Newbie to DSP - Phase shift detection

Hello everyone.

I'm completely new to DSP, coming from a background of PIC18.

I have two mics that I want to use for sound direction detection, and
I guess using fourier transforms to get the freqs    out from one
channel then looking for a rough match in the second channel over time
(and vice-versa) is the correct way to go.

I have no experience of dsPICs and I'm looking for
1. some pointers about learning the maths involved
2. some pointers about using dsPICs to get the job done painlessly
3. is there some IC commercially available that'll do this job.
4. am a barking up the correct tree

I've glanced over the comp.dsp faq list, but I think it may all be a
bit too advanced for me.  Not really up to A-level maths or above.
I'm looking for some application examples that'll show me the ropes
without pain.

Any advice / references / links / FAQs gratefully received.

On May 6, 7:00 pm, keithmarsh <keithmar...@gmail.com> wrote:
> Hello everyone.
>
> I'm completely new to DSP, coming from a background of PIC18.
>
> I have two mics that I want to use for sound direction detection, and
> I guess using fourier transforms to get the freqs    out from one
> channel then looking for a rough match in the second channel over time
> (and vice-versa) is the correct way to go.
>
> I have no experience of dsPICs and I'm looking for
> 1. some pointers about learning the maths involved
> 2. some pointers about using dsPICs to get the job done painlessly
> 3. is there some IC commercially available that'll do this job.
> 4. am a barking up the correct tree
>
> I've glanced over the comp.dsp faq list, but I think it may all be a
> bit too advanced for me.  Not really up to A-level maths or above.
> I'm looking for some application examples that'll show me the ropes
> without pain.
>
> Any advice / references / links / FAQs gratefully received.

Direction-finding isn't really a beginner's topic. I'm not meaning to
throw a wet blanket on you here, but you might be better served with
some more introductory DSP material, like a good book. Something with
good appeals to intuition is important if the math intimidates you.
You might try "Understanding Digital Signal Processing" by comp.dsp's
own Rick Lyons. It's a great introductory text.

Jason

On May 7, 6:59 am, cincy...@gmail.com wrote:
> On May 6, 7:00 pm, keithmarsh <keithmar...@gmail.com> wrote:
>
>
>
> > Hello everyone.
>
> > I'm completely new to DSP, coming from a background of PIC18.
>
> > I have two mics that I want to use for sound direction detection, and
> > I guess using fourier transforms to get the freqs    out from one
> > channel then looking for a rough match in the second channel over time
> > (and vice-versa) is the correct way to go.
>
> > I have no experience of dsPICs and I'm looking for
> > 1. some pointers about learning the maths involved
> > 2. some pointers about using dsPICs to get the job done painlessly
> > 3. is there some IC commercially available that'll do this job.
> > 4. am a barking up the correct tree
>
> > I've glanced over the comp.dsp faq list, but I think it may all be a
> > bit too advanced for me.  Not really up to A-level maths or above.
> > I'm looking for some application examples that'll show me the ropes
> > without pain.
>
> > Any advice / references / links / FAQs gratefully received.
>
> Direction-finding isn't really a beginner's topic. I'm not meaning to
> throw a wet blanket on you here, but you might be better served with
> some more introductory DSP material, like a good book. Something with
> good appeals to intuition is important if the math intimidates you.
> You might try "Understanding Digital Signal Processing" by comp.dsp's
> own Rick Lyons. It's a great introductory text.
>
> Jason

You may want to refer DSP using TMS#@)C6x by Chassing.

 There is a similar problem given as a example with source code.

On May 7, 11:00 am, keithmarsh <keithmar...@gmail.com> wrote:
> Hello everyone.
>
> I'm completely new to DSP, coming from a background of PIC18.
>
> I have two mics that I want to use for sound direction detection, and
> I guess using fourier transforms to get the freqs    out from one
> channel then looking for a rough match in the second channel over time
> (and vice-versa) is the correct way to go.
>
> I have no experience of dsPICs and I'm looking for
> 1. some pointers about learning the maths involved
> 2. some pointers about using dsPICs to get the job done painlessly
> 3. is there some IC commercially available that'll do this job.
> 4. am a barking up the correct tree
>
> I've glanced over the comp.dsp faq list, but I think it may all be a
> bit too advanced for me.  Not really up to A-level maths or above.
> I'm looking for some application examples that'll show me the ropes
> without pain.
>
> Any advice / references / links / FAQs gratefully received.

You sound like an amateur. Go study Elect Engineering first and come
back in 4 years. I often want to do brain surgery but they won't let
me.

