warbleR 1.1.36

Changes and additions:

• New function vocal_activity() estimates vocal activity across sound files based on selections.

• Function query_xc() now fails gracefully when the Xeno-Canto API is not available.

warbleR 1.1.35

Changes and additions:

• New function waveform_similarity() to calculate the similarity between waveforms
• Fix typos in vignette

warbleR 1.1.34

Changes and additions:

• Update requested by CRAN
• Remove querxc() examples from vignette

warbleR 1.1.33

Changes and additions:

• Added ‘peakt’ feature to spectro_analysis() to measure the time at which peak amplitude occurs
• Add message when URL resource not available in read_sound_files()

warbleR 1.1.32

Changes and additions:

• Updated system for letting users know the number of process in a routine have been upgraded

warbleR 1.1.31

Changes and additions:

• function freq_DTW() now parallelizes dynamic time warping computing
• new vignette ‘Annotation data format’
• confirm.extended argument in selection_table() has been deprecated.
• song_param() alternative name for song_analysis() has been deprecated.
• alternative names of must functions have been deprecated

warbleR 1.1.30

Changes and additions:

• gaps() now turns negative numbers (from overlapping annotations) into 0’s
• error message added to simulate_songs() if negative numbers in ‘gaps’ are found

Bug fixes:

• quer_xc() bug for searching on windows and macOS
• spectro_analysis() bug when ‘harmonicity = TRUE’ due to changes in soundgen::analyze() output names
• print.selection_table() now checks for possible errors in check_res() metadata
• bug fixed in multi_DTW() when measuring contours other than dominant frequency

warbleR 1.1.29

Changes and additions:

• full_spectrograms() new argument ‘only.annotated’ to only print pages that contain annotations.
• New method (‘time.closest’) in test_coordination() which estimate coordination between more than 2 individuals
• Default value of ‘confirm.extended’ argument in selection_table() set to FALSE. Will be deprecated in future versions.

warbleR 1.1.28

Changes and additions:

• Warning messages in check_sels() and check_sound_files() when having sound files/wave objects with different sampling rates
• Fix bug in file size estimation when ‘whole.recs = TRUE’ in selection_table()
• Box and axis can be removed in phylo_spectro()

New functions:

• by_element_est() convert a by-song extended selection table to by-element

warbleR 1.1.27

New functions:

• sound_pressure_level() to measure, well, sound pressure level
• find_clipping() to detect clipped selections
• wav_2_flac() to convert ‘.wav’ files into flac compressed lossless format
• gaps() for calculating time intervals (i.e. gaps) between selections
• read_sound_file() is now use to read sound files
• split_sound_files() now can also split associated selection tables

Changes and additions:

• read_sound_file() and consequently most other functions in warbleR will look at the ‘channel’ column for deciding which channel to read
• updated download URL in quer_xc()
• selection_table and extended_selection_table subsetting ([) drop behavior switched to FALSE
• overlapping_sels() has been updated to improve performance
• unique.labs argument in overlapping_sels() has been deprecated
• fix.selec argument added to check_sels() to fix ‘selec’ column labels
• warbleR now supports ‘.wav’, ‘.flac’, ‘.wac’ and ‘.mp3’ formats
• New argument ‘downsample’ to speed up processing in remove_silence()
• New printing methods for ‘autodetec.output’, ‘find.peaks.output’ and ‘xcorr.output’ object classes
• ‘templates’ and ‘surveys’ arguments added to cross_correlation() to simplify signal detection with this function
• auto_detec() output including envelopes is ~1/3 smaller to reduce memory usage

Bug fixes:

• error when resampling wavs in resample_est_waves()
• error when silences are not found in remove_silence()

warbleR 1.1.26

Changes and additions:

• df_DTW() and ff_DTW() were merged in a new function freq_DTW()
• sp.en.ts(), dfts() and ffts() were merged in a new function freq_ts()
• sel_tailor() can undo deletions
• button labels replaced by symbols in sel_tailor()
• many functions renamed (see new_function_names)

warbleR 1.1.25

Bug fixes:

• Negative gaps values in song_param()

New functions:

• optimize_autodetec() for tuning detection parameters in autodetec()
• envelope() a C-compiled version of seewave::env()
• new function find_peaks() to detect signals on sound files using xcorr() output
• read_sound_file() to read wav, wav and mp3 files as well as wave objects in extended selection tables

Changes and additions:

• Remove bioacoustics::resample() as it was deprecated (mp32wav() and fixwavs() affected)
• Remove comment columns in lbh_selec_table
• Selection table printing method uses huxtable package for text coloring
• ‘suffix’ argument added to lspec()
• autodetec() does not create image files with spectrograms any more
• new arguments added to autodetec(): ‘hold.time’ for merging selections, ‘thinning’ for decreasing envelope size
• error when having “songs” across multiple files in song_param()

warbleR 1.1.24

Changes and additions:

• lspec() allow to set resolution and custom dimensions of output image ifles
• split_wavs() can now save selections in a data frame of selection table as sound files
• read_wave() can now read .wav and .mp3 files from URLs
• ‘strength’ argument to control color in color_spectro()
• ‘method’ option to create spectrograms on the fly in xcorr()
• ‘img’ argument in autodetec() is being soft-deprecated
• lspec() now takes output from xcorr(), find_peaks() and autodetec() and plots cross-correlation scores or envelopes along with spectrograms
• ‘download’ option available in find_annotations()
• ‘leaflet’ interactive maps in xc_maps()
• ‘split_wavs()’ new argument ‘only.sels’ to only return selection tables
• ‘color_spectro()’ frequency range is not restricted by the selection table top and bottom frequency

Bug fixes:

• Negative gaps values in song_param()
• Trivial bug when downloading files in quer_xc()
• Fixed number of segments when using ‘sgmts’ argument in ‘split_wavs()’, it used to return ‘sgmts - 1’ segments.
• Fix mismatch in specreator() spectrograms and oscillograms
• Fix bug when no detections in x_corr()

warbleR 1.1.23

Changes and additions:

• add ‘am.amps’ argument to add amplitude modulation in sim_songs()

Bug fixes:

• fix ‘spectrograms()’ when plotting extended selection tables ‘by song’
• fix ‘compare_methods()’ when less than 100 rows in ‘X’

warbleR 1.1.22

Bug fixes:

• fix rbind in extended selection tables

warbleR 1.1.21

Changes and additions:

• print information of extended selection table includes sampling rates, bit depths and warbleR version
• warning message in ‘rbind.extended_selection_table()’ when some wave object names are found on both extended selection tables

warbleR 1.1.20

Changes and additions:

• mfcc cross-correlation implemented in ‘xcorr()’ (argument ‘type’)
• argument ‘cor.mat’ in ‘xcorr()’ has been deprecated
• new argument ‘output’ in ‘xcorr()’
• new example data ‘comp_matrix’ and ‘lbh_selec_table2’ for using on ‘find_peaks()’
• ‘xcorr()’ can run cross-correlation over entire sound files
• ‘seltailor()’ can take now extended selection tables

New functions:

• new function ‘image_to_wave()’ to convert images into wave objects
• new function ‘find_peaks()’ to detect signals on sound files using ‘xcorr()’ output

warbleR 1.1.19

Changes and additions:

• remove duplicated sound file names in spectrogram() images when ‘by.song’
• replace all class(.) == , by is(., *)
• new argument ‘fill’ to control the fill color of boxes in ‘specreator()’

New functions:

• new function rename_waves_est() to rename wave objects in extended selection tables

Bug fixes:

• box/lines plotting when no frequency columns are supplied

warbleR 1.1.18

Changes and additions:

• updates on specan() to improve harmonic detection
• xcorr() performance has been largely improved (5X faster)
• ‘compare.matrix’ argument in xcorr() to let user defined the pairwise comparisons to be conducted
• function xcorr.graph() has been deprecated (use corrplot package instead)
• arguments ‘dens’ and ‘bp’ in xcorr() have been deprecated
• jpeg and maps packages are not strong dependencies anymore

warbleR 1.1.17

Changes and additions:

• arguments ‘custom1’ and ‘custom2’ in autodetec() to let users enter their own data
• new vignette ‘Introduction to warbleR’

warbleR 1.1.16

New functions:

• sort_colms: moves required colms to left side of data frame. Originally found in Rraven package
• split_wavs: splits sound files in shorter segments

Changes and additions:

• ‘to’ replaced by ‘dest.path’ argument removed in mp32wav()
• read_wave() can also read regular .wav files
• quer_ml() temporarily removed while ebird-macaulay API is updated
• fix_wavs() now can use bioacoustics package or SOX for resampling
• global options (from warbleR_options()) are now taken by wav_info()
• ‘coor.test()’ now returns 0 for coordination score when the ratio is undefined
• ‘dB.threshold’ argument in ‘frange()’ and ‘frange_detec()’ for detecting frequency range based on a dB scale
• ‘min_indx’ and ‘max_indx’ argument in ‘song_param()’ renamed as ‘min_colm’ and ‘max_colm’ and now take names instead of indexes
• more accurate dominant frequency detection in ‘dfts()’, ‘trackfreqs()’ and ‘specan()’
• ‘elm_colm’ and ‘elm_fun’ argument to calculate element level parameters
• ‘unq.elemts’ and ‘mean.elemt.count’ parameters added to ‘song_param()’ output
• more error messages in check_sels()
• more info about file input in ‘auto_detec()’ documentation
• add ‘frange’ for bp in ‘dfts()’
• 2 new methods available in ‘compare.methods()’
• more error messages in check_sels()
• more info about file input in ‘auto_detec()’ documentation
• ‘frange’ available as bandpass in ‘dfts()’ and ‘mfcc_stats()’
• ‘bioacoustics’ dependency changed to ‘suggested package’
• ‘mfcc_stats()’ return NAs instead of error when MFCC can’t be calculated

Bug fixes:

• x_corr() on extended selection tables with extra column names generated and error
• system() commands for windows in ‘fix_wavs()’ and ‘resample_est()’
• normalization when downsampling in ‘mp32wav()’ was required
• error when creating extended selection tables ‘by song’ and ‘song’ column tags are repeated across sound files (selection_table())
  

warbleR 1.1.15

New functions:

• resample_est: resample wave objects in a extended selection table
• phylo_spectro: add spectrograms to the tips of a phylogeny
• wav_info: wrapper for ‘selection_table()’ that returns wave file information
• multi_DTW: wrapper on dtw for comparing signals described by multiple contours
• rm_channels: remove channels from wave files
• melfcc_stats: calculate descriptive statistics on Mel-frequency cepstral coefficients

Changes and additions:

• ‘normalize’ argument in ‘mp32wav()’ renamed as ‘dip.depth’
• ‘frange’ and ‘dfrange’ arguments in ‘xcorr()’ function deprecated
• ’selection_table` now can take files found in different directories and create extended selection tables
• new parameters measured by ‘specan()’
• wav.size.MB column names is now wav.size in ‘check_sels()’ output
• ‘title.labels’ argument in ‘specreator()’ to customized title

Bug fixes:

• Using ‘wav.path’ from ‘warbleR_options’ in ‘catalog()’
• clipping.edges option in ‘dfts()’ and ‘ffts()’

warbleR 1.1.14

Changes and additions:

• ‘open_wd()’ example fixed (requested by CRAN)

warbleR 1.1.13

Changes and additions:

• new function names (print(new_function_names))
• object class ‘extended_selection_table’ to create self-contained acoustic data sets
• ‘before’ and ‘eq.dur’ arguments in ‘snrspec’ function
• ‘ts.df’ argument in ‘seltailor’ function to adjust frequency contours
• ‘ts.df’ argument in ‘seltailor’ function to adjust frequency contours
• ‘by.song’ argument in ‘specreator’ to create full song spectrograms with labeled elements

Bug fixes:

• Change “http” to “https” protocol for accessing Xeno-Canto recordings in “querxc”

New functions:

