I know there are spectrum analysis programs that have a zoom function, ie. you can limit the graphical display to show only a small frequency range within the sampled bandwidth. What I am looking for, however, is a program that will only FFT that small frequency range (user specified), instead of the entire sampled bandwidth, so resolution within it can be maximized. Has anyone seen anything like this? Charles
Windowing spec. analyzer?
Started by ●July 13, 2004
Reply by ●July 13, 20042004-07-13
"Charles davis" <charlesdavis134@aol.com> wrote in message news:57n8f01t66r25cmit958guvh502ap3q40t@4ax.com...> I know there are spectrum analysis programs that have a zoom function, > ie. you can limit the graphical display to show only a small frequency > range within the sampled bandwidth. > > What I am looking for, however, is a program that will only FFT that > small frequency range (user specified), instead of the entire sampled > bandwidth, so resolution within it can be maximized. > > Has anyone seen anything like this?You'll find this in various instruments (spectrum/signal analyzers). However, I'm not aware of 'programs' that do this (by program I assume you mean SW that runs on a desktop PC and commonly available as shareware/freeware). Try searching for 'zoom FFT' - it will accomplish what you are looking for, but I'm not sure if it is available in canned programs. If you are writing your own program, then the zoom-FFT description will certainly help. Cheers Bhaskar> > Charles
Reply by ●July 13, 20042004-07-13
Charles davis wrote:> I know there are spectrum analysis programs that have a zoom function, > ie. you can limit the graphical display to show only a small frequency > range within the sampled bandwidth. > > What I am looking for, however, is a program that will only FFT that > small frequency range (user specified), instead of the entire sampled > bandwidth, so resolution within it can be maximized. > > Has anyone seen anything like this? > > CharlesThe closer together your FFTed points are, the longer a duration of samples you need. To resolve frequencies n Hz apart, you need a duration on the order of 1/n second. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●July 14, 20042004-07-14
Charles davis <charlesdavis134@aol.com> wrote in message news:<57n8f01t66r25cmit958guvh502ap3q40t@4ax.com>...> I know there are spectrum analysis programs that have a zoom function, > ie. you can limit the graphical display to show only a small frequency > range within the sampled bandwidth. > > What I am looking for, however, is a program that will only FFT that > small frequency range (user specified), instead of the entire sampled > bandwidth, so resolution within it can be maximized.I think you can achieve the same effect by zero-padding the window to some large length, FFT the resulting signal and then use the zoom. It will be slightly slower that the method you suggest, but is way easier to implement.> Has anyone seen anything like this?Nope. I use my own routines I've made with matlab.> CharlesRune
Reply by ●July 14, 20042004-07-14
If processing time isn't a concern just directly transform it. Then you can look at magnitudes of frequencies in between the Fourier series and pick your own window size. Though resolution bandwidth is still depicted by the sample rate/samples processed. It's easy to write, it's like ten or so lines of code. "Charles davis" <charlesdavis134@aol.com> wrote in message news:57n8f01t66r25cmit958guvh502ap3q40t@4ax.com...> I know there are spectrum analysis programs that have a zoom function, > ie. you can limit the graphical display to show only a small frequency > range within the sampled bandwidth. > > What I am looking for, however, is a program that will only FFT that > small frequency range (user specified), instead of the entire sampled > bandwidth, so resolution within it can be maximized. > > Has anyone seen anything like this? > > Charles