Richard Owlett wrote:
(snip)

> So, i'm frustrated
> And even if the group "dives me up a wall" I do appreciate the attempts 
> to answer the questions that many think I should be asking. Even if they 
> are not the questions I am asking.

Sometimes people don't ask the right question, so the only possibility
is to answer the question that they seem to be asking.

This happens a lot in comp.lang.* newsgroups, where someone posts
about a program that doesn't work (possibly in the language of
the newsgroup) but shows no code from the program.  One can then
guess what might have caused the given problem.

Sometimes a post reminds one of a question that should have
been asked but wasn't actually asked.

-- glen

dbd wrote:
> On Apr 16, 10:45 am, Richard Owlett <rowl...@atlascomm.net> wrote:
> 
> When you are in a room full of people with hammers, why are you
> surprised when they respond to your questions with nails?
> 
> Perhaps you have unrealistic expectations that have led you to an
> unproductive approach.
> 
> 
>>...
>>
>>If I pare it down to a bare bones math question I'm assaulted with
>>request for the application details.
> 
> 
> Yes, what's wrong with that? This is an engineering sort of forum and
> that is how the engineering mind proceeds when you say you have a
> problem. That's how we evaluate to arrive at solutions and how we know
> when we are done.
> 
> You began this discussion with a request for a 'better interpolation'.
> The 'better' can only be understood in some context, but you complain
> when someone asks for that context! That's either fuzzy or trollish.
> 
> 
>>If I give "details", either it generates "rabbit trail" subthreads ...
> 
> 
> Yes, what's wrong with that? This is an engineering sort of forum.
> 
> If the subthreads aren't doing what you want, post. Say so and say
> why. If you don't keep playing, why shouldn't we continue to discuss
> points that interest us? Of course, those subthreads will continue
> whatever you do, but if you don't keep giving feedback towards what
> you want to hear about, we can't respond with anything about the
> interests you have chosen to leave vague.
> 
> 
>>...
>>
>>So, i'm frustrated
>>And even if the group "dives me up a wall" I do appreciate the attempts
>>to answer the questions that many think I should be asking. Even if they
>>are not the questions I am asking.
>>...
> 
> 
> You are getting better than you are asking for. If you want something
> different, change what you ask for and how you ask for it or who you
> ask. If all you want is a warm fuzzy feeling, don't feed a bunch of
> technical people any suggestion of a "problem" and expect them to
> respond by trying to hold your hand.
> 
> Dale B. Dalrymple
> 

*LOL*

I ask for basic math
I get advanced engineering

OK so we are of different mindsets

b-b-b-b-b-b ut u got info i need/want ;/

On Apr 16, 10:45 am, Richard Owlett <rowl...@atlascomm.net> wrote:

When you are in a room full of people with hammers, why are you
surprised when they respond to your questions with nails?

Perhaps you have unrealistic expectations that have led you to an
unproductive approach.

> ...
>
> If I pare it down to a bare bones math question I'm assaulted with
> request for the application details.

Yes, what's wrong with that? This is an engineering sort of forum and
that is how the engineering mind proceeds when you say you have a
problem. That's how we evaluate to arrive at solutions and how we know
when we are done.

You began this discussion with a request for a 'better interpolation'.
The 'better' can only be understood in some context, but you complain
when someone asks for that context! That's either fuzzy or trollish.

>
> If I give "details", either it generates "rabbit trail" subthreads ...

Yes, what's wrong with that? This is an engineering sort of forum.

If the subthreads aren't doing what you want, post. Say so and say
why. If you don't keep playing, why shouldn't we continue to discuss
points that interest us? Of course, those subthreads will continue
whatever you do, but if you don't keep giving feedback towards what
you want to hear about, we can't respond with anything about the
interests you have chosen to leave vague.

> ...
>
> So, i'm frustrated
> And even if the group "dives me up a wall" I do appreciate the attempts
> to answer the questions that many think I should be asking. Even if they
> are not the questions I am asking.
> ...

You are getting better than you are asking for. If you want something
different, change what you ask for and how you ask for it or who you
ask. If all you want is a warm fuzzy feeling, don't feed a bunch of
technical people any suggestion of a "problem" and expect them to
respond by trying to hold your hand.

Dale B. Dalrymple

Jerry Avins wrote:
> Richard Owlett wrote:
> 
>   ...
> 
>> ARRRGHH ;)
>>
>> This group thinks my questions are more subtle than they are.
>> Although I haven't been in a classroom for 40 years, I'd guess my 
>> level of comprehension is that of a college sophomore or junior.
> 
> 
> Maybe you could word your questions more cautiously, so as not to evoke 
> images od subtlty in the reader. (More easily said than done)
> 
>   ...
> 
>> P.s. Re "You keep fuzzy logic alive at comp.dsp." I've stayed out of 
>> that thread cause the closest I've come to control systems is 
>> maintaining L&N chart recorders with chopper stabilized vacuum tube 
>> amplifiers in the loop. (so i show my age)
> 
> 
> Dale was gently implying that your thinking is fuzzy. Sometimes it is. 
> Sometimes mine is too. I don't mind getting called on it now and then.
> 
> Jerry

RAFL (similar to ROFL, physical therapist objects to reports of me being 
"on floor", so wheelchair has me "Rolling ABOVE Floor Laughing")

To take the second point first - my thinking isn't "fuzzy".
It's frequently totally disoriented and/or lost ;)

More seriously, this is primarily a group of engineers. And, although I 
completed 3 years towards a BSEE, my aptitude/interest profile just does 
not match a typical/average/"statistically normal" engineer.

