NEWS

---

voice 0.5.0

---

NEWS.md setup

Version 0.5.0 2025-03-31

• Class 'voice' created, all functions re-written using S4 method.

Version 0.4.29 2025-03-16

• Added function octave.R.

Version 0.4.28 2025-03-14

• Imports `ggplot2``.
• Adds ggplot2:: at aes, xlab, ylab and ggtitle.

Version 0.4.27 2025-03-14

• Minor corrections.

Version 0.4.26 2025-03-14

• Package title updated to 'Voice Analysis, Speaker Recognition and Mood Inference via music theory'.
• Removed 'mfcc','rf','rcf','rpf','rfc' from the default argument of extract_features.R
• Added 'gain' to the default argument of extract_features.R
• Added function piano_plot.R.
• Added function cut_audio.R.
• Added function assign_notes.R.

Version 0.4.25 2024-06-09

• Added try to line 209 of splitw.R.

Version 0.4.24 2024-03-25

• 'pyannote/speaker-diarization' set to 'pyannote/speaker-diarization-3.1' @ diarize.R
• Added speechbrain to https://raw.githubusercontent.com/filipezabala/voice/master/requirements.txt

Version 0.4.22 2022-MM-DD

• pycall set from '~/miniconda3/envs/pyvoice38/bin/python3.8' to '~/miniconda3/envs/pyvoice/bin/python' @ diarize.R and extract_features.R

Version 0.4.21 2023-04-20

• Dropped the section 3 from voicegnette_R after Prof Brian Ripley's email: "Dear maintainer, Please see the problems shown on https://cran.r-project.org/web/checks/check_results_voice.html. Please correct before 2023-05-03 to safely retain your package on CRAN. The CRAN Team"

Version 0.4.20 2023-04-20

• filesRange and features switched. Now filesRange is in the 3rd and features in the 2nd positions in the argument vector. This change impacts the examples of interp_df.R, interp_mc.R, smooth_df.R, tag.R.
• f0_praat, ..., f8_praat implemented in extract_features.R.
• pycall = '~/miniconda3/envs/pyvoice38/bin/python3.8' added as argument in extract_features.R.
• formants set to fmt in extract_features.R.
• windowShift set to 10 as default in extract_features.R and feat_summary.R.
• extract_features_py.R deprecated.
• voice::conv set to voice::interp at lines 45 and 50 of interp_mc.R.
• Functions get_ removed from spltw.R.
• More information given in voice::diarize documentation.

Version 0.4.19 2023-04-06

• All documentation verified.
• voice::conv set to voice::interp.
• voice::conv_mc set to voice::interp_mc.
• voice::conv_df set to voice::interp_df.
• gender set to sex as argument of voice::extract_features.
• gender set to sex as argument of voice::feat_summary.
• gender set to sex as argument of voice::tag.

Version 0.4.18 2023-04-05

• Added pycall argument to voice::extract_features_py.
• References updated in voice::extract_features.
• References updated in voice::extract_features_py.
• F0 column inferred via voice::extract_features_py set to f0_praat. Line 39 of extract_f0.py, (df_f0_long['f0_praat']) and line 83 of extract_features_py.R (dplyr::select(id, file_name, interval, F0) %>%).
• 1, column inferred via voice::extract_features_py set to f0_praat.
• Column mhs set to f0_mhs in voice::extract_features.
• Lean call to voice::extract_features: features = c('f0', 'formants', # Pitch and formants 'df', 'pf', # Formant dispersion and position 'rf', 'rpf', 'rcf', # Formant removals 'rfc', # (R)e(F)lection (C)oefficients 'mfcc'), # (M)el (Frequency (C)epstral (C)oefficients
• RPf - Formant Position Removal by Zabala (2023) disentangled from Pf by Puts et al (2012) in voice::extract_features.
• voice::diarize calls directly pyannote.audio via reticulate using a token. The download and call of libs.py and diarization-pyannote.py is no longer needed.
• Reconnect to Github after 5 months.

Version 0.4.17 2023-03-14

• voice now depends of R (>= 4.0.0), not R (>= 4.1.0) in order to allow Kaggle installation.
• Deprecated poetry (sent to draft/poetry.R), created R/diarize instead.
• Deprecated data/id_path.rda and man/id_path.Rd (sent to draft/id_path.rda and draft/id_path.Rd), created /data/mozilla_id_path.rda and man/mozilla_id_path.Rd instead.
• Updated .Rbuildignore: ^.*.Rproj$ ^.Rproj.user$ ^cran-comments.md$ ^CRAN-SUBMISSION$ ^requirements.txt$ ^draft$ ^revdep$
• rstudio.com became posit.co.
• Monterey became Ventura.
• voice::diarize exported to NAMESPACE.

