cAmdf

Description

This component computes the Average Magnitude Difference Function (AMDF) for each input frame. Various methods for padding or warping at the border exist.

Type hierarchy

cDataProcessorcVectorProcessorcAmdf

Fields

  • nameAppend (string) [default: amdf]

    A string suffix to append to the input field names (default: empty)

  • nLag (numeric) [default: 0.0]

    If nLag > 0, compute AMDF up to maximum lag 'nLag' (or maximum possible lag limited by framesize). If nLag=0, then always the maximum framesize will be used (this is the default). If nLag < 0, then nLag=framesize/((-1)*nLag) will be used.

  • method (string) [default: limit]

    The AMDF computation method (i.e. the handling of border conditions):

    • 'limit' : compute AMDF only in regions where the shifted windows overlap

    • 'warp' : compute cyclical AMDF by warping of input

    • 'zeropad' : zero pad missing samples

  • invert (numeric) [default: 0.0]

    1 = invert the AMDF output values (literally '1-amdf'), i.e. so that the behaviour of the AMDF output corresponds more to that of an autocorrelation function.