soundgen (version 1.7.0)

medianSmoother: Median smoothing

Description

Internal soundgen function.

Usage

medianSmoother(df, smoothing_ww, smoothingThres, inviolable = NULL)

Arguments

df

dataframe (each column is processed separately, so multiple contours can be fed into this function at once to speed things up)

smoothing_ww

width of smoothing window (points)

smoothingThres

tolerated deviance from moving median (semitones)

inviolable

a vector of indices of the rows rows of df that should not be modified (meant for manual pitch values)

Value

Returns a dataframe of the same dimensions as df.

Details

Internal helper function for smoothing pitch contours or other contours. Only outliers are modified, so it's not like smoothing with a kernel. NB: the expected input is pitch, so deviance is calculated on a log-scale.

Examples

# NOT RUN {
df = data.frame(a = rnorm(20, mean = 100, sd = 20),
                b = rnorm(20, mean = 100, sd = 10))
df1 = soundgen:::medianSmoother(df, smoothing_ww = 5,
      smoothingThres = 1,
      inviolable = c(rep(TRUE, 10), rep(FALSE, 10)))
plot(df[, 2], type='b')
lines(df1[, 2], type='b', col='blue', pch=3)
# }