Version 0.4.16 2022-09-14

• Allow R/conv_df.R to use unitary features argument.

Version 0.4.15 2022-09-07

• Removed duplicated 'must' from autoDir parameter documentation at poetry.R
• Updated vignettes/voicegnette_R.Rmd.
• Set verbose = FALSE:
  • R/feat_summary.R
  • R/tag.R
• Set to.data.frame = to.data.frame to cn.li call @ R/conv_df.R.

Version 0.4.14 2022-09-02

• As suggested by Benjamin Altmann:
  • Added @return to R/write_list.R. This implies adding a \value field to the corresponding .Rd file.

Version 0.4.12

• As suggested by Benjamin Altmann:
  • Reduced the length of the title to 64 characters, less than 65. In DESCRIPTION:
    nchar('Tools for Voice Analysis, Speaker Recognition and Mood Inference')
  • Package names, software names and API names written in single quotes in title and description.
  • enrich_rttm.R: '\code{voice::read_rttm}'
  • extract_features_py.R: 'Python's' 'Parselmouth'
  • feat_summary.R: 'voice::extract_features'
  • poetry.R: 'Python's' 'pyannote-audio'
  • DESCRIPTION: 'R' and 'Python'
• Reference "Zabala, F.J. (2022) to appear in..." removed from
  • extract_features.R
  • feat_summary.R
  • tag.R
• Added @return to the following .R files regarding exported methods and explaining the functions results in the documentation. This implies adding a \value field to the corresponding .Rd files.
  • R/audio_time.R
  • R/enrich_rttm.R
  • R/extract_features.R
  • R/extract_features_py.R
  • R/feat_summary.R
  • R/get_bit.R
  • R/get_dur.R
  • R/get_left.R
  • R/get_right.R
  • R/get_samp.rate.R
  • R/get_tbeg.R
  • R/get_tdur.R
  • R/is_mono.R
  • R/notes.R
  • R/poetry.R
  • R/read_rttm.R
  • R/splitw.R
  • R/tag.R
• Functions moved to draft directory:
  • draft/chords.R
  • draft/has_audio.R
  • draft/id_file.R
  • draft/is_can.R
  • draft/is_ext.R
  • draft/memory.R
  • draft/na_filter.R
  • draft/notes_summary.R
  • draft/plot_note.R
  • draft/plot_q.R
  • draft/rowProp.R
  • draft/rp.R
  • draft/spoken_time.R
• Functions in which \dontrun was replaced with or simply added \donttest:
  • R/conv_df.R
  • R/conv_mc.R
  • R/enrich_rttm.R
• In order to easily suppress information messages to the console, argument 'verbose' was set to \code{FALSE} as default @ R/extract_features.R .
• tempdir() used @:
  • R/enrich_rttm.R
  • R/get_tbeg.R
  • R/get_tdur.R
  • R/poetry.R
  • R/read_rttm.R
  • R/splitw.R
• \url added to references at
  • R/enrich_rttm.R
  • R/notes_freq.R
  • R/notes.R
  • R/read_rttm.R

Version 0.4.11 2022-08-28

• In order to reduce the examples running time:
  • The following variables were omitted from conv_df.R example: 'zcr','mhs','rms','gain','rfc','ac','cep','dft','css','lps','mfcc'.
  • \dontrun added to conv_mc.R.
  • \dontrun added to na_filter.R.

Version 0.4.10 2022-08-28

• "You also have to remove the licences file which is part of R anyway. Best, Uwe Ligges". License file removed by unlink('LICENSE') command.
• License set to the original 'GPL-3' (instead of 'MIT + file LICENSE', 'MIT' or 'GPL-3 + file LICENSE') @ DESCRIPTION file.

Version 0.4.9 2022-08-26

• License set to 'MIT + file LICENSE' (instead of either 'MIT', 'GPL-3' or 'GPL-3 + file LICENSE') @ DESCRIPTION file.

Version 0.4.8 2022-08-26

• License set to 'MIT' (instead of either 'GPL-3' or 'GPL-3 + file LICENSE') @ DESCRIPTION file.

