Similar to Calculating the PDF of a waveform from its samples, but for periodic sampled signals, and I just want the peak. Is there a way to calculate the highest inter-sample peak of the reconstructed signal?

For instance, the signal 0, 0, 1, 1, 0, 0 is symmetrical, so you know the intersample peak is between samples 2 and 3, at 2.5, and can be calculated as ... + sinc(-0.5) + sinc(0.5) + sinc(5.5) + sinc(6.5) + ... ≈ 1.24

But is there a way to do this for the general case, when you don't know where the peak occurs?

This page says the MLS has the lowest possible crest factor, but I think for analog use, this isn't actually true, since MLS has strong intersample peaks, >6 dB higher than the samples, and a sweep-like signal like the Newman method has better crest factor.

enter image description here

  • $\begingroup$ it's a good point (about the crest factor), if you're considering crest factor from the output of the bandlimited D/A converter. in writing that, i was considering only the peak and r.m.s. values of the samples themselves, not the reconstructed waveform. $\endgroup$ – robert bristow-johnson Apr 30 '14 at 3:20

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.