• selection_table: converts data frames into an object of classes ‘selection_table’ or ‘extended_selection_table’
• generic S3 methods for and ‘print’, ‘cbind’ and ‘[’ for classes ‘selection_table’ or ‘extended_selection_table’
• song_param: calculate acoustic parameters at the song (or other hierarchical) level
• spec_param: plot catalog of a single selection with varying spectrogram (display) parameters
• read_wave: wrapper for tuneR’s readWave that ignores extension file case mismatches
• warbleR_options: function to set global parameters *open_wd: open current working directory in wave browser

warbleR 1.1.12

Changes and additions:

• Parallel computing available in ‘querxc’ function when getting metadata
• Parallel computing available in ‘checksels’ function
• New arguments in ‘seltailor’ to allow tailoring frequency contours
• Parallel computing with and without progress bar in all iterative functions
• Parallel computing in windows OS with and without progress bar
• New arguments in ‘specreator’ to customize control lines/box around selections
• Replace internal “message” functions by “cat” (so warnings are printed in black, no red)

New functions:

• track_harm: track harmonic frequency contour
• sim_song: simulate songs

warbleR 1.1.10

Changes and additions:

• Example sound files have been moved to the new R package ‘NatureSounds’ (which is now a dependency)

Bug fixes:

• Problem with multiple tags in ‘catalog’

New functions:

• rm_sil: removes silence segments (i.e. segments with very low amplitude values) from wave files
• consolidate: copies (sound) files scattered in several directories into a single folder

warbleR 1.1.9

Changes and additions:

• “low.freq” and “high.freq” column names in default table format are now “bottom.freq” and “top.freq” respectively
• Moved ‘imp.syrinx’ and ‘imp.raven’ functions to new R package Rraven
• Added a new object class ‘selection.table’
• Split vignette into a series of three with new functions added to the workflow
• New arguments “title”, “by.row”, “prop.mar”, “box” and “rm.axes” in “catalog” function to further customize catalog setup
• New arguments “spec.mar”, “lab.mar” and “group.tag” in “catalog” function to color groups of selections
• “trackfreqs”, “dfts”, and “dfDTW” functions can use “frange.detec” internally to set bandpass limits (see “frange.detec” argument)
• New argument “index” in “seltailor” function to indicate which selection should be tailored
• New argument “wl.freq” in “specan” function to set windows length independently for time and frequency domain measurements
• New measurement “meanpeakf” in specan
• “imp.raven” and “imp.syrinx” now import low and high frequency in kHz (if all.data == FALSE)
• Mew argument “bp” (bandpass) and “wl” in “sig2noise” function to calculate signal-to-noise ratio within a frequency range

New functions:

• is.selection.table: checks if an object is of class ‘selection.table’
• make.selection.table: creates an object of class ‘selection.table’
• frange.detec: detects the frequency range of signals in wave objects
• frange: applies ‘frange.detec’ function iteratively on signals listed in a selection table
• move.imgs: copy/cut image files to folders
• ovlp_sels: find overlapping/duplicated selections
• cut_sels: generates individual sound files (.wav) for each selection in a selection table
• color_spectro: produces spectrograms in which signals are highlighted with colors
• exp.raven: exports selection tables to a .txt file that can be opened in Raven

Bug fixes:

• Reading wave files with no stereo slot in “specan” and “checksels”

warbleR 1.1.8

Bug fixes:

• installation in some linux distributions

warbleR 1.1.7

Changes and additions:

• “seltailor” function now can use external graphic devices (X11 or quartz)

Bug fixes:

• installation in windows OS fixed (pbmcapply conflict in previous version)

warbleR 1.1.6

New functions:

• catalog: create song catalogs (i.e. images with multiple spectrograms in several rows/columns that can be color-labeled)
• catalog2pdf: convert several catalog jpeg’s images into single pdf files
• fixwavs: convert corrupted files to a format that can be imported into R

Changes and additions:

• added new acoustic parameters to “specan” function related to distribution of energy in the time axis. Also time entropy (“time.ent”) and overall entropy (“entropy”) were added
• “centroid” and “mode” parameters were removed from “specan” function
• sig2noise function can return signal-to-noise ratio in dB (“in.dB” argument). It can also measure noise at both sides of the signal or just right before (“before” argument)
  