Version 0.4.7 2022-08-26

• License set to 'GPL-3' (instead of 'GPL-3 + file LICENSE') @ DESCRIPTION file. As I remove it, a new note was presented:
  'File LICENSE is not mentioned in the DESCRIPTION file.'

Version 0.4.6 2022-08-25

• Now depends: R (>= 4.1.0) in order to install on Windows.

Version 0.4.5 2022-08-24

• Following Uwe Ligges' list.
• URL: https://www.apple.com/macos/big-sur/ moved to https://www.apple.com/macos/monterey/ @ README.md.
• Using fully specified URLs starting with the protocol for https://www.xquartz.org/ @ README.md.
• Changed http --> https for https://ffmpeg.org/ and https://www.fon.hum.uva.nl/praat/ @ README.md. Note http://filipezabala.com/ and http://www.rob-mcculloch.org/ are non-secure protocols.
• The canonical URL of the CRAN page was updated for https://cran.r-project.org/package=wrassp, https://cran.r-project.org/package=tuneR, https://cran.r-project.org/package=seewave and https://cran.r-project.org/package=gm.
• Can't find where to omit "+ file LICENSE", the DESCRIPTION file requires it.

Version 0.4.4 2022-08-23

• The list.txt file was generated by write_list.R function. \dontrun environment was applied to solve the problem at devtools::check.

Version 0.4.3 2022-07-18

• Replaced \url by \doi at 10.1121/1.421048 and 10.1098/rspb.2011.0829 @ extract_features.R.
• Where is list.txt?????

Version 0.4.2 2022-07-18

• Removed “≈” U+2248 Almost Equal
• Adjusted Maintainer and Authors@R (Zabala Filipe J.): Maintainer: 'Zabala Filipe J. [email protected]' Authors@R: 'Zabala Filipe J. [email protected]'
• Where is list.txt?????

Version 0.4.1 2022-07-18

• Many improvements after devtools::release().
• Where is list.txt?????

Version 0.4.0 2022-07-17

• Many improvements after devtools::check().
• Where is list.txt?????

Version 0.3.22 2022-07-15

• Update documentation @ extract_features.R.

Version 0.3.21 2022-06-10

• Lost changes.

Version 0.3.20 2022-05-30

• Corrected verbose = TRUE @ tag.R and @ feat_summary.R.

Version 0.3.19 2022-05-30

• Added @ voice::extract_features in line 22: #' @param verbose Logical. Should the running status be showed?
• Verbose also added to feat_summary.R and tag.R.
• feat_summary2 set to feat_summary in line 31 @ feat_summary.R.

Version 0.3.18 2022-05-23

• Added dplyr::tibble in line 92 @ feat_summary.R.

Version 0.3.17 2022-05-22

• Solved summarizing problem when wav_path was a directory (and not a file name) @ feat_summary.R.

Version 0.3.16 2022-05-22

• Lost changes.

Version 0.3.15 2022-05-21

• Updated feat_summary.R normalizing dirnames @ Media and @ Extended (full path)
• Added extdata directory.

Version 0.3.14 2022-05-07

• Lost changes.

Version 0.3.13 2022-MM-DD

• Set id_path.Rda to mozilla_id_path.Rda

Version 0.3.12 2022-02-13

• Removed toLowerGroupBy @ feat_summary.R and tag.R.

Version 0.3.11 2022-02-11

• Set check.mono and stereo2mono to FALSE @ extract_features.R, feat_summary.R and tag.R.
• Added x <- dplyr::as_tibble(x) @ feat_summary.R.

Version 0.3.10 2022-02-11

• Set wavFiles <- do.call(rbind, wavFiles) to wavFiles <- do.call(rbind, as.list(unlist(wavFiles))) @ extract_features.R.

