westocl wrote:
>> westocl wrote:
>>
>>   ...
>>
>>> Negative group delay cant be real. Im just using a digital computer to
>>> create a situation where the math works out, and thats enough to
> satisfy
>>> causality.
>> Real negative group delay can occur. Its inevitable association with 
>> appropriate phase delay is the constraint on it that keeps it causal.
>>
>> Jerry
>> -- 
>> Engineering is the art of making what you want from things you can get.
>> ???????????????????????????????????????????????????????????????????????
> 
> Hmph. If we know the time it takes for a digital computer to do its
> multiplications, it it fesable to ever say that it is real when we know
> the response comes from mathematics?

To say that what is real, the multiplication? The time? Please clarify.

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

>westocl wrote:
>
>   ...
>
>> Negative group delay cant be real. Im just using a digital computer to
>> create a situation where the math works out, and thats enough to
satisfy
>> causality.
>
>Real negative group delay can occur. Its inevitable association with 
>appropriate phase delay is the constraint on it that keeps it causal.
>
>Jerry
>-- 
>Engineering is the art of making what you want from things you can get.
>???????????????????????????????????????????????????????????????????????

Hmph. If we know the time it takes for a digital computer to do its
multiplications, it it fesable to ever say that it is real when we know
the response comes from mathematics?

westocl wrote:

   ...

> Negative group delay cant be real. Im just using a digital computer to
> create a situation where the math works out, and thats enough to satisfy
> causality.

Real negative group delay can occur. Its inevitable association with 
appropriate phase delay is the constraint on it that keeps it causal.

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

>On 10 Mrz., 11:05, Rune Allnor <all...@tele.ntnu.no> wrote:
>> On Mar 10, 9:40=A0am, Andor <andor.bari...@gmail.com> wrote:
>>
>> > Rune Allnor wrote:
>> > > A far more intersting question is whether a negative group
>> > > delay is non-causal.
>>
>> > You seem to have comletely missed the whole point of the exercise.
>>
>> At first, certainly. But I think I sorted it out eventually.
>
>My reply was a bit short, sorry about that. Well, the idea was to
>start with a circuit which was definitely and unrefutably causal and
>show how negative group delay leads to prediction in that case. So non-
>causality was ruled out from the start.
>
>> > Sorry about the confusion part :-).
>>
>> Well, thanks for initiating that excercise.
>
>I sure enjoyed the exercise :-).
>
>Regards,
>Andor


Dr Andor.
 
Your explanation is indeed 'phase cloning'. Rune asked me to disclose the
'trick'.  And thats what phase cloning is all about. Its a seeming way to
predict the future.  If one knows what two signals go into the black box,
we can predict the future exactly and compensate for it.  You talked of
oversampling., I erroneously called it 'resolution', however, it is indeed
oversampling.  Thats why im having problems getting the number of taps
down.

That being said, i believe, if we stomach the oversampling issue, we can
create filters that have an exact magnitude and phase response we
specify.

Negative group delay cant be real. Im just using a digital computer to
create a situation where the math works out, and thats enough to satisfy
causality.

On 10 Mrz., 11:05, Rune Allnor <all...@tele.ntnu.no> wrote:
> On Mar 10, 9:40&#4294967295;am, Andor <andor.bari...@gmail.com> wrote:
>
> > Rune Allnor wrote:
> > > A far more intersting question is whether a negative group
> > > delay is non-causal.
>
> > You seem to have comletely missed the whole point of the exercise.
>
> At first, certainly. But I think I sorted it out eventually.

My reply was a bit short, sorry about that. Well, the idea was to
start with a circuit which was definitely and unrefutably causal and
show how negative group delay leads to prediction in that case. So non-
causality was ruled out from the start.

> > Sorry about the confusion part :-).
>
> Well, thanks for initiating that excercise.

I sure enjoyed the exercise :-).

Regards,
Andor

On Mar 10, 9:40&#4294967295;am, Andor <andor.bari...@gmail.com> wrote:
> Rune Allnor wrote:
> > A far more intersting question is whether a negative group
> > delay is non-causal.
>
> You seem to have comletely missed the whole point of the exercise.

At first, certainly. But I think I sorted it out eventually.

> Sorry about the confusion part :-).

Well, thanks for initiating that excercise.

Rune

Rune Allnor wrote:

> A far more intersting question is whether a negative group
> delay is non-causal.