This group is not just "application oriented" but " _specific_ 
application oriented" by mindset. That does *NOT* mean an individual 
deals with ONLY "speech recognition" OR "control systems". He may do 
both. !!!! *BUT* !!! he is most comfortable with a specific case.
[I hope that came out somewhere near right.]

As to the first, I can't win.

If I pare it down to a bare bones math question I'm assaulted with 
request for the application details.

If I give "details", either it generates "rabbit trail" subthreads or to 
paraphrase somebody's tag line, "Universal advice 'DON'T DO DAT' "

For example (snicker snicker LOL)

About two years ago I basically wanted to ask how to determine the 
average time of a process measured by a clock having femto-second 
accuracy *BUT* quarter HOUR resolution and report in a form grasped by 
liberal ARTS majors. Having been recently severely *CHASTISED* for lack 
of "detail", 
http://groups.google.com/groups/search?q=%22Time+cards+and+sampling+theorems%22
was the result. Not only did I not get an answer, but in a related 
thread or a personal email I was told "don't do that you will jeporadize 
your job"


So, i'm frustrated
And even if the group "dives me up a wall" I do appreciate the attempts 
to answer the questions that many think I should be asking. Even if they 
are not the questions I am asking.



http://groups.google.com/groups/search?q=%22Time+cards+and+sampling+theorems%22
Time cards and sampling theorems

Richard Owlett wrote:

   ...

> ARRRGHH ;)
> 
> This group thinks my questions are more subtle than they are.
> Although I haven't been in a classroom for 40 years, I'd guess my level 
> of comprehension is that of a college sophomore or junior.

Maybe you could word your questions more cautiously, so as not to evoke 
images od subtlty in the reader. (More easily said than done)

   ...

> P.s. Re "You keep fuzzy logic alive at comp.dsp." I've stayed out of 
> that thread cause the closest I've come to control systems is 
> maintaining L&N chart recorders with chopper stabilized vacuum tube 
> amplifiers in the loop. (so i show my age)

Dale was gently implying that your thinking is fuzzy. Sometimes it is. 
Sometimes mine is too. I don't mind getting called on it now and then.

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

dbd wrote:
> [snip]
>>
>>I asked here with a slightly different purpose in mind.
>>
>>Given a limited amount of data and some understanding of what the
>>underlying process "looks like":
>>1. How much/what can we extrapolate/interpolate/infer/... ?
>>2. How do you go about solving the problem?
> 
> 
> 
>>I've learned to be very careful in phrasing my questions on this group.
>>If I give too many details of the _specific_ instance that triggered the
>>question, I'll get a fish rather than learning to fish ;)
> 
> 
> You don't leave out just specifics, you leave out any clear statement
> of purpose or goal. You keep fuzzy logic alive at comp.dsp. We've
> grown accustomed to your inability to define a problem  when you hand
> wave about your "problem". Fortunately at comp.dsp we are so used to
> amusing ourselves talking about fishing that it doesn't matter that
> you never know whether to fish or cut bait:,)
> 
> Dale B. Dalrymple

ARRRGHH ;)

This group thinks my questions are more subtle than they are.
Although I haven't been in a classroom for 40 years, I'd guess my level 
of comprehension is that of a college sophomore or junior.

For someone with engineering interests, my math aptitude is below average.

I work on a problem of interest, and it triggers a *GENERAL* question.
There are usually multiple ways out of whatever bind I'm in.

But the question remains.

P.s. Re "You keep fuzzy logic alive at comp.dsp." I've stayed out of 
that thread cause the closest I've come to control systems is 
maintaining L&N chart recorders with chopper stabilized vacuum tube 
amplifiers in the loop. (so i show my age)

On Apr 15, 3:35 am, Richard Owlett <rowl...@atlascomm.net> wrote:
> glen herrmannsfeldt wrote:
> > Richard Owlett wrote:
>
> > (snip)
>
> >> The system has 4 independent signal sources. Two of which are
> >> comparable in magnitude of their fundamentals. Their frequencies
> >> differ by ~.2% .
> >> The other are 30-300+ times slower and much weaker. But ALL 4
> >> fundamental frequencies are known very accurately.
>
> > How about a least-squares fit to the four sines of known frequency
> > but unknown phase and amplitude?
>
> > Though it might be that a linear least-squares fit to a low order
> > polynomial would work, too.  Or maybe two sines plus a polynomial,
> > where the sines do the higher frequencies and the polynomial the
> > low frequencies.
>
> > -- glen
>
> I asked a different form of the question on sci.geo.satellite-nav. I
> eventually got a reply that one person uses Lagrange or Hermite
> interpolators depending on whether or not he has velocity data as well
> as position data. I have had a chance to look only at couple of
> Wikipedia articles. I've more reading to do. Any recommended sites - esp
> with worked examples?
>
> I asked here with a slightly different purpose in mind.
>
> Given a limited amount of data and some understanding of what the
> underlying process "looks like":
> 1. How much/what can we extrapolate/interpolate/infer/... ?
> 2. How do you go about solving the problem?


