I have a long sequence of time values (instants in time where events happened) and I would like to detect perodicities in this data. It's possible to sample these points onto a gigantic grid and take an FFT, but the times are known very accurately relative to the grid spacing available even for a 10M+ point FFT. So doing this entails a bunch of interpolating onto a coarse grid.

Are there other techniques to extract spectral/periodicity information?

I thought Lomb-Scargle could be used, but once I dug into it I realized that LS is geared toward getting a regular spectrum of unevenly sampled data. I have no amplitude data, just times, so if I put all ones for the amplitudes, it just gives me a flat spectrum back. I want the underlying assumption to be that the time signal is all zero except for impulses at the specified times.



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.