To use CEM or not.

[Magnus Danielson]

From: Haible Juergen <Juergen.Haible at nbgm.siemens.de>
Subject: RE: To use CEM or not.
Date: Wed, 29 Sep 1999 12:20:11 +0200

> 	>The 2045 is an OTA based filter , like the CEM3320, it adds very
> little
> 	>colouration to the sound, so doesn't really have much character.
> 
> Just a tiny remark to add:
> 
> I don't know about the 2045. But in attempt to characterize the 3320 and
> it's clean-ness, "OTA based filter" might not be exactly right.
> There *are* OTA based filters that add harmonics to the sound, depending
> on the feedback resistor dividers. The CEM3320 (unlike the CA3080)
> uses linearized gain cells. If memory serves, these are even current-in /
> current-out multipliers, so the "transconductance" in the word "OTA"
> would not fit anyway.

Excuse me for interrupting here Mr. Haible, but this was interesting...

Do you say, on you Analogue DIY honor, that it is the non-linearized
multiplier cores that adds the harmonics that people tends to wish for
from curtain filters?

Just to make sure that we talk the same languague, the linearizing done to
multiplier cores (like the Gilbert cell) is the input diodes which act
arctanh functions when you pull the diffrential current through them?

That is, when you skip those you end up with a tanh distorsion.

Interesting thing!
I didn't knew that... or rather, I didn't think of it before.

> I don't know what a good word to characterize the type of 3320 filters.
> "Linearized VCA based filter" should cover the character of its sound,
> so maybe that would fit.

So, rougthly you are suggesting that we divide these filters into the
two groups of unlinearized and linearized OTAs.

It would be neat to see how well this correlates with peoples view of
various filters.

Cheers,
Magnus