PSN-L Email List Message

Subject: Re: Most Recent Discover Magazine pp66-67 WaterFall Display
From: "Geoffrey" gmvoeth@...........
Date: Sat, 14 Jul 2007 03:57:52 -0700

Hey thats far out.
If only i knew how to program such stuff myself
using PowerBasic.

It takes me now 11 seconds to look at only 384 lines of data
covering about 30 minutes of time.

If I could somehow speed that up I would sure like to know how.


----- Original Message ----- 
From: "Mike Price" 
Sent: Sunday, July 08, 2007 16:06
Subject: Re: Most Recent Discover Magazine pp66-67 WaterFall Display

> Geoff,
> A 3.2 GHz Pentium 4 with a vectorized FFT implementation using SSE2 
> (special extended 128-bit MMX instruction set of the Pentium) can 
> execute a 1024 point FFT in well under 1mS. A 4096 FFT can be executed 
> in under 6mS. This would blow the doors off any ca. 1972 implementation 
> even if done entirely in HW.
> The number of points needed in an FFT is driven by the desired frequency 
> resolution (bin width = sample_frequency / number_of_points). You may be 
> able to get away with fewer number of points if you don't need fine 
> frequency resolution.
> The FFT produces a complex number: a + bi. Think of it as a vector. You 
> get magnitude and phase as follows:
> mag = sqrt( a^2 + b^2 )
> phase = atan ( b / a )
> Mike Price
> Geoffrey wrote:
>> Wonderful and thanks;
>> It has been a few years since I looked
>> and I never found anything like that.
>> It (FFT) was classified Secret in the US Navy in the 1970s
>> but after I saw my "secret" machine sold to foreign
>> countries in a commissiond officers magazine
>> (My dad was a retired officer) I no longer accepted the navys 
>> classification.
>> for that machine.
>> I now realize they want to keep it a secret
>> only till they market the machine wherever.
>> The secrecy is more for economic (manufacturing)
>> than military reasons. We were probably only
>> testing the device for Diagnostic Retreival Company.
>> A general purpose computer like mine that works in a serial fashion
>> can not match the speed of that parallel computing machine.
> __________________________________________________________
> Public Seismic Network Mailing List (PSN-L)
> To leave this list email PSN-L-REQUEST@.............. with 
> the body of the message (first line only): unsubscribe
> See for more information.

Public Seismic Network Mailing List (PSN-L)

[ Top ] [ Back ] [ Home Page ]