Re: hi-quality pitch shifter ?

[Ricky W. Hunt]

"Mad Scientist Jr" > wrote in message
om...
> i have a whammy ii pedal that i sometimes use, but when i play chords
> through it, they sound all mechanical and warped. this pedal is 10 yrs
> old so i am sure since then something better has come out. i am
> looking for a studio quality pitch shifter for recording that i can
> plug a guitar or other instrument into and change pitch upto a whole
> step without any loss in quality. someone suggested lexicon. can
> someone recommend a specific model, or perhaps something analog?
> thanks.

There's a world of difference in shifting a single note and shifting a
chord. There's really nothing real-time (at least that sounds good) that
will shift a chord more than about a whole step. And even then it's pretty
noticeable (both for the listener and the player).

[Mike Rivers]

In article > writes:

> I do not think an analogue pitch shifter exists - but there are analogue
> frequency shifters so maybe you are thinking of them.

Probably not any more, but we had analog pitch changers (and speed
changers) before digital technology was advanced enough to use it for
that application. There was a mechanical tape-based system that used a
rotating head assembly. A nightmare, but it worked. It was mostly used
for changing speed and keeping a constant pitch rather than the other
way around, to shorten radio programs that were just a little too
long.

I suppose that beating the input source with a oscillator could be
considered an analog frequency shifter. They do that to record birds
and whales.



--
I'm really Mike Rivers )
However, until the spam goes away or Hell freezes over,
lots of IP addresses are blocked from this system. If
you e-mail me and it bounces, use your secret decoder ring
and reach me here: double-m-eleven-double-zero at yahoo

[Mad Scientist Jr]

> I do not think an analogue pitch shifter exists - but there are analogue
> frequency shifters so maybe you are thinking of them.

what is the difference between a pitch and a freq shifter?
were the original octave pedals analog?
they could shift pitch one octave right?

[Mad Scientist Jr]

that's my point ... 10 yrs later maybe we now have some hi tech 32-bit
super fast chip that can do it. no ?

> There's a world of difference in shifting a single note and shifting a
> chord. There's really nothing real-time (at least that sounds good) that
> will shift a chord more than about a whole step. And even then it's pretty
> noticeable (both for the listener and the player).

[Phil Allison]

"Mad Scientist Jr"

> > I do not think an analogue pitch shifter exists - but there are
analogue
> > frequency shifters so maybe you are thinking of them.
>
>
> what is the difference between a pitch and a freq shifter?


** The first multiplies all frequencies by a number ( greater or less than
one ) while the second adds or subtracts a number from all frequencies.


> were the original octave pedals analog?
> they could shift pitch one octave right?


** They could not shift pitch a sound.

Their function was to filter out and use the fundamental frequency of a
note to trigger a simple 2:1 divider that produced a square wave output one
octave lower. This phantom note ( after some filtering ) would start and
stop with the original and could be set to any desired level with a knob.



............ Phil

[Scott Dorsey]

Mad Scientist Jr > wrote:
>> I do not think an analogue pitch shifter exists - but there are analogue
>> frequency shifters so maybe you are thinking of them.
>
>what is the difference between a pitch and a freq shifter?

A pitch shifter shifts everything up by a factor, so that it remains in tune.
A frequency shifter adds a fixed shift to each frequency, so that it is no
longer in tune at all.

>were the original octave pedals analog?
>they could shift pitch one octave right?

Those things were actually just bridge rectifiers. Every time there was a
peak on either the positive or negative-going side of the line, it was turned
into a positive-going peak. This basically doubles the frequency, at the
expense of massive distortion.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Mad Scientist Jr]

okay that explains it a little more.. thanks

now what about a digital pitch shifter.. hasnt anyone made a hi-res
fast enough one where you could do chords in realtime without any
noticable quality loss? 64 bit?

[Michael R. Kesti]

Phil Allison wrote:

>** Frequency shifters that move the whole audio spectrum a few Hz up or
>down are used to suppress acoustic feedback in PA systems. I know about
>them because I designed one few years back that was later published as a
>project in Australia's leading electronics magazine.

I am very interested in that article. Please cite the name of that
magazine and the date of the issue that includes the article.

Thanks!

