Autocorrelation with cycle

[Lars H]

Hi Simon,

in the recordings you mention that outocorrelation may be problematic because it “fades out” for higher values of τ. We are trying to find a value of τ which represents a good cycle, so can’t we just cycle around when we reach the end of the window? Obviously the window length may not be a multiple of τ, but at least we remove some of the fade-out. I upload a drawing to illustrate.

Attachments:

You must be logged in to view attached files.

[Simon]

OK, I get this idea. You are proposing to ‘loop’ the contents of the window, to effectively create a longer single.

I think the problem with this will be that when we cycle around, we will create a a discontinuity in the waveform (because we don’t loop in perfect multiples of the pitch period: the window is not generally aligned to the pitch period).

[Author]

Posts