Reply by "Fel...@yahoo.com [audiodsp]"●November 13, 20142014-11-13
Hi,
You can use fdatool in Matlab to design a bandpass filter. Yo have several
choices, such as Elliptic, Chebyshev, etc.
As you have an idea of how is your signal, you can determine empirically which
filter and parameters fit best to your data.
Best,
Felipe
Reply by "sen...@yahoo.co.za [audiodsp]"●November 10, 20142014-11-10
I have 24 bit ADC data with dithering noise on it. My original signals into the
ADC were quite small. The dithering noise therefore makes out a large part of
the amplitude of the ADC data. I want to remove the noise to recover as much as
possible of the actual signals. I find lots of websites that say that the
dithering noise can be filtered out but no description of how this can be
done.
Unfortunately I know very little about DSP filters so I do not know which type
of filter would be best. This is the first time I have tried to implement DSP
filters on real data. I used a brick wall FFT filter in Excel but found that the
final signals had a much lower amplitude if I selected only one frequency bin or
had ringing when I selected more frequency bins . I tried using Hamming and
Chebychef filters in Matlab but got out totally distorted signals.
Of all the methods I have tried, moving average has worked best but it seems
like it should be possible to filter out more noise.
My source data is a low level 63Hz sine wave with the dithering noise added. It
would help me a lot if somebody can just tell me what type of filter I should
use. Then I can go read up how it works and hopefully figure out how to
implement it to remove the dithering noise.