Hi, this might be a very basic question, but it is not clear to me. What is the relation between a filter and determining the peak (or RMS) of a signal? If you sample a signal and search for its peak, is it that the frequency at which you return the peak of a signal, represents the cut-off frequency of a filter? If this statement is wrong, then what is the relation?
peak detect = filter?
Started by ●January 18, 2007
Reply by ●January 18, 20072007-01-18
dtsao wrote:> Hi, > > this might be a very basic question, but it is not clear to me. What is > the relation between a filter and determining the peak (or RMS) of a > signal? If you sample a signal and search for its peak, is it that the > frequency at which you return the peak of a signal, represents the cut-off > frequency of a filter? If this statement is wrong, then what is the > relation?The peak value that a signal occurs at a particular instant of time. Where doe frequency come into that? I think you're asking for a relation that doesn't exist. Jerry -- Engineering is the art of making what you want from things you can get. ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
Reply by ●January 18, 20072007-01-18
Oh, ok. I consider that the input signal is a sine wave, but where it's amplitude might not always be exactly the same (just off slightly at each peak from the previous). So there is a frequency here.
Reply by ●January 19, 20072007-01-19
dtsao wrote:> Oh, ok. I consider that the input signal is a sine wave, but where it's > amplitude might not always be exactly the same (just off slightly at each > peak from the previous). So there is a frequency here.It's not a pure sin if the peak amplitude changes. A Fourier transform will tell you what the additional frequencies are, but nothing about time. You are combining concepts here in an incompatible way. I haven't figured out yet how to separate them for you. Jerry -- Engineering is the art of making what you want from things you can get. ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
Reply by ●January 19, 20072007-01-19
Jerry Avins wrote:> dtsao wrote: >> Oh, ok. I consider that the input signal is a sine wave, but where it's >> amplitude might not always be exactly the same (just off slightly at each >> peak from the previous). So there is a frequency here. > > It's not a pure sin if the peak amplitude changes. A Fourier transform > will tell you what the additional frequencies are, but nothing about > time. You are combining concepts here in an incompatible way. I haven't > figured out yet how to separate them for you. > > JerryA sine, even of [apparent] constant frequency, that changes it's amplitude does indeed have frequency components apart from the fundamental. Just what those are depend on *how* the sine varies. As Jerry notes, a sine that varies in any way is no longer a pure sine. But merely reading the peak value does not tell one the *extra* frequencies involved. Look at it this way: An amplitude modulated sine will have peak amplitudes that vary at the modulating rate, so if you captured all successive peaks, you could indeed state *a* modulating frequency, but you would not necessarily have all the necessary information about the signal - you have only one piece of information. A phase modulated signal has amplitude modulation components *and* frequency modulation components. As always, a complete view requires that the entire signal (not merely peak amplitudes) be viewed. That's not as clear as I would like; perhaps tomorrow after I've recovered from the week :) Cheers PeteS
Reply by ●January 19, 20072007-01-19
PeteS wrote:> Jerry Avins wrote: >> dtsao wrote: >>> Oh, ok. I consider that the input signal is a sine wave, but where it's >>> amplitude might not always be exactly the same (just off slightly at >>> each >>> peak from the previous). So there is a frequency here. >> >> It's not a pure sin if the peak amplitude changes. A Fourier transform >> will tell you what the additional frequencies are, but nothing about >> time. You are combining concepts here in an incompatible way. I >> haven't figured out yet how to separate them for you. >> >> Jerry > > A sine, even of [apparent] constant frequency, that changes it's > amplitude does indeed have frequency components apart from the > fundamental. Just what those are depend on *how* the sine varies. As > Jerry notes, a sine that varies in any way is no longer a pure sine. > > But merely reading the peak value does not tell one the *extra* > frequencies involved. > > Look at it this way: > > An amplitude modulated sine will have peak amplitudes that vary at the > modulating rate, so if you captured all successive peaks, you could > indeed state *a* modulating frequency, but you would not necessarily > have all the necessary information about the signal - you have only one > piece of information. A phase modulated signal has amplitude modulation > components *and* frequency modulation components. > > As always, a complete view requires that the entire signal (not merely > peak amplitudes) be viewed. > > That's not as clear as I would like; perhaps tomorrow after I've > recovered from the week :)Pete, There is no variation of the peak amplitude at all with pure phase or frequency modulation. Jerry -- Engineering is the art of making what you want from things you can get. ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
Reply by ●January 19, 20072007-01-19
