zeroes in unit circle, minimum phase

[anyfoo]

Hello,

When I look at the "System Response" of a measurement I took inside my car, it consists almost entirely of zeroes inside the unit circle. Doesn't this mean that the system is, by definition, "mostly" minimum phase[1], and so reversible? (Cancel each zero with a pole inside the unit circle, the system remains stable and the overall system should be zero phase.)

Could REW help me make such a filter?

[1] "Mostly" because I don't expect the derived system to be completely accurate, and indeed I see some narrow non-minimum phase regions in the group delay, but I figured it's okay to ignore them for now.

 

[John Mulcahy]

Any invertible response can be inverted (e.g. the trace arithmetic 1/A operation) to create an inverse filter, but the filter is likely to have high gains wherever the response has attenuation and to only be effective at the exact point of measurement, making it useless from a practical perspective.

 

[anyfoo]

Ah, yeah, my bad. It's a sampled spectrum, dependent on how long the FFT is. I guess artisanal and hand-crafted is still the way to go. 8)