> Don't use math as an excuse not to think. (I use thinking as an excuse > not to do math. That way bites too!) Remember: we suppose that the > analog signal contains no energy above half the lower sampling > frequency, so decimating won't cause aliasing. Otherwise, we would need > to low-pass before decimating, melting down the whole argument.Yes but quantization noise is white so the *noise* does extend to the 10*fs/2 frequency and does get aliased back in when we decimate. The question was really whether this noise that is aliased back in is greater than, equal to, or less than the noise that would be present without oversampling and dropping samples. The answer appears to be "equal-to" and this makes sense - it's kind of a conservation of energy thing (where else would the noise power go). This is more of a hypothetical question of what happens to quantization noise if we don't lpf before drop-sampling - I would never do this in practice.
Oversampling w/ Drop interpolation
Started by ●August 17, 2004
Reply by ●August 19, 20042004-08-19
Reply by ●August 19, 20042004-08-19
> Hi Jerry, > I wonder if this is a homework problem. > > [-Rick-]Wow - I'm a bit offended by that assertion! Nicely done. My only homework these days is changing diapers and playing with blocks, which turns out to be more challenging than the academic kind. Of course maybe it sounds like a HW question because it makes no practical sense. Seriously, it is juat a "what if" that I'm having troubles with. Thanks all for the different perspective - sometimes we just don't look at things from the right viewpoint and it makes simple things hard.
Reply by ●August 19, 20042004-08-19
Andor Bariska <an2or@nospam.net> wrote in message news:<412372d0$1@pfaff2.ethz.ch>...> Dennis M wrote: > .. > > But doesn't drop sampling have a LPF effect? > > It's the other way around: you have to lowpass _before_ you drop > samples. The effect of just droping samples is as follows:I understand that - I'm quite familiar with how to resample a signal. This is not something I would do in practice, just a "what-if" type of thing. Anyway, let me try again: if I oversample (10xfs) a bandlimited signal (fs/2) the quantizatuin noise will be spread from (0 to 5xfs). Now if I drop-sample decimate this signal will the noise power in the resulting signal be the same as if I had simply sampled at fs?
Reply by ●August 19, 20042004-08-19
I just rememberred the other thing that got me going down the line of thought that drop sampling had a lpf effect: http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&threadm=_TE08.5373%24pe.2878%40nwrddc01.gnilink.net&rnum=6&prev=/groups%3Fq%3Ddrop%2Bsample%2Bfrequency%2Bresponse%26hl%3Den%26lr%3D%26ie%3DUTF-8%26selm%3D_TE08.5373%2524pe.2878%2540nwrddc01.gnilink.net%26rnum%3D6 This made a lot of sense to me from a signals point of view. From a common sense point of view it does not (as Jerry's example flushed out). - Dennis
Reply by ●August 19, 20042004-08-19
Dennis M wrote:> Andor Bariska <an2or@nospam.net> wrote in message news:<412372d0$1@pfaff2.ethz.ch>... > >>Dennis M wrote: >>.. >> >>>But doesn't drop sampling have a LPF effect? >> >>It's the other way around: you have to lowpass _before_ you drop >>samples. The effect of just droping samples is as follows: > > > I understand that - I'm quite familiar with how to resample a signal. > This is not something I would do in practice, just a "what-if" type of > thing. > > Anyway, let me try again: if I oversample (10xfs) a bandlimited signal > (fs/2) the quantizatuin noise will be spread from (0 to 5xfs). Now if > I drop-sample decimate this signal will the noise power in the > resulting signal be the same as if I had simply sampled at fs?Not just the noise, nut also, if you arrange it that way, also the actual samples. Why is it so hard to see that? Jerry -- ... the worst possible design that just meets the specification - almost a definition of practical engineering. .. Chris Bore ������������������������������������������������������������������������
Reply by ●August 19, 20042004-08-19
Dennis M wrote:> I just rememberred the other thing that got me going down the line of > thought that drop sampling had a lpf effect: > > http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&threadm=_TE08.5373%24pe.2878%40nwrddc01.gnilink.net&rnum=6&prev=/groups%3Fq%3Ddrop%2Bsample%2Bfrequency%2Bresponse%26hl%3Den%26lr%3D%26ie%3DUTF-8%26selm%3D_TE08.5373%2524pe.2878%2540nwrddc01.gnilink.net%26rnum%3D6 > > This made a lot of sense to me from a signals point of view. From a > common sense point of view it does not (as Jerry's example flushed > out). > > - DennisI think it's a stretch to describe a zero-order hold as drop-sample decimation, and I can't see it as any kind of interpolation (What is drop-sample interpolation anyway?) no matter how askance I look. Jerry -- ... the worst possible design that just meets the specification - almost a definition of practical engineering. .. Chris Bore ������������������������������������������������������������������������
Reply by ●August 20, 20042004-08-20
> I think it's a stretch to describe a zero-order hold as drop-sample > decimation, and I can't see it as any kind of interpolation (What is > drop-sample interpolation anyway?) no matter how askance I look.I'm not sure if your question was rhetorical but from my view, drop sample interpolation means picking the sample that is closest to your interpolation point. I may have thrown this thread way out of whack by generalizing the question too much. The system in my minde may not be an integer factor of oversampling and there are some other factors as to where your interpolation point may lie. - Dennis
Reply by ●August 20, 20042004-08-20
> However, if quantization noise extends the spectrum beyond B then the result > is obviously different. This seems to be the case you're interested in. > "Drop sampling" / decimation doesn't reduce the higher frequencies at all. > It disguises them because they've been folded into lower frequencies - thus > incereasing the lower frequency noise energy.Thanks for hanging with me on this one. There seems to be a real are to asking the question correctly and I think I've blown it. You seem to be onto what I was originally asking. I'm assuming bandlimited input B and fs = 2*B and oversampling rate fs` = K x fs (where K can even be non-integer, thus the term drop sample interpolation). With not prefiltering the noise power due to quantization noise would be the same - I see that now. But when that quantization noise gets folded back in due to aliasing would the spectrum of the noise look the same as if it had simply been sampled at the lower frequency (i.e. not oversampled). Not in my mind! I think you're right that I just need to sit down run some Matlab on it. That will probably clear it up much quicker. I'll do that and hopefully I can put together a post at some later date that describes why this is not as simple of a problem as it might appear at first. - Dennis
Reply by ●August 20, 20042004-08-20
"Jerry Avins" <jya@ieee.org> wrote in message news:4124de92$0$21752$61fed72c@news.rcn.com...> Dennis M wrote: > > > This made a lot of sense to me from a signals point of view. From a > > common sense point of view it does not (as Jerry's example flushed > > out). > > > > - Dennis > > I think it's a stretch to describe a zero-order hold as drop-sample > decimation, and I can't see it as any kind of interpolation (What is > drop-sample interpolation anyway?) no matter how askance I look.In my experience, drop-sample is a pretty common way of describing things. Personally, I call it "skip and repeat" sample rate conversion (for decimation, it would be just skipping). This ends up happening frequently in digital audio systems if the sample rate clocks of 2 devices are not exactly aligned. But that is usually an "automatic" feature of the buffering/FIFIO of the digital audio receiver chip, not something that is intentionally added. I probably wouldn't call it "drop-sample interpolation" but I might say "drop-sample decimation/conversion". But for the upsampling equivalent, "repeat-sample interpolation" does make some sense--you are interpolating between the samples using a zero-order hold interpolator, aka just repeating samples. This does have a low pass filtering effect, but it's a pretty lousy filter!
Reply by ●August 21, 20042004-08-21
On 19 Aug 2004 09:11:54 -0700, dennis.merrill@thermo.com (Dennis M) wrote:>> Hi Jerry, >> I wonder if this is a homework problem. >> >> [-Rick-] > > >Wow - I'm a bit offended by that assertion! Nicely done. My only >homework these days is changing diapers and playing with blocks, which >turns out to be more challenging than the academic kind. > >Of course maybe it sounds like a HW question because it makes no >practical sense. Seriously, it is juat a "what if" that I'm having >troubles with. Thanks all for the different perspective - sometimes >we just don't look at things from the right viewpoint and it makes >simple things hard.Hi Dennis, please don't be offended. It just "seemed" like a homework problem. Please take my comment as a compliment because your question sounded, well, so academic. I understand your feeling though. A couple of years ago I posted a math question here (I needed help because my math skills are so weak) and my question appeared so fundamental that one of our pals here thought my question was a homework problem. I was offended. Your words: "sometimes we just don't look at things from the right viewpoint and it makes simple things hard." made me laugh. Your words describe, perfectly, about 75% of my career! See Ya, [-Rick-]
