[sdiy] Sampling Aliases and Images

[Richie Burnett]

> Would it be correct to say that half of the images are mirror images?

I would say Yes.  Half of the infinitely repeating images of the original 
spectrum in a sampled data set are spectrally reversed or mirrored.

> In terms of high frequencies being output lower than low input 
> frequencies?

I don't think they are output at a lower frequency than the input frequency. 
They just repeat over and over again all the way up to infinity Hertz!

> Between Nyquist and the sample rate, would you call those frequencies 
> "images" or "aliases"?

I'd definitely call them images if we are talking about spectral content in 
the output of a DAC that hasn't been processed with an anti-imaging 
"reconstruction" filter yet.  Here I'm talking about that unwanted HF hash 
from those sharp little steps in the output of a DAC that would get removed 
by the subsequent low-pass filtering action of the reconstruction filter.

But if instead we are talking about some tones that you are feeding into an 
ADC that are anywhere above the Nyquist frequency (half the sample rate) 
then I would say that these tones *will result in aliasing taking place*, 
and I would call the resulting tones that are now below the Nyquist 
frequency *aliases*.

> I assume we can all agree that frequencies above the sample rate are 
> aliased down below the sample rate.

I would say that frequencies above *half* the sample rate (Nyquist limit) 
are aliased down below *half* the sample rate when you sample a 
continuous-time signal.

> Sounds like I'm going to have to review some terminology.

I don't claim to have all the terminology down and right :-)  I currently 
work in a University environment and only recently started to pay close 
attention to the meaning of these terms when I started talking to staff and 
students about this stuff, because I was worried I'd say something wrong or 
mislead them!

Best regards,

-Richie, 


-- 
This email has been checked for viruses by AVG.
https://www.avg.com