<snip>

--
================================================== ======================
Michael Kesti | "And like, one and one don't make
| two, one and one make one."
| - The Who, Bargain

[Mike Diack]

"Michael R. Kesti" > wrote in
:

> Phil Allison wrote:
>
>>** Frequency shifters that move the whole audio spectrum a few Hz up
>>or down are used to suppress acoustic feedback in PA systems. I know
>>about them because I designed one few years back that was later
>>published as a project in Australia's leading electronics magazine.
>
> I am very interested in that article. Please cite the name of that
> magazine and the date of the issue that includes the article.
>
> Thanks!
>
> <snip>
>

The idea is as old as the hills - There was one published in "Wireless
World"

"Frequency Shifter for "Howl" Suppression" by M Hartley Jones, July
1973 p317.

Used a pair of MC1495 4Q multipliers and a 5Hz quadrature oscillator.
I built it and it turned feedback squeals into Vogon noises.
(Before Mr Adams invented Vogons)
M

[Phil Allison]

"Mike Diack" <
> > Phil Allison wrote:
> >
> >>** Frequency shifters that move the whole audio spectrum a few Hz up
> >>or down are used to suppress acoustic feedback in PA systems. I know
> >>about them because I designed one few years back that was later
> >>published as a project in Australia's leading electronics magazine.
> >
> > I am very interested in that article. Please cite the name of that
> > magazine and the date of the issue that includes the article.
> >
>
> The idea is as old as the hills - There was one published in "Wireless
> World"
>
> "Frequency Shifter for "Howl" Suppression" by M Hartley Jones, July
> 1973 p317.
>
> Used a pair of MC1495 4Q multipliers and a 5Hz quadrature oscillator.
> I built it and it turned feedback squeals into Vogon noises.


** The first frequency shifters used RF techniques ( crystal oscillators,
modulators etc) and date from the early 1960s. Later designs in the
early 1970s used the "outphaser" method with multiplier ICs like the
C1495 - which finally became obsolete and unobtainable in the mid 1990s.
It was this fact that prompted me to design an audio shifter for EA
magazine based on the Analog Devices AD633 multiplier which proved to be
ideal for the job.

The M. Hartley Jones 5 Hz shifter you mention is rather primitive - more
like a concept demonstration than a useable product. Electronics Today
published an improved version of this design in 1977 also using MC1495s -
it proved to be very effective at increasing the gain before feedback
threshold in speech reinforcement systems operating in reverberant rooms. I
know, because I built quite a few and sold them to PA system operators.

If, as it seems, you built a primitive frequency shifter and tested it in
in a small, non reverberant room then the results are not even worth
mentioning. You have enlightened no-one by doing so.





............... Phil

[Mike Rivers]

In article > writes:

> The idea is as old as the hills - There was one published in "Wireless
> World"
>
> "Frequency Shifter for "Howl" Suppression" by M Hartley Jones, July
> 1973 p317.

I remember that. There was a commercial product built from the same or
similar circuit in the mid-70's. Taj Mahal (the singer) asked what was
wrong with the monitors.



--
I'm really Mike Rivers )
However, until the spam goes away or Hell freezes over,
lots of IP addresses are blocked from this system. If
you e-mail me and it bounces, use your secret decoder ring
and reach me here: double-m-eleven-double-zero at yahoo

[Peter B.]

Info on Freq shifters....


US Patent: 3,800,088:
Apparatus For Producing Special Audio Effects Utilizing Phase Shift
Techniques
Harald Bode
Filed: 28 August 1972
Issued: 26 March 1974


A High-Accuracy Frequency Shifter for Professional Audio Applications
Harald Bode, Robert A. Moog
Preprint 865, 42nd AES Convention, May 1972
AES Journal, Vol. 20, No. 6, p453, 1972


There is a big write up on frequency shifters in Bernie Hutchins
'Electronotes' newsletter.

[Pooh Bear]

Ben Bradley wrote:

> On 8 Aug 2004 10:53:16 -0400, (Mike Rivers) wrote:
>
> >
> >In article > writes:
> >
> >> I do not think an analogue pitch shifter exists - but there are analogue
> >> frequency shifters so maybe you are thinking of them.
> >
> >Probably not any more, but we had analog pitch changers (and speed
> >changers) before digital technology was advanced enough to use it for
> >that application. There was a mechanical tape-based system that used a
> >rotating head assembly. A nightmare, but it worked.
>
> You can frequency-modulate the clock of a bucket-brigade-device
> with a sawtooth to get (moderately crude) pitch shifting (the clocking
> and charge-shifting parts are digital, but it's still "mostly
> analog"). Isn't that (or something similar) how the earlier
> Harmonizers worked?

