acorr#

class diffsptk.AutocorrelationAnalysis(acr_order, frame_length, norm=False, acf='none')[source]#

See this page for details. Currently, spectrum input is not supported.

Parameters:
acr_orderint >= 0 [scalar]

Order of autocorrelation, \(M\).

frame_lengthint > M [scalar]

Frame length, \(L\).

normbool [scalar]

If True, normalize autocorrelation.

acf[‘none’, ‘biased’, ‘unbiased’]

Type of autocorrelation function.

forward(x)[source]#

Estimate autocorrelation of input.

Parameters:
xTensor [shape=(…, L)]

Framed waveform.

Returns:
rTensor [shape=(…, M+1)]

Autocorrelation.

Examples

>>> x = diffsptk.ramp(4)
>>> acorr = diffsptk.AutocorrelationAnalysis(3, 5)
>>> r = acorr(x)
>>> r
tensor([30.0000, 20.0000, 11.0000,  4.0000])

See also

frame levdur lpc