hi folks, Is there any literature out there analyzing the effects of running simple peak detection algorithms on different classes of signals? I've found peak detection to be a very useful operation in my DSP work, and a web search returns many links to algorithms that perform this operation, but (perhaps due to its nonlinearity) I haven't been able to find any papers describing the mathematical properties of the peak detection operation. I understand that peak detection is a bit of a black art and no general results may exist, but I'd appreciate any pointers to even highly simplified cases. Thanks, Sina
peak detection literature
Started by ●July 28, 2008
Reply by ●July 28, 20082008-07-28
On Jul 28, 7:13�pm, sina.tootoon...@gmail.com wrote:> hi folks, > > Is there any literature out there analyzing the effects of running > simple peak detection algorithms on different classes of signals? I've > found peak detection to be a very useful operation in my DSP work, and > a web search returns many links to algorithms that perform this > operation, but (perhaps due to its nonlinearity) I haven't been able > to find any papers describing the mathematical properties of the peak > detection operation. I understand that peak detection is a bit of a > black art and no general results may exist, but I'd appreciate any > pointers to even highly simplified cases. > > Thanks, > > SinaSorry to ask something stupid, but what do you hope to accomplish with "peak detection"? Are you trying to detect the signal, estimate its delay?
Reply by ●July 29, 20082008-07-29
On Jul 28, 8:32�pm, julius <juli...@gmail.com> wrote:> On Jul 28, 7:13�pm, sina.tootoon...@gmail.com wrote: > > > hi folks, > > > Is there any literature out there analyzing the effects of running > > simple peak detection algorithms on different classes of signals? I've > > found peak detection to be a very useful operation in my DSP work, and > > a web search returns many links to algorithms that perform this > > operation, but (perhaps due to its nonlinearity) I haven't been able > > to find any papers describing the mathematical properties of the peak > > detection operation. I understand that peak detection is a bit of a > > black art and no general results may exist, but I'd appreciate any > > pointers to even highly simplified cases. > > > Thanks, > > > Sina > > Sorry to ask something stupid, but what do you hope to accomplish > with "peak detection"? �Are you trying to detect the signal, estimate > its delay?Not to derail, but lets say I'm trying to detect a coded signal. Say a gold code arbitrarily, but really I'm more interested in general coding theory. What would be a good free, online resource to consult?
Reply by ●July 29, 20082008-07-29
<sina.tootoonian@gmail.com> wrote in message news:e1903747-0b89-42aa-b856-1adf0c82e413@r15g2000prh.googlegroups.com...> hi folks, > > Is there any literature out there analyzing the effects of running > simple peak detection algorithms on different classes of signals? I've > found peak detection to be a very useful operation in my DSP work, and > a web search returns many links to algorithms that perform this > operation, but (perhaps due to its nonlinearity) I haven't been able > to find any papers describing the mathematical properties of the peak > detection operation. I understand that peak detection is a bit of a > black art and no general results may exist, but I'd appreciate any > pointers to even highly simplified cases.Peak detector is the one of the classic items in the radio location theory. Just about any textbook on the radio location has a chapter about it. Vladimir Vassilevsky DSP and Mixed Signal Consultant www.abvolt.com
Reply by ●July 29, 20082008-07-29
On Jul 28, 10:38 pm, wellwatch <wellwa...@gmail.com> wrote:> > Not to derail, but lets say I'm trying to detect a coded signal. Say > a gold code arbitrarily, but really I'm more interested in general > coding theory. What would be a good free, online resource to consult?There's a big literature on "detection theory" and "estimation theory". To simplify a bit, in an AWGN channel, the optimal detector is a cross-correlator followed by magnitude (if there's phase ambiguity) and threshold. How to set the threshold is determined by the signal in question, and the noise level. Not sure about online, but try a search using the keywords above.