You can't use that method to frequency shift. It's an issue of samples in vs
samples out. With a fixed sampling rate in - each sample equates to a fixed time
period. If the out is clocked faster or slower then you will end up running out of
samples ( faster ) or introduce a time delay if slower ( same as slowing a tape
down for example ).

The kind of pitch shifting you refer to requires a DSP algorithm that either throws
away some samples or introduces new ones.


Graham

[Bob Cain]

Pooh Bear wrote:


> The kind of pitch shifting you refer to requires a DSP algorithm that either throws
> away some samples or introduces new ones.

Or does things _way_ more complicated than that. :-)

One can Google on "pitch scaling" to get a sample of the
research work. The really good stuff, though, doesn't get
published. It ends up in products.

The most easily understood is based on what's called a phase
vocoder but it doesn't sound all that good on anything but
moderate quality voice where the artifacts don't
signifigantly hinder the application.


Bob
--

"Things should be described as simply as possible, but no
simpler."

A. Einstein

[Phil Allison]

<Kurt Riemann>
> >
> >> Phil, you are totally ****ing irritating and unhelpful..


** Kurt - you are a serious public menace.

I do hope the police have a watch on you.





............ Phil

[Phil Allison]

<Kurt Riemann>

** This dude is out there - he may be in YOUR neighbourhood.

Be afraid.

Kurt is a serious public menace.

I do hope the police have a watch on him.





............. Phil

[Mike Rivers]

In article > writes:

> You can frequency-modulate the clock of a bucket-brigade-device
> with a sawtooth to get (moderately crude) pitch shifting (the clocking
> and charge-shifting parts are digital, but it's still "mostly
> analog"). Isn't that (or something similar) how the earlier
> Harmonizers worked?

That's how the Marshall Time Modulator worked. You can't get a
constant pitch shift that way, only a pitch modulation.

Harmonizers were cut-and-splice.


--
I'm really Mike Rivers )
However, until the spam goes away or Hell freezes over,
lots of IP addresses are blocked from this system. If
you e-mail me and it bounces, use your secret decoder ring
and reach me here: double-m-eleven-double-zero at yahoo

[Mike Rivers]

In article > writes:

> It's definitely not acceptable for any musical applications, but when you
> think about it, an awful lot of sound systems are never used for music
> anyway.

That was usually discovered after the sound company set it up in the
PA system and the band was on stage.


--
I'm really Mike Rivers )
However, until the spam goes away or Hell freezes over,
lots of IP addresses are blocked from this system. If
you e-mail me and it bounces, use your secret decoder ring
and reach me here: double-m-eleven-double-zero at yahoo

[Mike Rivers]

In article > writes:

> There is a big write up on frequency shifters in Bernie Hutchins
> 'Electronotes' newsletter.

Hey! I've got some of those. I'll have to look it up and see if he
mentions Phil.

--
I'm really Mike Rivers )
However, until the spam goes away or Hell freezes over,
lots of IP addresses are blocked from this system. If
you e-mail me and it bounces, use your secret decoder ring
and reach me here: double-m-eleven-double-zero at yahoo

[Peter B.]

"Phil Allison" > wrote in message >...


>
> ** That would be the same Harald Bode who's name is associated with so
> many plots ????
>

>
> ............. Phil


I had wondered the same thing in the past and never got around to
looking it up. I think I may have found some answers today.

Here is some info on Hendrik Bode of 'bode plot' fame:

http://www.geocities.com/neveyaakov/electro_science/bode.html


There is a nice quote from Hendrik Bode on this website and some
interesting history on feedback...

http://www.geocities.com/edug2406/apuntes_hp.htm

There is a nice quote near the top of the page.

"The engineer who embarks on the design of a feedback amplifier must
be a creature of mixed emotions."
—Hendrik Bode





Further reseach found that Harald Bode also worked at Bell Labs, here
is a good link to some information on him:

http://www.obsolete.com/120_years/machines/melochord/

[Kurt Albershardt]

Scott Dorsey wrote:
> Bob Cain > wrote:
>
>
>> Feedback supression is still one hell of an application
>> awaiting a really good, real time, low latency frequency
>> scaling solution.
>
>
> What's wrong with the Urei Little Dipper? You hear it ring, you turn
> the knob, it stops ringing. Works in realtime, and the latency is as
> low as the operator, which can be pretty low.

