Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronic Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

problem regarding decimation of samples

Status
Not open for further replies.

shameem

Member level 5
Joined
Oct 27, 2005
Messages
92
Helped
5
Reputation
10
Reaction score
3
Trophy points
1,288
Activity points
2,042
hi,
i am working on the dab receiver designing.i am having a signal containing 8.192 mega samples of 8 bit adc, but i want to use fft window size 2048. So i want to decimate the samples by a factor 4. i want to know that wheather a fir decimation filter is necessary to decimate or i can decimate the samples simply by taking 1 samples in group of 4 samples.
if anybody have an idea, please help me.
 

Circuit_seller

Full Member level 5
Joined
Sep 23, 2002
Messages
302
Helped
17
Reputation
34
Reaction score
7
Trophy points
1,298
Activity points
2,354
Hi

I can not understand why you should decimate by four. You can directly perfrom a 256-point fft on 8.192 MSPS signal.
 

pop48m

Junior Member level 2
Joined
Oct 13, 2005
Messages
20
Helped
5
Reputation
10
Reaction score
2
Trophy points
1,283
Activity points
1,415
shameem said:
hi,
i am working on the dab receiver designing.i am having a signal containing 8.192 mega samples of 8 bit adc, but i want to use fft window size 2048. So i want to decimate the samples by a factor 4. i want to know that wheather a fir decimation filter is necessary to decimate or i can decimate the samples simply by taking 1 samples in group of 4 samples.
if anybody have an idea, please help me.

It depends on tha ratio of the highest spectral component of the signal and the sampling frequency.
If this ratio is less than 1/8, then you can safely take one of four samples.
If this ratio is higher than 1/8, some form of filtering should be done before the decimation. The simplest filter would be to make an average of a group of four samples and teke this average as a new decimated sample.

Hope this helps.

Pop
 

alizadeh_m

Junior Member level 2
Joined
Oct 29, 2005
Messages
24
Helped
1
Reputation
2
Reaction score
0
Trophy points
1,281
Activity points
1,475
It depends whether the sampling rate used is above 4 time the nyquist rate or below. If it is above, dropping three samples out of four, wont cause a problem.
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Top