├── .gitignore
├── LICENCE
├── LICENCE.pyDAWG
├── README.md
├── align.py
├── alignannotation.praat
├── aligntier.praat
├── book
    ├── .chktexrc
    ├── .gitignore
    ├── Makefile
    ├── book.bib
    ├── book.tex
    └── preamble.tex
├── cleaninterval.praat
├── generatedict.praat
├── generatedict.py
├── install.sh
├── install_win.bat
├── mkrelease.sh
├── par.dutch
    ├── DICT
    ├── HVITECONF
    ├── MMF
    ├── MONOPHONES
    └── PRECONFIGNIST
├── par.spanish
    ├── DICT
    ├── HVITECONF
    ├── MMF
    ├── MONOPHONES
    └── PRECONFIGNIST
├── phonetizer.py
├── procs.praat
├── pyDAWG.py
├── settings.praat
├── settings_ni.praat
└── setup.praat


/.gitignore:
--------------------------------------------------------------------------------
1 | *.pyc
2 | *.tar.xz
3 | *.zip
4 | settings
5 | isettings
6 | phon.txt
7 | test
8 | 


--------------------------------------------------------------------------------
/LICENCE:
--------------------------------------------------------------------------------
 1 | The MIT License (MIT)
 2 | 
 3 | Copyright (c) <year> <copyright holders>
 4 | 
 5 | Permission is hereby granted, free of charge, to any person obtaining a copy
 6 | of this software and associated documentation files (the "Software"), to deal
 7 | in the Software without restriction, including without limitation the rights
 8 | to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
 9 | copies of the Software, and to permit persons to whom the Software is
10 | furnished to do so, subject to the following conditions:
11 | 
12 | The above copyright notice and this permission notice shall be included in
13 | all copies or substantial portions of the Software.
14 | 
15 | THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16 | IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17 | FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
18 | AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19 | LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
20 | OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
21 | THE SOFTWARE.
22 | 


--------------------------------------------------------------------------------
/LICENCE.pyDAWG:
--------------------------------------------------------------------------------
 1 | # encoded in utf-8.
 2 | 
 3 | Copyright (c) 2011, Wojciech Muła
 4 | All rights reserved.
 5 | 
 6 | Redistribution and use in source and binary forms, with or
 7 | without modification, are permitted provided that the following
 8 | conditions are met:
 9 | 
10 | * Redistributions of source code must retain the above copyright
11 | notice, this list of conditions and the following disclaimer.
12 | * Redistributions in binary form must reproduce the above
13 | copyright notice, this list of conditions and the following
14 | disclaimer in the documentation and/or other materials
15 | provided with the distribution.
16 | * Neither the name of the Wojciech Muła nor the names of its
17 | contributors may be used to endorse or promote products
18 | derived from this software without specific prior written
19 | permission.
20 | 
21 | THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND
22 | CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES,
23 | INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
24 | MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
25 | DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR
26 | CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
27 | SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
28 | LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF
29 | USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
30 | AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
31 | LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
32 | IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF
33 | THE POSSIBILITY OF SUCH DAMAGE.
34 | # eof
35 | 


--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
 1 | # Praatalign: an interactive Praat plug-in for performing phonetic forced alignment
 2 | 
 3 | #### Due to licencing the english and sampa models have been removed.
 4 | 
 5 | ### Version 2.0a
 6 | 
 7 | Praatalign is a plug-in for Praat that can be used to do forced phonetic
 8 | alignment on speech signals and in particular free speech. Praatalign combines
 9 | the powerful HTK toolkit with the interactivity and modularity of Praat to
10 | create an interactive, intuitive and extendable application. Currently
11 | Praatalign uses the acoustic models trained by Schiel et al. used in
12 | MAUS\[1,2\].
13 | 
14 | ### Releases
15 | 
16 | [Latest manual](https://github.com/dopefishh/praatalign/releases/download/v2.0/book_2.0.pdf)
17 | 
18 | Latest plugin:
19 | [.tar.xz](https://github.com/dopefishh/praatalign/releases/download/v2.0a/praatalign_2.0a.tar.xz)
20 | [.zip](https://github.com/dopefishh/praatalign/releases/download/v2.0a/praatalign_2.0a.zip)
21 | 
22 | [Older releases](https://github.com/dopefishh/praatalign/releases)
23 | 
24 | ### How to cite
25 | 
26 | ```tex
27 | @misc{praatalign2.0a,
28 | 	author={Lubbers, Mart and Torreira, Francisco},
29 | 	title={Praatalign: an interactive Praat plug-in for performing phonetic forced alignment},
30 | 	howpublished={\url{https://github.com/dopefishh/praatalign}},
31 | 	year={2013-2018},
32 | 	note={Version 2.0}
33 | }
34 | ```
35 | 
36 | - IEEE: `M. Lubbers and F. Torreira, Praatalign: an interactive Praat plug-in for performing phonetic forced alignment. 2016.`
37 | - APA: `Lubbers, M., & Torreira, F. (2016). Praatalign: an interactive Praat plug-in for performing phonetic forced alignment. Retrieved from https://github.com/dopefishh/praatalign`
38 | 
39 | ### Authors
40 | 
41 | Mart Lubbers (mart@martlubbers.net) and Francisco Torreira
42 | 
43 | ### References
44 | 
45 | 1. http://www.bas.uni-muenchen.de/Bas/BasMAUS.html
46 | 2. Schiel F. (1999): Automatic Phonetic Transcription of Non-Prompted Speech, Proc. of the ICPhS 1999. San Francisco, August 1999. pp. 607-610.
47 | 


