Search Engine **www.edaboard.com**

11 Threads found on edaboard.com: **Average Fft**

I know that I**fft**/**fft** are used in standard ofdm systems as a modulating/demodulating bases.Recently i've read that the orthogonality feature of DFT is also provided by Discrete Cosine Transform and by replacing I**fft** stage with IDCT can give better peak to **average** power reduction.Can anyone explain how this happen?

Digital communication :: 02-25-2013 15:41 :: Marina90 :: Replies: **0** :: Views: **712**

Hi Tunelabguy,
Refering to the steps in your previous post:
= In step 4 the **average** must be taken or not depending on the scaling factor of the **fft** algorithm you are using. The important fact is that you need total power.
= In step 5 you should multiply by 10 rather than 20 beacuse you have a power level yet
Nevertheless, it seems that the A-weig

Digital Signal Processing :: 10-30-2012 12:18 :: zorro :: Replies: **9** :: Views: **4317**

Calculate the RMS (Root Mean Square) it is the same as the standard deviation (1/n*Sum(value-**average**)^2

Digital Signal Processing :: 03-20-2012 21:32 :: klystron :: Replies: **2** :: Views: **795**

Hello!
Using **fft** - multiplication - I**fft** is called fast convolution.
I don't know if it's your intention, but it would be rather silly to use fast convolution for a
moving **average**. If you do a n-tap moving **average**, then you don't need a convolution.
You just add the current sample s and subtract the nth element s. Therefore (...)

Digital Signal Processing :: 04-11-2011 01:41 :: doraemon :: Replies: **5** :: Views: **4481**

I do not know what your "positive**fft**" is and a few other variables. At any rate, here is a simple code that calculates the **average** **fft**. I do not have matlab with me, so you might have to tweak it a little bit:Y_**average**=zeros(1,n);
w=sum(windowHann);
for m=0:9
Y=**fft**(x(1+m*n:n+m*n).*windowHann)*(2/w); (...)

Digital Signal Processing :: 06-03-2010 16:41 :: JoannesPaulus :: Replies: **6** :: Views: **3564**

Please not the textbook answer (my **fft** is 2^k (hardware reason) so because of guard bands it could be only a compact subset less than 2^k but greater than 2^(k-1))!
I have found that there exists good preamble symbols for 2^k used carriers which have very low peak to **average**. Also binary channel subdivision for OFDMA is easier and adaptivity i

Digital communication :: 03-29-2009 13:35 :: rfsystem :: Replies: **2** :: Views: **982**

quantization noise can get from **fft** with trans sim, but you can't get others such as thermal noise , 1/f noise since the **average** to time is zero

Analog Circuit Design :: 08-19-2008 09:48 :: iexplorer :: Replies: **3** :: Views: **1203**

Seems you have converted your waveform into the frequency domain by using the **fft**, which doesnt look to be necessary to caculate wavelgnth (λ).
To get the wavelength, better to count zero crossings (after you have smoothened out your signal) and **average** over a sufficently long interval. That will give you the frequency. Assuming this is so

Digital Signal Processing :: 06-17-2008 08:25 :: bulx :: Replies: **1** :: Views: **1006**

Hello DSP-ers,
I am new to DSP and MATLAB and I have a project in which I am to take a signal and partition it into blocks of length L (# of samples) and perform an **fft** of each block and then ensemble **average** the **fft** of each. Anywhere to start would be helpful. I am able to read in the signal into MATLAB.
I was also wondering if the (...)

Digital Signal Processing :: 06-03-2008 20:17 :: highQ :: Replies: **0** :: Views: **1240**

hi there...
can anyone help me in answering that what type of mean(**average**) is taken in selecting a center(carrier) frequency in the entire AM (audio frequency band) at which sampling can be performed for extracting a particular channel(radio frequency channel)...
i'll be highly greatful if someone helped me in...

Digital communication :: 08-07-2007 08:28 :: Nosheen :: Replies: **4** :: Views: **1108**

Hi
Can someone guide me to the right algorithm to **average** magnitude in **fft** spectrum. Can this be done with n-point moving **average** filter without need to store last n frames of magnitude spectrum? Now I?m using exponential moving **average** algorithm ? Xavr = alpha*Xavr + (1-alpha)*Xnew. Is there similar formula for n-point (...)

Digital Signal Processing :: 02-24-2006 18:31 :: stroma :: Replies: **2** :: Views: **3068**

Previous
1
Next

Last searching phrases:

take course | take course | keep out | keep out | nor not | cant get | three way | seven | seven | nobody

take course | take course | keep out | keep out | nor not | cant get | three way | seven | seven | nobody