You seem to have comletely missed the whole point of the exercise.
Sorry about the confusion part :-).

Regards,
Andor

On Mar 8, 11:18 am, Andor <andor.bari...@gmail.com> wrote:
> From a recent discussion here:
>
> > >if i could generate some coefficents that had a 'negative' group delay for
> > >a period of time, would you think that 'phase cloning' was new and
> > >intersting??
>
> > A time machine would be pretty revolutionary, yes.
>
> > Negative group delay means that the output appears before the input
> > arrives.
>
> Fascinating concept, isn't it? I was curious enough to dig into the
> topic for a while and write up what I found out. You can read about it
> here:
>
> http://www.dsprelated.com/showarticle/54.php
>
> Regards,
> Andor

There is another way to do this. Take an integrator say K/s and put a
pure time-delay in it's feedback path exp(-sT).

Hence G(s) =K/s and H(s)=exp(-sT). At low frequencies we can easily
show that the closed-loop system G/(1+GH) is approx 1/exp(-sT) or
exp(sT)!!
Now in actual fact before people start jumping up and down you can
make this closed -loop system stable - but not stable enough to
predict the future. For a given bandwidth the overshoot you get
appears to be always greater than the delay (or advance) T.You can
introduce phase-lead to further stabilise but I didn't see any vortex
opening..

Hardy

On Mar 9, 7:20&#4294967295;pm, Andor <andor.bari...@gmail.com> wrote:
> Ron wrote:
> > A filter with negative group delay can produce the output
> > before the input if the input has been on and continues on
> > a highly predictable trajectory due to staying completely
> > within a certain constrained bandwidth.
>
> > No violation of causality becaues of the "if" clause.
>
> That is a great nutshell explanation. Thanks!

Well, in the abstract of the article you state that

    "This article aims at pouring oil in the fire and
     causing yet more confusion :-)."

and from that perspective, the above summary might be OK.

A far more intersting question is whether a negative group
delay is non-causal. In that case, the above is not quite
as helpful, simply because it has no practical relevance,
it is only philosophical speculation.

I have explained why in an earlier post:

http://groups.google.com/group/comp.dsp/msg/a10e64239d48d374

Rune

On Mon, 10 Mar 2008 08:04:53 +0800, Steve Underwood <steveu@dis.org>
wrote:

>Eric Jacobsen wrote:
>> On Sat, 8 Mar 2008 06:57:15 -0800 (PST), Mark <makolber@yahoo.com>
>> wrote:
>> 
>>> On Mar 7, 5:18 pm, Andor <andor.bari...@gmail.com> wrote:
>>>> From a recent discussion here:
>>>>
>>>>>> if i could generate some coefficents that had a 'negative' group delay for
>>>>>> a period of time, would you think that 'phase cloning' was new and
>>>>>> intersting??
>>>>> A time machine would be pretty revolutionary, yes.
>>>>> Negative group delay means that the output appears before the input
>>>>> arrives.
>>>> Fascinating concept, isn't it? I was curious enough to dig into the
>>>> topic for a while and write up what I found out. You can read about it
>>>> here:
>>>>
>>>> http://www.dsprelated.com/showarticle/54.php
>>>>
>>>> Regards,
>>>> Andor
>>> thank you
>>>
>>> Fig 6 was facinating...
>>>
>>> Could you reproduce Fig 6 except start at T=0?
>>>
>>> It would be interesting to see what happens at the very
>>> (unpredictable) start of the input signal.
>>>
>>> I'm going to build one of these and go to Wall Street  :-).
>>> Too bad stock numbers are not bandlimted  :-(
>>>
>>> Mark
>> 
>> This guy seems to think they are:
>> 
>> http://www.xyber9.com/Xyber9/Home.aspx
>> 
>> I like the "nominated for the Nobel Prize in Economics" marketing.   I
>> think the threshold for nomination isn't very tough to cross.   We
>> should all nominate each other for some big prize, or at least some
>> prize with a big sounding name.
>
>Even economists accept that if you laid them all end to end they 
>couldn't reach a conclusion. What kind of bar could possibly be low 
>enough for a nomination threshold?
>
>Steve

I don't know how low that bar would need to be, but if it has an
elevator and good beer on tap, I'd still find it acceptable. 


;)

Eric Jacobsen
Minister of Algorithms
Abineau Communications
http://www.ericjacobsen.org