I have to say I really miss my Ashly SC-68.

I'd be positively thrilled to see one of the small console manufacturers add a couple of notch filters to their master section.

[Peter B.]

"Phil Allison" > wrote in message >...

>
>
>
> ** That would be the same Harald Bode who's name is associated with so
> many plots ????
>
>
> ............. Phil



Some time ago I wondered the same thing but never got around to
finding out. Here is what I dug up on the internet today:




Hendrik Bode

'Bode Plot'


http://www.geocities.com/neveyaakov/electro_science/bode.html


Interesting reading on feedback:

http://www.geocities.com/edug2406/apuntes_hp.htm

Quote on feedback:

The engineer who embarks on the design of a feedback amplifier must be
a creature of mixed emotions.
—Hendrik Bode



Harald Bode

'Frequency Shifter'


http://www.obsolete.com/120_years/machines/melochord/




Both happened to work at Bell Labs.



Peter

[Mad Scientist Jr]

What I was wondering is, if an octave pedal cust the frequency in
half, which equals one octave, could one be altered to cut or boost
the frequency by whatever rate equals half a step? Or does half a step
vary depending on what note it is (eg there is no half step between B
and C notes, I am rusty on my music theory)?

Also, I am wondering if it is possible to fix the fact that a digital
pitch shifter craps out on a chord, by first breaking the chord into
its component parts/frequencies and then operating on these
individually (one pitch shifter per frequency or part), then
recombining them into the final pitch shifted chord?

[ScotFraser]

<< What I was wondering is, if an octave pedal cust the frequency in
half, which equals one octave, could one be altered to cut or boost
the frequency by whatever rate equals half a step?>>

Halving the frequency electronically is extremely simple. Any other division is
extremely complex.

<<Or does half a step
vary depending on what note it is >>

No, a half step is a ratio relationship of 1.0594 to 1. Doesn't matter where in
the frequency spectrum or scale you start, the ratio is always the same.

<<(eg there is no half step between B
and C notes, I am rusty on my music theory)?>>

B to C IS a half step, so there is no half step between them.

<<Also, I am wondering if it is possible to fix the fact that a digital
pitch shifter craps out on a chord, by first breaking the chord into
its component parts/frequencies and then operating on these
individually (one pitch shifter per frequency or part), then
recombining them into the final pitch shifted chord? >>

There's no algorithm currently available, other than a well trained human
brain, which can separate the notes out of a chord. Suppose your chord consists
of every note on a piano played simultaneously. I think you can imagine that no
device now or in the future could ever discern & separate all 88 notes out of
this chord.



Scott Fraser

[Mike Rivers]

In article > writes:

> What I was wondering is, if an octave pedal cust the frequency in
> half, which equals one octave, could one be altered to cut or boost
> the frequency by whatever rate equals half a step?

No. An "octave pedal" works by dividing the input frequency in half
using a flip-flp. This is a circut which changes state when a specific
voltage is reached going in a spicific direction (either
negative-to-positive or positive-to-negative).

Let's say a positive-going input level of +1V causes it to change
state, either from off to on or from on to off. Since the input signal
(let's say it's a sine wave with an amplitide of 3V peak-to-peak,
swinging from +1.5V to =1.5V) reaches +1V while the signal is going
positive only once per cycle, the flip-flop will turn on for the first
cycle, then off for the second cycle, then on again for the third
cycle, etc. This means that the frequency of the flip-flop is exaclty
half that of the input signal. There's no way you can change this
relationship to be some other multiple of the input signal using the
"octave dividing" technique.

> does half a step
> vary depending on what note it is (eg there is no half step between B
> and C notes, I am rusty on my music theory)?

With an equal-tempered scale, the ratio between to half steps is the
twelfth root of two, about 1.06 as I recall. You can't do that with a
flip-flop.

> Also, I am wondering if it is possible to fix the fact that a digital
> pitch shifter craps out on a chord, by first breaking the chord into
> its component parts/frequencies and then operating on these
> individually (one pitch shifter per frequency or part), then
> recombining them into the final pitch shifted chord?