> I've learned to be very careful in phrasing my questions on this group.
> If I give too many details of the _specific_ instance that triggered the
> question, I'll get a fish rather than learning to fish ;)

You don't leave out just specifics, you leave out any clear statement
of purpose or goal. You keep fuzzy logic alive at comp.dsp. We've
grown accustomed to your inability to define a problem  when you hand
wave about your "problem". Fortunately at comp.dsp we are so used to
amusing ourselves talking about fishing that it doesn't matter that
you never know whether to fish or cut bait:,)

Dale B. Dalrymple

glen herrmannsfeldt wrote:
> Richard Owlett wrote:
> 
> (snip)
> 
>> The system has 4 independent signal sources. Two of which are 
>> comparable in magnitude of their fundamentals. Their frequencies 
>> differ by ~.2% .
>> The other are 30-300+ times slower and much weaker. But ALL 4 
>> fundamental frequencies are known very accurately.
> 
> 
> How about a least-squares fit to the four sines of known frequency
> but unknown phase and amplitude?
> 
> Though it might be that a linear least-squares fit to a low order
> polynomial would work, too.  Or maybe two sines plus a polynomial,
> where the sines do the higher frequencies and the polynomial the
> low frequencies.
> 
> -- glen
> 

I asked a different form of the question on sci.geo.satellite-nav. I 
eventually got a reply that one person uses Lagrange or Hermite 
interpolators depending on whether or not he has velocity data as well 
as position data. I have had a chance to look only at couple of 
Wikipedia articles. I've more reading to do. Any recommended sites - esp 
with worked examples?

I asked here with a slightly different purpose in mind.

Given a limited amount of data and some understanding of what the 
underlying process "looks like":
1. How much/what can we extrapolate/interpolate/infer/... ?
2. How do you go about solving the problem?

I've learned to be very careful in phrasing my questions on this group. 
If I give too many details of the _specific_ instance that triggered the 
question, I'll get a fish rather than learning to fish ;)

Richard Owlett wrote:

(snip)

> The system has 4 independent signal sources. Two of which are comparable 
> in magnitude of their fundamentals. Their frequencies differ by ~.2% .
> The other are 30-300+ times slower and much weaker. But ALL 4 
> fundamental frequencies are known very accurately.

How about a least-squares fit to the four sines of known frequency
but unknown phase and amplitude?

Though it might be that a linear least-squares fit to a low order
polynomial would work, too.  Or maybe two sines plus a polynomial,
where the sines do the higher frequencies and the polynomial the
low frequencies.

-- glen

dbd wrote:
> On Apr 13, 7:52 am, Richard Owlett <rowl...@atlascomm.net> wrote:
> 
>>c...@claysturner.com wrote:
>>
>>>>Yes to knowing period from other sources.
>>>>Points 1 -> 97 cover exactly 24 hours.
>>>>The period is 23 hours 56 minutes and some seconds.
>>
>>>Aha a sidereal period. Richard, there are quite a few sources that
>>>will let you find the osculating (not oscillating!) coefficients. A
>>>good book on orbital mechanics will likely prove useful.
>>
>>I've a limited budget (read as ~null ;)
>>What would be good keywords for Google searching?
>>
>>
>>
>>
>>>Given your simple case of data and limited interval for extrapolation,
>>>you can try a simple LPC method. It should be good enough.
>>
>>Yep. And Dale used a large enough hammer to wake me up to already
>>"knowing" the answer ;)
>>
>>
>>>Clay
> 
> 
> Richard
> 
> Given the additional information you have provided about the signals
> in your samples, I doubt that the DFT decomposition approach for your
> sample size will prove appropriate. The maximum likelihood approach
> might. Knowing the number of expected signals is an advantage. If you
> are interested in it try reading the second section of the last
> reference I provided. The last two references are theses and one of
> the advantages of theses over transactions papers is that they provide
> a description of the technical background instead of just references
> to it. Also the cost of the material from the source I cited is within
> your budget.
> 
> An interesting follow up on your effort might be to subtract the model
> of your signal (however obtained) from your samples and look for the
> frequencies that have peaks in the residual. Are there effects where
> you might expect them?
> 
> Dale B. Dalrymple

Downloaded both. Have some reading do ;)
May have more than enough reading time, retirement may come earlier than 
planned.