autocorrelate suspiciously never has deviants

[rcrath]

[ruohoruotsi]

Interesting, what do you mean by deviants? Do you have an example autocorrelation analysis?

[rcrath]

Late to the party. There are tolerances set so that any wavecycles that fall outside of the tolerances -- I.e. the wave cycle is too long or to short - can optionally be excluded from the construction of the wavetable. These are marked deviant, in my best 19th century social science language.

[ruohoruotsi]

Well autocorrelation has 2 parts, (1) the multiply & sum, slide signals ... to generate the new signal (2) pick peaking, to find the highest peak, and distances between peaks is the wave cycle length ... so these two processes are numerical and are not EXACT (i.e. wave cycle drift) so if you never have "deviants" ...there must be something else going on.

I'm sure we can get to the bottom of this in a quick huddle and a few examples! 🚀

[rcrath]

I'm clawing my way back to the wavetabler. It needs a good going over as there are problems with the z-cross method too. Can I check back in with you next week sometime? Thnx for offering.

[ruohoruotsi]

Sure thing sir, I'm around!! Hit me up on Skool for quicker response and we can hope on a call to huddle on details!