Jerry Avins wrote:> PeteS wrote: >> Jerry Avins wrote: >>> dtsao wrote: >>>> Oh, ok. I consider that the input signal is a sine wave, but where it's >>>> amplitude might not always be exactly the same (just off slightly at >>>> each >>>> peak from the previous). So there is a frequency here. >>> >>> It's not a pure sin if the peak amplitude changes. A Fourier >>> transform will tell you what the additional frequencies are, but >>> nothing about time. You are combining concepts here in an >>> incompatible way. I haven't figured out yet how to separate them for >>> you. >>> >>> Jerry >> >> A sine, even of [apparent] constant frequency, that changes it's >> amplitude does indeed have frequency components apart from the >> fundamental. Just what those are depend on *how* the sine varies. As >> Jerry notes, a sine that varies in any way is no longer a pure sine. >> >> But merely reading the peak value does not tell one the *extra* >> frequencies involved. >> >> Look at it this way: >> >> An amplitude modulated sine will have peak amplitudes that vary at the >> modulating rate, so if you captured all successive peaks, you could >> indeed state *a* modulating frequency, but you would not necessarily >> have all the necessary information about the signal - you have only >> one piece of information. A phase modulated signal has amplitude >> modulation components *and* frequency modulation components. >> >> As always, a complete view requires that the entire signal (not merely >> peak amplitudes) be viewed. >> >> That's not as clear as I would like; perhaps tomorrow after I've >> recovered from the week :) > > Pete, > > There is no variation of the peak amplitude at all with pure phase or > frequency modulation. > > JerryPhase modulators generate the familiar sidebands of AM for small modulating signals. That's something I've known (and worked with) for almost 40 years :) My comment was meant to show that the signal is impossible to characterise without proper analysis. Cheers PeteS
Reply by ●January 19, 20072007-01-19
PeteS wrote:> Jerry Avins wrote: >> PeteS wrote: >>> Jerry Avins wrote: >>>> dtsao wrote: >>>>> Oh, ok. I consider that the input signal is a sine wave, but where >>>>> it's >>>>> amplitude might not always be exactly the same (just off slightly >>>>> at each >>>>> peak from the previous). So there is a frequency here. >>>> >>>> It's not a pure sin if the peak amplitude changes. A Fourier >>>> transform will tell you what the additional frequencies are, but >>>> nothing about time. You are combining concepts here in an >>>> incompatible way. I haven't figured out yet how to separate them for >>>> you. >>>> >>>> Jerry >>> >>> A sine, even of [apparent] constant frequency, that changes it's >>> amplitude does indeed have frequency components apart from the >>> fundamental. Just what those are depend on *how* the sine varies. As >>> Jerry notes, a sine that varies in any way is no longer a pure sine. >>> >>> But merely reading the peak value does not tell one the *extra* >>> frequencies involved. >>> >>> Look at it this way: >>> >>> An amplitude modulated sine will have peak amplitudes that vary at >>> the modulating rate, so if you captured all successive peaks, you >>> could indeed state *a* modulating frequency, but you would not >>> necessarily have all the necessary information about the signal - you >>> have only one piece of information. A phase modulated signal has >>> amplitude modulation components *and* frequency modulation components. >>> >>> As always, a complete view requires that the entire signal (not >>> merely peak amplitudes) be viewed. >>> >>> That's not as clear as I would like; perhaps tomorrow after I've >>> recovered from the week :) >> >> Pete, >> >> There is no variation of the peak amplitude at all with pure phase or >> frequency modulation. >> >> Jerry > > Phase modulators generate the familiar sidebands of AM for small > modulating signals. That's something I've known (and worked with) for > almost 40 years :)What kind of phase modulators do you commonly use?> My comment was meant to show that the signal is impossible to > characterise without proper analysis.That's as true as a statement can be! :-) To the extent that the amplitude varies, the signal is distorted. An AM signal with shallow modulation generates small sidebands. If we sample the phasors every time the carrier advances one cycle, the two sideband phasors, each of the same amplitude, revolve around the tip of the carrier phasor; the upper sideband clockwise and the lower, counterclockwise. The sideband phasors both point in the direction at the same instant (yielding the largest resultant) and in the opposite direction, yielding the smallest. A little geometry shows that the resultant amplitude varies sinusoidally. Now leave the picture the same, except that the carrier phasor is rotated 90 degrees. The geometry for this case shows that the resultant moves sinusoidally along a line perpendicular to the unmodulated carrier phasor's "snapshot" position, effectively varying the phase. Strictly, the phase variation, being an arctangent, is not sinusoidal, nor is the amplitude, a secant, constant. For "small enough" sideband amplitudes, the phase variation is *nearly* sinusoidal and the amplitude is *nearly* constant. This happens when the approximation sin(theta) = theta = tan(theta) is adequate. Using approximations without confining their application to their range of validity leads to the sort of error we had here. Jerry -- Engineering is the art of making what you want from things you can get. ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
Reply by ●January 19, 20072007-01-19