Version 0.3.9 2022-02-10

• Tidy up tag.R documentation, removing @param subj.id, @param media.id and @param subj.id.simplify.
• Added @param wavPath in feat_summary.R and tag.R documentation.
• Added @param wavPathName in feat_summary.R and tag.R documentation.
• Normalized dirnames @ feat_summary.R

Version 0.3.8 2022-02-09

• dplyr::vars @ feat_summary.R

Version 0.3.7 2022-02-09

• Make Media and Extended datasets consistent in documentation.
• extract_features.R
  • Drop 'file_name_ext' column @ output.
  • Add 'wavPath' as the third column @ extract_features.R output.
  • Admit directories or files in the main argument x.
• feat_summary.R
  • path > wav_path, id_seq > slice_seq, id_seq_file > slice_seq_file
  • Added 'groupBy' argument
  • Check tag.R applied on Canonical data
• tag.R
  • Create examples
  • Remove arguments subj.id, media.id and subj.id.simplify
  • Disable audio_time

Version 0.3.6 2022-01-02

• Added @references Ardila et al (2019).
• Added voice::plot_q.R, a plot_q[uality] function.
• Imports magrittr at DESCRIPTION file.

Version 0.3.5 2022-01-02

• Removed voice_ex1.rda, 7.5MB.
• Deleted NAMESPACE.

Version 0.3.4 2022-01-01

• Code: id_path <- E %>% select(client_id:path) save(id_path, file = '~/Dropbox/D_Filipe_Zabala/pacotes/voice/data/id_path.rda', compress = 'xz') voice::id_path

Version 0.3.3 2022-01-01

• Code: Eid <- E %>% select(client_id:path) save(Eid, file = '~/Dropbox/D_Filipe_Zabala/pacotes/voice/data/id_path.rda', compress = TRUE) data(voice::id_path)

Version 0.3.2 2021-12-15

• Harmonize voice::extract_features.R and voice::tag.R arguments.

Version 0.3.1 2021-12-03

• extract_features2 set to voice::extract_features @ line 24 @ feat_summary.R.

Version 0.3.0 2021-12-03

• Set argument 'directory' to 'x' @ extract_features.R.
• Set features colnames to lowercase @ extract_features.R.
• Added df, pf, rf, rcf and rpf as argument of 'features' @ extract_features.R.
• Deprecate argument 'extraFeatures' @ extract_features.R and, feat_summary.R and tag.R.
• Contemplate ZCR, RFC, AC, and MFCC (numbers) @ feat_summary.R.

Version 0.2.11 2021-11-27

• Added dplyr::bind_cols to line 602 @ extract_features.R.

Version 0.2.10 2021-11-27

• RCf <- paste0('RCf', 2:8)
• RPf <- paste0('RPf', 2:8)

Version 0.2.9 2021-11-27

• Formant Dispersion Removal (RDf) set to Formant Cumulated Removal (RCf), line 53 @ feat_summary.R
• dat$Dfn <- dat$Fn-dat$F1 set to dat$Dfn <- (dat$Fn-dat$F1)/(n-1) @ Df - Formant Dispersion by Fitch (1997) @ extract_features.

Version 0.2.8 2021-11-27

• Formant Dispersion Removal (RDf) set to Formant Cumulated Removal (RCf).
• RCf set from i=1,...,8 to i=2,...,8.

Version 0.2.7 2021-11-10

• Added suppressMessages to line 61 @ smooth_df.R.
• Summary generalized to all features obtained from voice::extract_features @ voice::feat_summary.

Version 0.2.6 2021-11-07

• Added 'rule = 2' to line 52 @ conv.R.

Version 0.2.5 2021-10-25

• Added voice::audio_time to voice::tag.
• New tag order @ voice::feat_summary: mean, sd, vc, median, iqr, mad.

Version 0.2.4 2021-10-14

• Added tag_F0_iqr = IQR(F0, na.rm = TRUE) @ feat_summary.R.
• Added tag_F0_mad = mad(F0, na.rm = TRUE) @ feat_summary.R.

Version 0.2.3 2021-10-13

• Set x[,subj.id] <- as.character(cumsum(!duplicated(x[,subj.id]))) @ tag.R.

Version 0.2.2 2021-10-13

• Added parameter subj.id.simplify and all voice::extract_features parameters @ tag.R.

Version 0.2.1 2021-10-13

• tag.R incorporates parameter mc.cores = 1 and dropped i = 4.
• set tag.R from feat_summary2 to feat_summary at line 26.
• feat_summary.R running voice::extract_features instead of extract_features.
• feat_summary.R return(M_summ) instead of return(M).

Version 0.2.0 2021-10-13

• Added has_audio.R.
• Added audio_time.R.
• Added spoken_time.R
• Added feat_summary.R
• Added tag.R.

Version 0.1.1 2021-09-27

• audio_id.R set to id_file.R.
• audio_id(x, i = 5, drop_fn = FALSE) set to id_file(x, col = NULL, pattern = '[_.]', i = 5, drop_col = FALSE)
• id set to id_seq, id_file set to id_seq_file @ extract_features.R.