Reply by ●July 29, 20082008-07-29
On Jul 28, 5:32 pm, julius <juli...@gmail.com> wrote:> On Jul 28, 7:13 pm, sina.tootoon...@gmail.com wrote: > > > hi folks, > > > Is there any literature out there analyzing the effects of running > > simple peak detection algorithms on different classes of signals? I've > > found peak detection to be a very useful operation in my DSP work, and > > a web search returns many links to algorithms that perform this > > operation, but (perhaps due to its nonlinearity) I haven't been able > > to find any papers describing the mathematical properties of the peak > > detection operation. I understand that peak detection is a bit of a > > black art and no general results may exist, but I'd appreciate any > > pointers to even highly simplified cases. > > > Thanks, > > > Sina > > Sorry to ask something stupid, but what do you hope to accomplish > with "peak detection"? Are you trying to detect the signal, estimate > its delay?Hi Julius, I didn't have a particular application in mind. I've used peak detection for different reasons in the past e.g. to distinguish different signal categories in a recording, and have found it to work quite well. However since I haven't come across any general characterizations of the PD operation, I've always felt a little uneasy using it since I can't say anything more rigorous than 'it seems to work well on my test signals' to justify my using it. - Sina
Reply by ●July 29, 20082008-07-29
On Jul 29, 3:01 am, "Vladimir Vassilevsky" <antispam_bo...@hotmail.com> wrote:> <sina.tootoon...@gmail.com> wrote in message > > news:e1903747-0b89-42aa-b856-1adf0c82e413@r15g2000prh.googlegroups.com... > > > hi folks, > > > Is there any literature out there analyzing the effects of running > > simple peak detection algorithms on different classes of signals? I've > > found peak detection to be a very useful operation in my DSP work, and > > a web search returns many links to algorithms that perform this > > operation, but (perhaps due to its nonlinearity) I haven't been able > > to find any papers describing the mathematical properties of the peak > > detection operation. I understand that peak detection is a bit of a > > black art and no general results may exist, but I'd appreciate any > > pointers to even highly simplified cases. > > Peak detector is the one of the classic items in the radio location theory. > Just about any textbook on the radio location has a chapter about it. > > Vladimir Vassilevsky > DSP and Mixed Signal Consultantwww.abvolt.comThanks for the pointer Vladimir, I'll take a look. - Sina
Reply by ●July 29, 20082008-07-29
On Jul 29, 9:49 am, sina.tootoon...@gmail.com wrote:> Hi Julius, > > I didn't have a particular application in mind. I've used peak > detection for different reasons in the past e.g. to distinguish > different signal categories in a recording, and have found it to work > quite well. However since I haven't come across any general > characterizations of the PD operation, I've always felt a little > uneasy using it since I can't say anything more rigorous than 'it > seems to work well on my test signals' to justify my using it. > > - SinaAre you saying that you are using peak detection to discriminate between different audio signals? Well, the effectiveness depends first of all on what classifications you want to make. If I were you, I'd forget all about peak detection, think about how you would describe the signal classes you are trying to discriminate, and then see what the best method is. Sounds like you have a hammer you have been using, and you are not sure what a nail looks like :-). Maybe you should now look at how a nail really looks like, and then see if a hammer is the correct tool for it. Cheers, Julius
Reply by ●July 29, 20082008-07-29
wellwatch wrote:> On Jul 28, 8:32 pm, julius <juli...@gmail.com> wrote: >> On Jul 28, 7:13 pm, sina.tootoon...@gmail.com wrote: >> >>> hi folks, >>> Is there any literature out there analyzing the effects of running >>> simple peak detection algorithms on different classes of signals? I've >>> found peak detection to be a very useful operation in my DSP work, and >>> a web search returns many links to algorithms that perform this >>> operation, but (perhaps due to its nonlinearity) I haven't been able >>> to find any papers describing the mathematical properties of the peak >>> detection operation. I understand that peak detection is a bit of a >>> black art and no general results may exist, but I'd appreciate any >>> pointers to even highly simplified cases. >>> Thanks, >>> Sina >> Sorry to ask something stupid, but what do you hope to accomplish >> with "peak detection"? Are you trying to detect the signal, estimate >> its delay? > > Not to derail, but lets say I'm trying to detect a coded signal. Say > a gold code arbitrarily, but really I'm more interested in general > coding theory. What would be a good free, online resource to consult?Well, you are derailing the thread. Why not start a new one? Do you need to learn how? Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●July 29, 20082008-07-29
sina.tootoonian@gmail.com wrote:> On Jul 28, 5:32 pm, julius <juli...@gmail.com> wrote: >> On Jul 28, 7:13 pm, sina.tootoon...@gmail.com wrote: >> >>> hi folks, >>> Is there any literature out there analyzing the effects of running >>> simple peak detection algorithms on different classes of signals? I've >>> found peak detection to be a very useful operation in my DSP work, and >>> a web search returns many links to algorithms that perform this >>> operation, but (perhaps due to its nonlinearity) I haven't been able >>> to find any papers describing the mathematical properties of the peak >>> detection operation. I understand that peak detection is a bit of a >>> black art and no general results may exist, but I'd appreciate any >>> pointers to even highly simplified cases. >>> Thanks, >>> Sina >> Sorry to ask something stupid, but what do you hope to accomplish >> with "peak detection"? Are you trying to detect the signal, estimate >> its delay? > > Hi Julius, > > I didn't have a particular application in mind. I've used peak > detection for different reasons in the past e.g. to distinguish > different signal categories in a recording, and have found it to work > quite well. However since I haven't come across any general > characterizations of the PD operation, I've always felt a little > uneasy using it since I can't say anything more rigorous than 'it > seems to work well on my test signals' to justify my using it.A peak detector is a really bad way to look at a digital signal that is not heavily oversampled. With normal sampling, the peak is likely to fall between sample instants. In any case, you need some sort of average to compare with if you want to use it for characterizing a signal. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
