27 Mar
2010
27 Mar
'10
8:18 a.m.
On Fri, Mar 26, 2010 at 11:04:00PM -0700, Ken Restivo wrote:
Throughout the decades, I've read various
engineers and
producers waxing poetically and rapturously about the Neve
console filters and how great they sound.
Same about Helios, Cadac, Midas, ...
I worked on big Neve consoles for years, and sure I do
like the filters. But see below.
Filters are filters, and there must be some way to
measure
their their frequency characteristics and emulate it in
software. Has anyone done that? In LADSPA on Linux?
We're close to myths and pseudoscience here.
Filters must not add distortion, noise etc. They only change
the frequency response. Now classic analog filters used in
audio consoles are (with maybe some rare exceptions) all of
the 1st and 2nd order minimum phase type. Which means that
(given the required controls and ranges) they all do the same
thing.
What makes them different is how they *feel* - the relation
between feeling the knobs move and the actual response.
For shelf filters 'gain' and 'frequency' will interact in
some way, and for parametrics 'gain' and 'bandwidth' will.
There are different ways to define these interactions, and
most engineers will prefer one or the other.
Unless Neve (or any of the others) are doing something really
exceptional (and then we'd know it), I'm pretty sure that if
you take e.g. the Neve circuit, replace the potentiometers
with some other brand that have either more or less friction,
change the type, size and color of the knobs, and modify the
layout a bit, nobody would recognise it as 'the Neve filter'.
Neve are still in business, you can buy the equalisers. If
you do it would be interesting to measure them, but I'd be
surprised if anything special would turn up.
Ciao,
--
FA
O tu, che porte, correndo si ?
E guerra e morte !