Archival Electronics

[Martin Czech]

:::Don't get me wrong I love my analog stuff and do prefer it over anything
:::else, but in ten years time there won't be a need anymore.

I'm waiting for a reasonable amp/speaker model for guitar sound now for
over 10 years. Up to now, nothing comes even close to the real thing.

Sounds like a simple problem, but when you model the speaker (membrane
breakup, ie. you have to compute a finite element 2D or even 3D model),
radiation resistance, , the transformer, the tubes, it's getting really
nasty. I don't think that simplistic waveguide delay based methods will
lead anywhere with this kind of nonlinear operation, one would have to
solve the real equations. And even if the computing power would suffice,
the parameter extraction and model fitting and of course convergence
problems (ever dealt with spice?) would make it a real big project.

I think that building a real amp is still cheaper (in spite of the fact
that people want to play silent WITH distortion).

Nonlinearity is the fun and the name of the game, look how poor todays
discrete time systems perform in this area today. I think that dealing
with dynamic nonlinearitys would require rediculous high internal
sampling frequencys, 10x or even 100x higher then the external nyquist
requirement.

>From this point of view the future of analog gear still seems to be bright.

m.c.