You do the first part and the second part will be easy. <g>

--
I'm really Mike Rivers )
However, until the spam goes away or Hell freezes over,
lots of IP addresses are blocked from this system. If
you e-mail me and it bounces, use your secret decoder ring
and reach me here: double-m-eleven-double-zero at yahoo

[Mad Scientist Jr]

thanks for your answer... i guess i meant, is it possible to
*electronically* separate a sound into its component waves - not
necessarily by notes as we know them, but maybe by EQ range or divided
up by the spectrum? i am not a physicist but i know this is all based
on waves and it can get very complex, esp in realtime. the idea is if
a sound could be split up somehow so then each of the parts could be
worked on separately...

> There's no algorithm currently available, other than a well trained human
> brain, which can separate the notes out of a chord. Suppose your chord consists
> of every note on a piano played simultaneously. I think you can imagine that no
> device now or in the future could ever discern & separate all 88 notes out of
> this chord.
>
>
>
> Scott Fraser

[ScotFraser]

<< i guess i meant, is it possible to
*electronically* separate a sound into its component waves - not
necessarily by notes as we know them, but maybe by EQ range or divided
up by the spectrum? i am not a physicist but i know this is all based
on waves and it can get very complex, esp in realtime. the idea is if
a sound could be split up somehow so then each of the parts could be
worked on separately...>>

Yes, although this isn't the same as separating notes out of a chord. An easy
implementation of this is a crossover, used to divide frequencies into several
bands to feed woofers, mids, & tweeters individually. A complex version of this
is multiband noise reduction software that often splits the spectrum into 1,000
or more bands & applies expanders to each, according to the amplitude of energy
within each band.



Scott Fraser

[Mad Scientist Jr]

> Yes, although this isn't the same as separating notes out of a chord.

but it wouldn't have to be or would it? if you could get a *realtime*
spectrum or freq splitter to divide a chord up into small enough
pieces that could be
handled by a pitch shifter without noticable quality loss (since pitch
shifters work ok on single notes, maybe a single frequency or a narrow
freq range could be akin to a single note), then you could just run a
bunch of pitch shifters in parallel on all the pieces, then mix the
shifted pieces back together for the final sound. ?

[Mad Scientist Jr]

> Yes, although this isn't the same as separating notes out of a chord.

but it wouldn't have to be or would it? if you could get a *realtime*
spectrum or freq splitter to divide a chord up into small enough
pieces that could be
handled by a pitch shifter without noticable quality loss (since pitch
shifters work ok on single notes, maybe a single frequency or a narrow
freq range could be akin to a single note), then you could just run a
bunch of pitch shifters in parallel on all the pieces, then mix the
shifted pieces back together for the final sound. ?

[Mad Scientist Jr]

> > Also, I am wondering if it is possible to fix the fact that a digital
> > pitch shifter craps out on a chord, by first breaking the chord into
> > its component parts/frequencies and then operating on these
> > individually (one pitch shifter per frequency or part), then
> > recombining them into the final pitch shifted chord?
> You do the first part and the second part will be easy. <g>

see above, someone mentioned a spectrum or frequency splitter....

[Mad Scientist Jr]

> > Also, I am wondering if it is possible to fix the fact that a digital
> > pitch shifter craps out on a chord, by first breaking the chord into
> > its component parts/frequencies and then operating on these
> > individually (one pitch shifter per frequency or part), then
> > recombining them into the final pitch shifted chord?
> You do the first part and the second part will be easy. <g>

see above, someone mentioned a spectrum or frequency splitter....

[Bob Cain]

Mad Scientist Jr wrote:

> thanks for your answer... i guess i meant, is it possible to
> *electronically* separate a sound into its component waves - not
> necessarily by notes as we know them, but maybe by EQ range or divided
> up by the spectrum? i am not a physicist but i know this is all based
> on waves and it can get very complex, esp in realtime. the idea is if
> a sound could be split up somehow so then each of the parts could be
> worked on separately...

No because it isn't really made of component waves. It's
just what it is. We can pretend that it is and proceed from
their in obvious ways but the results show the validity of
the assumption. They aren't good.


Bob
--

"Things should be described as simply as possible, but no
simpler."

