Yep, that is not really correct, because:
when using a frame size of 1024, the lowest tone that can be measured is:
s_rate/framsize --> 44100 / 1024 = 43.06640625
Thats why I tested the amplitude formula with a 43 hz sine wave, and a 86 hz
tone (which would be the second harmonic) even if I use a 86 hz tone, the
seconde table index returns an amplitude vector of 2... even if there is no
43 hz information in the signal that would cause such a dramatic amplitude
on that frequency.......
But I'm also still wondering wat the max value is of the im and re tables,
if I measure the amplitude of the 43 hz tone with the following formula:
amp[index] = pow (pow (tableread(re, index), 2) + pow(tableread(im, index),
2) , 0.5);
amp[1] = pow(pow (tableread(re, 1), 2) + pow(tableread(im, 1), 2)), 0.5);
I get 4.12313 out, which would mean that:
sqrt( pow(4.12313, 2) / 2) = 2.195
The imaginairy component of the vector would be over 2?????
What type of value is that?
Peter
----- Original Message -----
From: John Lazzaro <lazzaro@CS.Berkeley.EDU>
To: <peter.maas@student-kmt.hku.nl>; <saol-dev@media.mit.edu>
Sent: Saturday, September 16, 2000 8:52 PM
Subject: Re: fft in saol
> > if I use this formula I get amplitude values which don't make sense, if
> > i put a 43 hz sine into a 1024 samples framed fft opcode
>
> Hmm, one thing comes to mind -- 43 Hz is a low frequency, you may not
> be getting many cycles in your 1024 frame, which will make amplitude
> estimation noisy. Try the experiment first with 1Khz and see if the
> numbers make sense ... although I'm really not a spectral estimation
> guy, so this could be an incorrect analysis ...
>
> --jl
>
> -------------------------------------------------------------------------
> John Lazzaro -- Research Specialist -- CS Division -- EECS -- UC Berkeley
> lazzaro [at] cs [dot] berkeley [dot] edu www.cs.berkeley.edu/~lazzaro
> -------------------------------------------------------------------------
This archive was generated by hypermail 2b29 : Mon Jan 28 2002 - 12:03:59 EST