K.

On 7 Mai, 01:00, keithmarsh <keithmar...@gmail.com> wrote:
> Hello everyone.
>
> I'm completely new to DSP, coming from a background of PIC18.
>
> I have two mics that I want to use for sound direction detection, and
> I guess using fourier transforms to get the freqs &#4294967295; &#4294967295;out from one
> channel then looking for a rough match in the second channel over time
> (and vice-versa) is the correct way to go.

No, it's not.

> 1. some pointers about learning the maths involved

The maths required includes

- Complex numbers
- Fourier transforms
- Z transforms

and while not an absolute must, knowledge in linear
algebra will significantly ease your pains with the
engineering.

> 2. some pointers about using dsPICs to get the job done painlessly
> 3. is there some IC commercially available that'll do this job.

Almost certainly.

> 4. am a barking up the correct tree

What tree? Maths? You need to learn a lot if you want to do
things yourself. DSP? You are certainly in the right place,
although you need to learn a lot before discussing specifics.
Getting an off-the-shelf device? Maybe, maybe not. Depends
on why you want this sort of device and how much $$ you have
to spend.

> I've glanced over the comp.dsp faq list, but I think it may all be a
> bit too advanced for me. &#4294967295;Not really up to A-level maths or above.
> I'm looking for some application examples that'll show me the ropes
> without pain.

You're out of luck. DSP is applied maths. I don't know
the maths then you can't do these sorts of things yourself.
The only painless way is to buy your way out, either by
buying a canned device or somebody to d things for you.

Rune

On Tue, 6 May 2008 16:00:56 -0700 (PDT), keithmarsh
<keithmarsh2@gmail.com> wrote:

>Hello everyone.
>
>I'm completely new to DSP, coming from a background of PIC18.
>
>I have two mics that I want to use for sound direction detection, and
>I guess using fourier transforms to get the freqs    out from one
>channel then looking for a rough match in the second channel over time
>(and vice-versa) is the correct way to go.
>
>I have no experience of dsPICs and I'm looking for
>1. some pointers about learning the maths involved
>2. some pointers about using dsPICs to get the job done painlessly
>3. is there some IC commercially available that'll do this job.
>4. am a barking up the correct tree
>
>I've glanced over the comp.dsp faq list, but I think it may all be a
>bit too advanced for me.  Not really up to A-level maths or above.
>I'm looking for some application examples that'll show me the ropes
>without pain.
>
>Any advice / references / links / FAQs gratefully received.

Hi Keith,
  To echo what Jason (cincydsp) said, you've picked 
an "awfully high signal processing mountain 
to climb." 

Google on "direction finding". You'll get thousands
of hits.  Review some of the tutorial material 
on the web and you'll start to develop a feeling 
for what's involved in "direction finding".
The material on the web will mostly describe 
radio frequency (RF) direction finding, but the 
basic ideas, the basic principles, you encounter 
on the web will also apply to audio direction finding.

I know almost nothing about direction finding, but 
I do know that it's NOT a simple problem.

Good Luck,
[-Rick-] 
 

On May 7, 3:43 am, Rune Allnor <all...@tele.ntnu.no> wrote:
> > 1. some pointers about learning the maths involved
>
> The maths required includes
>
> - Complex numbers
> - Fourier transforms
> - Z transforms
>
> and while not an absolute must, knowledge in linear
> algebra will significantly ease your pains with the
> engineering.

For limited purposes, it's possible to formulate the Fourier
transform without complex numbers, just a lot of trigonometry
and calculus (although complex algebra does make it possible
to simplify and generalize the results).

For a general introduction to direction finding at the
popular science level, I recommend the book:
"Blip, Ping, and Buzz: Making Sense of Radar and Sonar"
by Mark Denny
Johns Hopkins University Press 2007



IMHO. YMMV.
--
rhn A.T nicholson d.0.t C-o-M

kronecker@yahoo.co.uk wrote:

   ...

> You sound like an amateur. Go study Elect Engineering first and come
> back in 4 years. I often want to do brain surgery but they won't let
> me.

What a meat-handed, narrow-minded thing to say! Isaac Newton was an 
amateur too.

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

Thanks for your input all.

Yes, I know it's a big hill to climb, but the path starts here, and
some of you have pointed the right route to take.

I came across this interesting device in my search.  May save me 4
years, eh?

http://www.analog.com/en/prod/0,,770_847_AD8302,00.html

Thanks all

Keith

Oh, and this proved very useful as well.

http://www.scribd.com/word/full/5871?access_key=ccldgmcq6puwm