200e

in case anyone is still interested in this rather tangential topic,
here's some more OL action that went down... the short answer is no, the
210e doesn't perform phase reversal...

> Re: 210e -- audio signal phase reversal?
  ///
> doesn't produce delay, which is the essential component of phasers,
etc.

you're thinking of flanging. phasing is different because it has
arbitrary notch and peak positions, and a small number of them, too.
it's typically done with chains of all pass filters each one inverting
the phase above some sperarate frequency so the center frequency of each
stage sets a notch or peak F when recombined with the original signal.
then it's sort of normal to sweep all the F's up and down together. I
always wondered why no one moved each one around as they saw fit, but
probably because it's hard to think of something nicer.. the stages are
cool looking and I never learned how to analyse them formally, but the
idea is that you feed the signal to both the inverting and non inverting
inputs on an opamp, but stick a cap before of the - and a resistor
before the +. the cap blocks DC so that all goes to the + and come out
uninverted. as you go up in frequency more and more signal goes through
the cap.. I think the CF is set by the RC time constant. then you VC the
C or the R. I forget the curve the phase shift follows as F goes up..
and as to why it shifts at all instead of dissapearing at the CF it has
some thing to do with the impedance (voltage-current lead-lag), but i
never got that far.. unsatisfactory I know yah but each APF is like a
kinky little delay line, with feedback and feedforward. in a digital
implementation this is quite literal; in analog i guess the caps and
opamps provide inverted and uninverted history of some sort, but the
whole process of course seems more opaque to me.

the point is that you have to have SOME propagation through time to do
filtering or phase transformation. it seemed that the original poster of
this now somewhat baroque thread was hoping to find a short kinky delay
somewhere in the 200e, that could be exploited for weird filtering
effects. alas, no...


AND another thing I don't understand this:

> phase delay = p = 2/freq * arctan( (1-a)/(1+a) *tan(freq/2) ),

doesn't that make the phase shift totally frequency dependent? or is
there something about it that keeps it flat at low enough F's? I tried
plotting it out, but couldn't really see what was what. can this be made
flatter and flatter by adding terms in some known way? I'd like a good
way of making fixed phase shifts. the only digital method I knew of was
a hilbert FIR (only 90 degrees) that was pretty long (but maybe bacause
it was really flat?). I bet lumpy stuff could be just as useful..
yeah there's some phase distortion w/ that algorithm. it is sensitive to
frequency (the lower the better) and the desired delay (the closer to a
multiple of 1 sample the better), but at low frequencies it's quite
flat. at the nyquist frequency everything converges to an integer
multiple of the sampling period. oversampling helps.

the thing to plot would be delay ("p") against frequency for a given
parameter value ("a").

i personally think the phase distortion is kinda nice when using this as
a chorus or karplus-strong component. it's not exactly easy to hear...

-eb


--- In 200e@yahoogroups.com, "ezra buchla" <ezra.buchla@...> wrote:
>
> yeah, quadrature introduces a 90deg phase shift in an envelope.
nothing to
> do w/ audio.
>
> btw, it's pretty easy to make a digital APF with almost-linear
arbitrary
> phase response. combine a 1sample feedforward and 1sample feedback
loop:
>
> y = a*x + x_last - a*y_last
>
> phase delay = p = 2/freq * arctan( (1-a)/(1+a) *tan(freq/2) ),
>
> so a = sin((1-p) * freq/2) / sin((1+p) * freq / 2))
> , so for freq small relative to sample rate, a = approx. (1-p)/(1+p)
>
> but all this doesn't have much to do with the 200e. experiments with
digital
> filter architectures are probably best carried out on a computer.
>
> -eb
>
>
> On Mon, Apr 28, 2008 at 12:11 PM, nicholas_kent zaum@... wrote:
>
> >   Doesn't make much sense to me (But I don't own any modules). The
227e
> > is a quad channel device. In general no one defaults audio as far as
I
> > know to 90 degree phase shifts to do quad. All you will wind up with
> > is more potential of cancellation in an acoustic space. And 90
> > degrees for that matter probably requires a pair of allpass filters
to
> > achieve. It's not something you can get without filters as far as I
> > know compared to just inverting the polatity for 180 degrees (though
> > filters could do that too).
> >
> > My guess is people are thinking along these lines because if you are
> > mixing quad audio you might very well want to hit your VCAs with 90
> > degree quadrature of a *control voltage* in order to do "circular"
pan
> > effects. Or perhaps someone just wants to identify the speakers in
> > numbers rather than ABCD if they represent a quad space
> >
> >
> > --- In 200e@yahoogroups.com <200e%40yahoogroups.com>, "mritenburg"
> > mritenburg@ wrote:
> > >
> > > I might be wrong, but I have always assumed that on the 227e,
positions
> > > A, B, C, and D are 90 degrees out of phase from one another. So C
is
> > > 180 degrees out of phase from A and D is 180 degrees out of phase
from
> > > B.
> > >
> > > Matt
> > >
> > >
> > >
> > > --- In 200e@yahoogroups.com <200e%40yahoogroups.com>, "Matt
Carpenter"
> > <matfhew.carpenfer@>
> > > wrote:
> > > >
> > > > Hi all,
> > > > I was just wondering if the 210e allows phase reversal of the
audio
> > > > signal or if phase reversal is possible via other means,
specifically
> > > > for the purposes of feedback/feedforward experimentation.
> > > > Thanks!
> > > > Matt
> > > >
> > >
> >
> >
> >
>
>
> [Non-text portions of this message have been removed]
>



[Non-text portions of this message have been removed]