plotPSD: Plot Power Spectral Density

Description

Plots the distribution of summarised sound levels across frequency, either as lines of quantile levels or a heatmap showing the full distribution. Multiple PSD sources can be combined and plotted as long as they have identical frequency levels.

Usage

plotPSD(
  x,
  style = c("quantile", "density"),
  scale = c("log", "linear"),
  q = 0.5,
  color = "black",
  freqRange = NULL,
  dbRange = NULL,
  dbInt = 1,
  densityRange = NULL,
  units = "dB re: 1uPa^2/Hz",
  cmap = viridis_pal()(25),
  by = NULL,
  referenceLevel = NULL,
  facet = NULL,
  ncol = NULL,
  title = NULL,
  returnData = FALSE,
  progress = TRUE
)
prepPSDData(
  x,
  freqRange = NULL,
  style = c("density", "quantile"),
  by = NULL,
  dbInt = 1,
  compression = 10000,
  progress = TRUE
)

Value

a ggplot object for plotPSD, see details for prepPSDData

Arguments

x: a dataframe or list of dataframes, or file path or vector of file paths, or the output from prepPSDData
style: character specifying plot style to create, either "quantile", "density", or a vector with both
scale: scale to use for frequency axis, one of "log" or "linear"
q: quantile to plot
color: color for quantile
freqRange: range of frequencies to plot
dbRange: range of dB values to plot
dbInt: bin interval size for density plot
densityRange: optional range of values for density color scale
units: units for dB axis of plot
cmap: color map to use for density plot
by: optional column to plot different quantile lines by, only affects style='quantile'. If x is a data.frame, by can also be one of 'hour', 'month', or 'year' and that column will be created automatically if not present.
referenceLevel: only used together with by. A value of the by column to use as a reference for all other levels. The plot will then show the difference between the other levels and the reference
facet: optional column to facet the plots by
ncol: number of columns to use when plotting with facet
title: optional title for plot
returnData: if TRUE then no plot will be generated, instead the dataframe that would normally be used to make the plot will be returned
progress: logical flag to show progress bar
compression: compression factor for tdigest, lower values are less accurate but will compute faster. Only relevant for style='quantile' when loading and combining multiple datasets

Author

Taiki Sakai taiki.sakai@noaa.gov

Details

prepPSDData is called by the plotting code, and does not necessarily need to be called separately from plotPSD. Loading PSD data can be time consuming, so it may be useful to load the data first, then it is easier to spend time adjusting plot settings.

The output of prepPSDData is a list with 5 elements:

frequency: - the frequency values of the input data
freqRange: - the value of the "freqRange" parameter if it was supplied
dbVals: - the dB values of breakpoints used for "density" plotting
quantileData: - the data used for quantile plots. These are stored as "tidgest" objects serialized using as.list.tdigest, from which quantiles can be computed
densityData: - the data used fro quantile plots. These are stored as a matrix of bin counts - each column corresponds to the "frequency" output, each row corresponds to bins defined using "dbVals" as boundaries

Examples

Run this code


psd <- loadSoundscapeData(system.file('extdata/PSDSmall.csv', package='PAMscapes'))
# Plotting only first 1000 columns for brevity
plotPSD(psd[1:1000], style='density')
plotPSD(psd[1:1000], style='quantile', q=.05)

Run the code above in your browser using DataLab