Great test, thanks everyone involved.

I listened a few days ago, but I was less interested in how accurate the models matched the original mic samples, due the unknown variations between the instances of the mics modelled by each maker.

So the real question seems to be, do they capture the vibe of the original, and the presence/depth/smoothness etc, in a useful way? And then, as I'm considering it mainly as a vocal mic for myself (ie. not so much as a multi-purpose/voice studio asset), I asked myself which virtual mic's sound I preferred subjectively - ie. presumably the character of the raw mic, shining through each model. Or perhaps the 'sound' of each modelling technology.

No confirmation bias, because I like the look and price of the Slate best - but I actually ended up preferring the overall vibe of the Townsend! I wonder if the multi-pattern emulation has something to do with it, or maybe I just like the vibe of the raw mic? Purely subjective of course - but also sucks for me, as it's so much more expensive (and I was already looking at a used Slate to keep cost manageable).

I also like that you can get the Townsend software without dongles, so it's easier to pass mixes around and/or remix old projects even if you do sell on the mic. But the cost is well out of my reach, damnit …

And yes I agree that the original mics do still sound slightly more natural, more smooth. Which makes me wonder how digital algorithms can improve to nail that last bit of 'analogueness'. After all, even the modelling mics are real mics digitally captured. So theoretically (I'm a programmer so it's an educated guess) the modelling processing may improve in the future to take off that slight digital signature.