> Assume i receive a signal and perform direct conversion from passband to
> baseband with a front end analog filter with bandwidth +/- B for my
> signal.
This must be homework, because only in a homework problem or a classroom
do you encounter a perfectly bandlimited signal. They simply don't occur
in nature (feel free to explain why).
> In my ADC I sample at 2B and have a signal plus noise limited
> to the band +/-B with each noise sample assumed to be uncorrelated. If i
> instead decide to sample the signal at the output of the analog filter
> at 10B i will have correlated noise on the samples.
Will you? Where is the noise coming from? Will _all_ the noise in a
_real_ system come from _just_ the sources that are spelled out in a
homework assignment?
> Does this mean that if i have a simulation with oversampling greater
> than 2B, i have to add non-white noise prior to performing any signal
> processing on the samples?
Maybe. What are you simulating?
--
www.wescottdesign.com
Reply by CW●May 22, 20092009-05-22
Assume i receive a signal and perform direct conversion from passband
to baseband with a front end analog filter with bandwidth +/- B for my
signal. In my ADC I sample at 2B and have a signal plus noise limited
to the band +/-B with each noise sample assumed to be uncorrelated.
If i instead decide to sample the signal at the output of the analog
filter at 10B i will have correlated noise on the samples.
Does this mean that if i have a simulation with oversampling greater
than 2B, i have to add non-white noise prior to performing any signal
processing on the samples?
CW