Re: Trying to follow the math behind wavelets

Martin,

Thanks for the interest.  I haven't given up on my research, but
other things keep interrupting.

On 11 Oct 2008 19:03:50 GMT, Martin Eisenberg <martin.eisenberg@[EMAIL PROTECTED]
>
wrote:
> Fr**** McKenney wrote:
> 
>> On the off chance that someone muight be curious, I thought I'd
>> post a sample of a WWV waveform being "decoded" using wavelets. 
> 
> So you haven't drowned after all.

No, but we haven't had much rain lately, so there's still hope.
<grin!>

> ... With hindsight, do you see any 
> benefits of wavelets over other approaches to this task?

The "demo" I posted shows me that I _can_ use wavelets to clearly
pick WWV's digital time-code information out of an ideal (e.g.
Scilab-generated) signal.

On the other hand, using the same techniques on parts of a WAV file
recorded off a Heathkit MAC-II with an indoor antenna shows me that
this data is a lot harder to pick out from a RealWorld(tm) signal.
The audible ****tions of the recording are moderately clear most of
the time, so I'm assuming I didn't lose too much of the 100Hz during
recording, and I _think_ I can spot the data bits here and there,
but they come and go as I work my way through the WAV file.

Next step:  use wavelets to try reducing the noise.

Several of the papers in my collection describe what seems like a
reasonable technique:  dropping the signal gain at points where
there is "wideband" activity, that is, the wavelet "filtering" shows
presence simultaneously across a wide range of scales.

Hm.  And, just for grins (since it only now occurred to me), I think
I'll check the 100Hz performance of the recording system's audio
sup****t; all the K8N-E book says is "Realtek ALC850", so I may have
to dig out my signal generator.


Frank McKenney, McKenney Associates
Richmond, Virginia / (804) 320-4887
Munged E-mail: frank uscore mckenney ayut mined spring dawt cahm (y'all)