Internal soundgen function.
pathfinder(
pitchCands,
pitchCert,
pitchSource,
manual = NULL,
certWeight = 0.5,
pathfinding = c("none", "fast", "slow")[2],
annealPars = list(maxit = 5000, temp = 1000),
interpolWin_bin = 3,
interpolTol = 0.05,
interpolCert = 0.3,
manualCert = 1,
snakeStep = 0.05,
snakePlot = FALSE
)a matrix of multiple pitch candidates per fft frame. Each column is one fft frame, each row is one candidate (the last row is always "manual")
a matrix of the same dimensionality as pitchCands specifying our certainty in pitch candidates
(0 to 1) in pitch postprocessing, specifies how much we prioritize the certainty of pitch candidates vs. pitch jumps / the internal tension of the resulting pitch curve
method of finding the optimal path through pitch
candidates: 'none' = best candidate per frame, 'fast' = simple heuristic,
'slow' = annealing. See soundgen:::pathfinder
a list of control parameters for postprocessing of
pitch contour with SANN algorithm of optim. This is
only relevant if pathfinding = 'slow'
when interpolating pitch candidates, the median is
calculated over plus-minus interpolWin_bin
when interpolating pitch candidates, the criterion
for needing to interpolate is the absence of pitch candidates with values
within 1 plus-minus interpolTol of the median of pitch center of
gravity over the interpolation window. For ex., if interpolTol
is .05, we look for values from 0.95 to 1.05 time the median value over
interpolation window.
when interpolating pitch candidates, all generated pitch
candidates are assigned a certainty equal to interpolCert
the certainty in manually added pitch candidates
optimized path through pitch candidates is further
processed to minimize the elastic force acting on pitch contour. To
disable, set snakeStep = 0
if TRUE, plots the snake
Returns a numeric vector of pitch values representing the best found path through pitch candidates.
Internal helper function for postprocessing pitch contour. Starts with a reasonable guess and computes the more-or-less optimal pitch contour (not quite the very optimal - too computationally expensive).