• scale argument in dfDTW and ffDTW to allow to focus dynamic time warping comparisons on the frequency contours while ignoring differences in pitch
• NAs can be removed from xcorr function output. They can also be ignored when using XCORR in compare.methods function
• filtersels can return selections with or without image files (“missing” argument). It can also return a data frame or the row index of the selections (“index” argument)
• imp.raven function can import raven selections from files that do not share all the columns and extract the sound file name from the selection table name
• FF function (from tuneR package) now can be used for measuring fundamental frequency (in trackfreqs, specan, ffts and ffDTW functions). Seems to be more accurate and faster than fund function from seewave
• window overlap can now be defined by users for fundamental and dominant frequency measures in specan function
• More arguments to control dynamic time warping settings in dfDTW and ffDTW
• Progress bar is available for parallel computing (when parallel > 1) in Linux and for some functions also in iOS. It works using the function mcpbapply::pbmclapply *Spectrogram creating functions (manualoc, specreator, trackfreqs, dfts, ffts, dfDTW, ffDTW, lspec, catalog) can perform faster (4-10 times) using a different image algorithm. This can be called with the argument “fast.spec”. Note that not all spectrogram options are available when fast.spec = TRUE
• selections longer than 20 s can be analyzed with specan function (previously it returned an error, this is going to be slow anyways)
• “clip.edges” argument in compare.methods function to set removing missing values at the start and end of signals when using dfDTW and ffDTW methods
  
• “threshold” argument in compare.methods function to set the amplitude detection threshold for ffDTW, dfDTW and SP methods
  
• “exclude” argument in imp.syrinx function to exclude selection files that cannot be read
• “rm.imcomp” argument in coor.test function to exclude singing events that only have one individual
• “cutoff” argument in coor.test function to determine the minimum number of signals per individual needed for a singing event to be included
• “rm.solo” argument in coor.test to control if signals that are not intercalated at the start or end of the sequence are removed. For instances the sequence of signals A-A-A-B-A-B-A-B-B-B (in which A and B represent different individuals) would be subset to A-B-A-B-A-B
• “incl.wav” argument in filtersels function to indicate if the sound file extension is included in the image files
• 3 different options (formulas) for calculating signal-to-noise ratio are now available (sig2noise function). In addition the “eq.dur” argument allows to measure a noise segment of the same duration than the signal
• grid can be removed from spectrograms in compare.methods function

Bug fixes:

• error produced when calculating frequency limits based on dominant frequency contours in xcorr (when frange is not provided)
• error for identifying readable sound files in some specific .wav formats in checkwavs function
• error when comparing signals with a high duration difference in xcorr

warbleR 1.1.5

Changes and additions:

• “frange” argument in seltailor to define also the frequency range (low.f and high.f)
• trackfreqs, specan now can use low.f and high.f as bandpass limits
• trackfreqs and specreator can plot boxes around signals if low.f and high.f are provided
• manualoc.df data frame example replaced by selec.table (which includes low.f and high.f columns)
• complex searches are now allowed in querxc() following xeno-canto advance query syntax
• added ‘clip.edges’ argument to dfts/ ffts/ dfDTW/ ffDTW/sp.en.ts functions to remove missing values at the start or end of signals
• more detailed checksels output (now including, sampling rate, duration, number of samples and bits)

New functions:

• sp.en.ts: extracts the spectral entropy across signals as a time series iteratively

warbleR 1.1.4

Changes and additions:

• “pb” argument to allow users to knock down progress bars and messages (in looping functions)
• modification to (slightly) improve ‘lspec’ performance
• new “sp” argument in compare.methods to allow users to input their own spectral parameters
• new “custom.contour” argument in trackfreqs to allow users to input their own frequency contours

Bug fixes:

• bug fix for setting ‘path’ argument in several functions
• bug fix to due to missing values (undetected) in amplitude detection in trackfreqs, ffts, dfts, ffDTW, and dfDTW

New functions:

• filtersels: subsets selection data frames based on manually filtered image files
• lspec2pdf: combines lspec images to single pdf files iteratively