Jerry Avins wrote:> PeteS wrote: >> Jerry Avins wrote: >>> PeteS wrote: >>>> Jerry Avins wrote: >>>>> dtsao wrote: >>>>>> Oh, ok. I consider that the input signal is a sine wave, but where >>>>>> it's >>>>>> amplitude might not always be exactly the same (just off slightly >>>>>> at each >>>>>> peak from the previous). So there is a frequency here. >>>>> >>>>> It's not a pure sin if the peak amplitude changes. A Fourier >>>>> transform will tell you what the additional frequencies are, but >>>>> nothing about time. You are combining concepts here in an >>>>> incompatible way. I haven't figured out yet how to separate them >>>>> for you. >>>>> >>>>> Jerry >>>> >>>> A sine, even of [apparent] constant frequency, that changes it's >>>> amplitude does indeed have frequency components apart from the >>>> fundamental. Just what those are depend on *how* the sine varies. As >>>> Jerry notes, a sine that varies in any way is no longer a pure sine. >>>> >>>> But merely reading the peak value does not tell one the *extra* >>>> frequencies involved. >>>> >>>> Look at it this way: >>>> >>>> An amplitude modulated sine will have peak amplitudes that vary at >>>> the modulating rate, so if you captured all successive peaks, you >>>> could indeed state *a* modulating frequency, but you would not >>>> necessarily have all the necessary information about the signal - >>>> you have only one piece of information. A phase modulated signal has >>>> amplitude modulation components *and* frequency modulation components. >>>> >>>> As always, a complete view requires that the entire signal (not >>>> merely peak amplitudes) be viewed. >>>> >>>> That's not as clear as I would like; perhaps tomorrow after I've >>>> recovered from the week :) >>> >>> Pete, >>> >>> There is no variation of the peak amplitude at all with pure phase or >>> frequency modulation. >>> >>> Jerry >> >> Phase modulators generate the familiar sidebands of AM for small >> modulating signals. That's something I've known (and worked with) for >> almost 40 years :) > > What kind of phase modulators do you commonly use? > >> My comment was meant to show that the signal is impossible to >> characterise without proper analysis. > > That's as true as a statement can be! :-) > > To the extent that the amplitude varies, the signal is distorted. An AM > signal with shallow modulation generates small sidebands. If we sample > the phasors every time the carrier advances one cycle, the two sideband > phasors, each of the same amplitude, revolve around the tip of the > carrier phasor; the upper sideband clockwise and the lower, > counterclockwise. The sideband phasors both point in the direction at > the same instant (yielding the largest resultant) and in the opposite > direction, yielding the smallest. A little geometry shows that the > resultant amplitude varies sinusoidally. > > Now leave the picture the same, except that the carrier phasor is > rotated 90 degrees. The geometry for this case shows that the resultant > moves sinusoidally along a line perpendicular to the unmodulated carrier > phasor's "snapshot" position, effectively varying the phase. Strictly, > the phase variation, being an arctangent, is not sinusoidal, nor is the > amplitude, a secant, constant. For "small enough" sideband amplitudes, > the phase variation is *nearly* sinusoidal and the amplitude is *nearly* > constant. This happens when the approximation sin(theta) = theta = > tan(theta) is adequate. > > Using approximations without confining their application to their range > of validity leads to the sort of error we had here. > > JerryMuch depends on the phase modulator as to the level of the sidebands. A precision [as precise as you want] modulator will still generate sidebands though. It's late [and I have an incipient cold] and I don't want to pressure the mathematical side of my brain right now, but I'll happily continue this tomorrow :) Cheers PeteS
Reply by ●January 21, 20072007-01-21
"Jerry Avins" <jya@ieee.org> wrote in message news:RfCdnSKBgLTbrizYnZ2dnUVZ_t_inZ2d@rcn.net...> PeteS wrote: > > Jerry Avins wrote: > >> dtsao wrote: > >>> Oh, ok. I consider that the input signal is a sine wave, but whereit's> >>> amplitude might not always be exactly the same (just off slightly at > >>> each > >>> peak from the previous). So there is a frequency here. > >> > >> It's not a pure sin if the peak amplitude changes. A Fourier transform > >> will tell you what the additional frequencies are, but nothing about > >> time. You are combining concepts here in an incompatible way. I > >> haven't figured out yet how to separate them for you. > >> > >> Jerry > > > > A sine, even of [apparent] constant frequency, that changes it's > > amplitude does indeed have frequency components apart from the > > fundamental. Just what those are depend on *how* the sine varies. As > > Jerry notes, a sine that varies in any way is no longer a pure sine. > > > > But merely reading the peak value does not tell one the *extra* > > frequencies involved. > > > > Look at it this way: > > > > An amplitude modulated sine will have peak amplitudes that vary at the > > modulating rate, so if you captured all successive peaks, you could > > indeed state *a* modulating frequency, but you would not necessarily > > have all the necessary information about the signal - you have only one > > piece of information. A phase modulated signal has amplitude modulation > > components *and* frequency modulation components. > > > > As always, a complete view requires that the entire signal (not merely > > peak amplitudes) be viewed. > > > > That's not as clear as I would like; perhaps tomorrow after I've > > recovered from the week :) > > Pete, > > There is no variation of the peak amplitude at all with pure phase or > frequency modulation. > >Yes but there is when there are two mixed FM signals ie co-channel interference. F. -- Posted via a free Usenet account from http://www.teranews.com
