Hi all, I read on a book stating that "if the amplitude characteristic of the transfer function of a system is antenuating at high frequency, then it is good for combating noise at high freqency" which seems to me is that the more antennuation at high frequency, the better in terms of noise performance... What does this mean? ( I have never designed a system so I have never thought about noise under high frequency...) Why should the noise under high frequency be concerned? How is it more significant than noise under low frequency? So a high pass filter is not good? A differentiator is a low pass filter, right? A integrator is a high pass filter, so it is bad? Thanks a lot!
What's important about noise in high frequency domain?
Started by ●October 7, 2005
Reply by ●October 7, 20052005-10-07
"lucy" <losemind@yahoo.com> wrote in message news:1128665632.243188.300390@g47g2000cwa.googlegroups.com...> Hi all, > > I read on a book stating that "if the amplitude characteristic of the > transfer function of a system is antenuating at high frequency, then it > is good for combating noise at high freqency" > > which seems to me is that the more antennuation at high frequency, the > better in terms of noise performance... > > What does this mean? ( I have never designed a system so I have never > thought about noise under high frequency...) > > Why should the noise under high frequency be concerned? How is it more > significant than noise under low frequency? > > So a high pass filter is not good? A differentiator is a low pass > filter, right? A integrator is a high pass filter, so it is bad? >Context helps to understand why an author might have said that. The quote you give is a trivial truism. "If you attenuate high frequencies then you won't have high frequencies (as much)" I think what this probably alludes to is this: In many systems the physical properties are such that high frequency signals *and* disturbances (noise, etc.) won't be reacted to. Many or most physical systems are ultimately lowpass in nature. So, if high frequency noise is initially a concern, should it be a concern? That's the issue it seems the quote would be addressing. Any attenuation of frequency ranges that are not important or are not of interest will reduce the inevitable noise that combines from those frequency ranges. So, a highpass filter can do this just as readily as a lowpass filter (see DC blocking for example). It is just that so many physical systems naturally filter high frequencies and sometimes there are important sources of high frequency noise (power line noise on very low frequency systems comes to mind) and, combined, the situation is advantageous .. that's all. Fred
Reply by ●October 7, 20052005-10-07
"Fred Marshall" <fmarshallx@remove_the_x.acm.org> wrote in message news:hv2dnb9PC6DEGtvenZ2dnUVZ_tKdnZ2d@centurytel.net...> > "lucy" <losemind@yahoo.com> wrote in message > news:1128665632.243188.300390@g47g2000cwa.googlegroups.com... >> Hi all, >> >> I read on a book stating that "if the amplitude characteristic of the >> transfer function of a system is antenuating at high frequency, then it >> is good for combating noise at high freqency" >> >> which seems to me is that the more antennuation at high frequency, the >> better in terms of noise performance... >> >> What does this mean? ( I have never designed a system so I have never >> thought about noise under high frequency...) >> >> Why should the noise under high frequency be concerned? How is it more >> significant than noise under low frequency? >> >> So a high pass filter is not good? A differentiator is a low pass >> filter, right? A integrator is a high pass filter, so it is bad? >> > > Context helps to understand why an author might have said that. > The quote you give is a trivial truism. > "If you attenuate high frequencies then you won't have high frequencies > (as much)" > > I think what this probably alludes to is this: > In many systems the physical properties are such that high frequency > signals *and* disturbances (noise, etc.) won't be reacted to. Many or > most physical systems are ultimately lowpass in nature. > > So, if high frequency noise is initially a concern, should it be a > concern? That's the issue it seems the quote would be addressing. > > Any attenuation of frequency ranges that are not important or are not of > interest will reduce the inevitable noise that combines from those > frequency ranges. So, a highpass filter can do this just as readily as a > lowpass filter (see DC blocking for example). It is just that so many > physical systems naturally filter high frequencies and sometimes there are > important sources of high frequency noise (power line noise on very low > frequency systems comes to mind) and, combined, the situation is > advantageous .. that's all. >And I suppose that if you sample after this Low-Pass Effect it helps reduce the amount of high frequency noise that aliasses back in ? Best of Luck - Mike
Reply by ●October 7, 20052005-10-07
a differentiator is a high pass filter with 6 dB per octave slope upwards an integrator is a low pass filter with a 6 dB per octave slope downward.. Noise filtering depends on the relationship between the noise and the signal you want. Mark
Reply by ●October 7, 20052005-10-07
lucy wrote:> Hi all, > > I read on a book stating that "if the amplitude characteristic of the > transfer function of a system is antenuating at high frequency, then it > is good for combating noise at high freqency" > > which seems to me is that the more antennuation at high frequency, the > better in terms of noise performance... > > What does this mean? ( I have never designed a system so I have never > thought about noise under high frequency...) > > Why should the noise under high frequency be concerned? How is it more > significant than noise under low frequency? > > So a high pass filter is not good? A differentiator is a low pass > filter, right? A integrator is a high pass filter, so it is bad? > > Thanks a lot!What is the context? Vinyl records and FM radio boost ("preemphasize") the highs to improve S/N there, then restore the overall response ("equalize"), attenuating also high-frequency noise. The noise in such media tends to be white -- equal power in equal amounts of bandwidth -- so there is as much noise in the top octave as there is in all the signal below it. That's why high-frequency noise dominates. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