A. Einstein

[Bob Cain]

Mad Scientist Jr wrote:

> thanks for your answer... i guess i meant, is it possible to
> *electronically* separate a sound into its component waves - not
> necessarily by notes as we know them, but maybe by EQ range or divided
> up by the spectrum? i am not a physicist but i know this is all based
> on waves and it can get very complex, esp in realtime. the idea is if
> a sound could be split up somehow so then each of the parts could be
> worked on separately...

No because it isn't really made of component waves. It's
just what it is. We can pretend that it is and proceed from
their in obvious ways but the results show the validity of
the assumption. They aren't good.


Bob
--

"Things should be described as simply as possible, but no
simpler."

A. Einstein

[Mad Scientist Jr]

again, i am not an engineer and am going on layman's terms here... so
dividing a sound up with a spectrum splitter isn't really dividing it
correctly ? i don't know much about chord theory, but isn't it
something like a chord is made up of a root note and harmonics and
overtones. can these be somehow divided?

> No because it isn't really made of component waves. It's
> just what it is. We can pretend that it is and proceed from
> their in obvious ways but the results show the validity of
> the assumption. They aren't good.

[Mad Scientist Jr]

again, i am not an engineer and am going on layman's terms here... so
dividing a sound up with a spectrum splitter isn't really dividing it
correctly ? i don't know much about chord theory, but isn't it
something like a chord is made up of a root note and harmonics and
overtones. can these be somehow divided?

> No because it isn't really made of component waves. It's
> just what it is. We can pretend that it is and proceed from
> their in obvious ways but the results show the validity of
> the assumption. They aren't good.

[Pooh Bear]

Mad Scientist Jr wrote:

> again, i am not an engineer and am going on layman's terms here... so
> dividing a sound up with a spectrum splitter isn't really dividing it
> correctly ? i don't know much about chord theory, but isn't it
> something like a chord is made up of a root note and harmonics and
> overtones.

Errr .... no.


Graham

[Pooh Bear]

Mad Scientist Jr wrote:

> again, i am not an engineer and am going on layman's terms here... so
> dividing a sound up with a spectrum splitter isn't really dividing it
> correctly ? i don't know much about chord theory, but isn't it
> something like a chord is made up of a root note and harmonics and
> overtones.

Errr .... no.


Graham

[Mad Scientist Jr]

Maybe you could expand on this? Smart aleck posts generally aren't that helpful.

[Bob Cain]

Mad Scientist Jr wrote:

> Maybe you could expand on this? Smart aleck posts generally aren't that helpful.

While the signal can certainly be decomposed into a Fourier
space equivalent, the problem of assigning meaning to the
result is called the "blind separation" problem which you
can google on. It's generally not possible and there is
still a lot of research ongoing to find methods that are
even somewhat effective.


Bob
--

"Things should be described as simply as possible, but no
simpler."

A. Einstein

[Ricky W. Hunt]

"Bob Cain" > wrote in message
...
>
>
> Mad Scientist Jr wrote:
>
> > Maybe you could expand on this? Smart aleck posts generally aren't that
helpful.
>
> While the signal can certainly be decomposed into a Fourier
> space equivalent, the problem of assigning meaning to the
> result is called the "blind separation" problem which you
> can google on. It's generally not possible and there is
> still a lot of research ongoing to find methods that are
> even somewhat effective.

Right. It's not even a matter of computer power at this point.

[Mad Scientist Jr]

Well that is at least some information to go on.
Thanks for the reply...

[Mad Scientist Jr]

I take it that it's more an AI type problem, since you are dealing
with waves that are occurring simultaneously, to "listen" and separate
a chord into notes or some other meaningful parts? Is that even close?

Has anyone gotten any results in this area?

> > While the signal can certainly be decomposed into a Fourier
> > space equivalent, the problem of assigning meaning to the
> > result is called the "blind separation" problem which you
> > can google on. It's generally not possible and there is
> > still a lot of research ongoing to find methods that are
> > even somewhat effective.
> Right. It's not even a matter of computer power at this point.

[Bob Cain]

Mad Scientist Jr wrote:

> I take it that it's more an AI type problem, since you are dealing
> with waves that are occurring simultaneously, to "listen" and separate
> a chord into notes or some other meaningful parts? Is that even close?

Spot on.

>
> Has anyone gotten any results in this area?

I dunno exactly where it stands right now. Google on "blind
separation."


Bob
--

"Things should be described as simply as possible, but no
simpler."

A. Einstein