Version 0.1.0 2021-09-26

• extract_features creates a general id and an id_file, counting by file.
• 'formant nested removal' (RNf) set to 'formant dispersion removal' (RDf) by Zabala (2021/2022) @ extract_features.R and @ data.R.
• Removed person("Salum Jr.", "Giovanni A.", email = "[email protected]", role = "aut").
• splitw must show the elapsed time
• drop_fn = TRUE set to drop_fn = FALSE

Version 0.0.0.9047 2021-09-21

• audio_id.R function created.
• Exported voice_ex1.rda and data.R.
• toFile = F set to toFile = FALSE @ extract_features.R.
• features.list[[i]] <- dplyr::tibble() set to lapply(features.list, dplyr::tibble) @ extract_features.R.
• Added 'formant dispersion' (Df) by Fitch (1997) (doi:10.1121/1.421048)
• Added 'formant position' (Pf) by Puts, Apicella & Cárdenas (2011) (https://doi.org/10.1098/rspb.2011.0829)
• Added 'formant removal' (Rf) by Zabala (2021/2022) @ extract_features.R.
• Added 'formant nested removal' (RNf) by Zabala (2021/2022) @ extract_features.R.
• Added 'formant position removal' (RPf) by Zabala (2021/2022) @ extract_features.R.

Version 0.0.0.9046 2021-09-15

• Added .name_repair = 'unique' to tibble::as_tibble @ smooth_df.R. This solved the Warning: The x argument of as_tibble.matrix() must have unique column names if .name_repair is omitted as of tibble 2.0.0. Using compatibility .name_repair.

Version 0.0.0.9045 2021-09-15

• names(rttm) <- basename(rttmFiles), or file/base names inserted as names in list @ read_rttm.R.
• Added columns ~black and ~Black to notes_freq.R.
• Added method == 'black' and method == 'Black' @ notes.R.
• Added distance <- diff(freq) @ notes.R.
• method set to measure @ notes.R.
• frequency set to freq @ notes_freq.R. . New col order @ notes_freq.R: freq, spn.lo, spn.hi, spn, midi, black, Black, wavelength . as_tibble set to tibble::as_tibble @ smooth_df.R. . xs_li set to snum_li @ smooth_df.R. . xs_df set to snum_df @ smooth_df.R. . xs_final set to xs @ smooth_df.R. . Ordered Imports: dplyr, e1071, ellipse, ggfortify, ggplot2, R.utils, RColorBrewer, reticulate, seewave, tibble, tidyr, tuneR, VIM, wrassp, zoo

Version 0.0.0.9044 2021-09-12

• Parameters fromWav, fromRttm admits either file or directory @ splitw.R.

Version 0.0.0.9043 2021-09-12

• Added parameter as.tibble = TRUE @ enrich_rttm.R.

Version 0.0.0.9042 2021-09-12

• Changed column name from 'id.split' to 'id_split @ enrich_rttm.R.
• is.num <- unlist(lapply(x, class)) == 'numeric' changed to is.num <- sapply(x, class) %in% c('integer', 'numeric') @ conv_df.
• Added smooth_df.R.
• Drop by.filter parameter @ conv_df.R.
• id and colnum shifted @ conv_df.R documentation.
• Added parameter as.tibble = TRUE @ enrich_rttm.R.

Version 0.0.0.9041 2021-09-10

• Changed mc.cores = parallel::detectCores() to mc.cores = 1 @ conv_df.R, conv_mc.R, extract_features.R.
• Argument silence.gap set from 0.4 to 0.5 (universal standard) @ splitw.R, enrich_rttm.R.
• Changed column name from 'id.min.time' to 'id.split' @ enrich_rttm.R.

Version 0.0.0.17

• Solved the recursive issue in diarization-pyannote.py (by Sola):

import os import re

file_list = [] for root, dirs, files in os.walk('./'): file_list = [f for f in files if re.findall(r'[.]wav', f, re.IGNORECASE)] break

print(file_list)

• os.listdir?
• os.walk(pathfrom, level=1)?
• os.walk(pathfrom, recursive = FALSE)?
• https://stackoverflow.com/questions/229186/os-walk-without-digging-into-directories-below

Version 0.0.0.1 2018-11-21 08:30:00 GMT-3