the limited (but widely applicable case) of what's called
a Linear Time-Invariant function.  That means the mic being modelled
responds the same from moment to moment.  Most things work pretty much like
this.  An example of something that doesn't is a compressor, which therefore
can't be modelled as an impulse response.  But most other audio things
(filters, rooms, mics, amps, instruments, most acoustic things, reverbs) can
be.




> So ... what this 
> boils down to is how we come by the various functions, no? Well, that's a
> whole other issue in itself, but as far as I can see an input response
> FUNCTION is not a sample. One can criticize the MM for making some pretty
> unrealistic assumptions about the nature of its input functions and for the
> arbitrariness with which it selects the impulse responses, but I don't see
> how one criticize it for not being a sample when it isn't??!

Sonic Foundry's Acoustic Modeller does the same thing but allows the user to
measure anything and apply it as an impulse response (filter).  So in a way,
sampling is subtly involved, since you excite whatever your mic or whatever
is, and sample that, then feed it to the program which processes the sample
into an impulse response.