--------------------------------------------------------------------------------
/align.py:
--------------------------------------------------------------------------------
  1 | #!/usr/bin/env python
  2 | # -*- coding: utf-8 -*-
  3 | 
  4 | import codecs
  5 | import itertools
  6 | import logging
  7 | import os
  8 | import phonetizer as ph
  9 | import subprocess
 10 | import sys
 11 | 
 12 | 
 13 | def force(phonetizer, utterance, starttime, endtime, wavefile,
 14 |           soxbinary, hvitebinary, hcopybinary, parameterdir,
 15 |           basename, hdr=True, code='w'):
 16 |     """Wrapper for the _force function that writes the status to a file for
 17 |     feedback via praat module.
 18 | 
 19 |     statusses can be:
 20 |         done     - Alignment was successfull.
 21 |         missox   - Sox binary not found.
 22 |         mishcopy - HCopy binary not found.
 23 |         mishvite - HVite binary not found.
 24 |     """
 25 |     status = _force(
 26 |         phonetizer, utterance, starttime, endtime, wavefile,
 27 |         soxbinary, hvitebinary, hcopybinary, parameterdir,
 28 |         basename, hdr, code)
 29 |     with open('{}.status'.format(basename), 'w') as f:
 30 |         f.write(status)
 31 |     return not status
 32 | 
 33 | 
 34 | def _force(phonetizer, utterance, starttime, duration, wavefile,
 35 |            soxbinary, hvitebinary, hcopybinary, parameterdir,
 36 |            basename, hdr, code):
 37 |     """
 38 |     Force aligns the given utterance, all parameters are passed by kwarg
 39 |     """
 40 |     logging.info('Starting to align: {}'.format(code))
 41 |     logging.info('Removing old files')
 42 |     for suffix in ['dot', 'htk', 'nis', 'rec', 'slf', 'status']:
 43 |         try:
 44 |             os.remove('{}.{}'.format(basename, suffix))
 45 |         except:
 46 |             pass
 47 | 
 48 |     # Open the preconfig and extract the sourcerate
 49 |     sourcerate = str(1e7/625)
 50 | 
 51 |     # Load the phonetizer
 52 |     pron = phonetizer.phonetize(utterance) or [[['<nib>']]]
 53 | 
 54 |     canonical = [''.join(p[0]) for p in pron]
 55 |     logging.info('Utterance phonetized spawned')
 56 | 
 57 |     # Create the graph file
 58 |     dawg = phonetizer.todawg(pron)
 59 |     with open('{}.slf'.format(basename), 'w') as f:
 60 |         f.write(phonetizer.toslf(*dawg))
 61 |     with open('{}.dot'.format(basename), 'w') as f:
 62 |         f.write(phonetizer.todot(*dawg))
 63 |     logging.info('SLF file created')
 64 | 
 65 |     # Run the sound processing
 66 |     soxcommand = [
 67 |         soxbinary, wavefile,
 68 |         '-t', 'sph',
 69 |         '-e', 'signed-integer',
 70 |         '-b', '16', '-c', '1',
 71 |         '{}.nis'.format(basename),
 72 |         'trim', starttime, duration,
 73 |         'rate', '-s', '-a', sourcerate]
 74 |     logging.info(' '.join(soxcommand))
 75 |     proc = subprocess.Popen(soxcommand,
 76 |                             stderr=subprocess.PIPE, stdout=subprocess.PIPE)
 77 |     out, err = proc.communicate()
 78 |     if proc.returncode == 127:
 79 |         return 'Sox binary couldn\'t be found...\n'\
 80 |             'It is searched for in {}'.format(hcopybinary)
 81 |     logging.info('Sox ran({}):\n\tout: {}\n\terr: {}'.format(
 82 |         proc.returncode, out, err))
 83 | 
 84 |     # Run the HCopy process
 85 |     hcopycommand = [
 86 |         hcopybinary,
 87 |         '-T', '0',
 88 |         '-C', os.path.join(parameterdir, 'PRECONFIGNIST'),
 89 |         '{}.nis'.format(basename), '{}.htk'.format(basename)]
 90 | 
 91 |     logging.info(' '.join(hcopycommand))
 92 |     proc = subprocess.Popen(hcopycommand,
 93 |                             stderr=subprocess.PIPE, stdout=subprocess.PIPE)
 94 |     out, err = proc.communicate()
 95 |     if proc.returncode == 127:
 96 |         return 'HCopy binary couldn\'t be found...\n'\
 97 |             'It is searched for in {}'.format(hcopybinary)
 98 |     logging.info('HCopy ran({}):\n\tout: {}\n\terr: {}'.format(
 99 |         proc.returncode, out, err))
100 | 
101 |     # Run the HVite actual alignment
102 |     hvitecommand = [
103 |         hvitebinary,
104 |         '-C', os.path.join(parameterdir, 'HVITECONF'),
105 |         '-w', '-X', 'slf',
106 |         '-H', os.path.join(parameterdir, 'MMF'),
107 |         '-s', '7.0',
108 |         '-p', '6.0',
109 |         os.path.join(parameterdir, 'DICT'),
110 |         os.path.join(parameterdir, 'MONOPHONES'),
111 |         '{}.htk'.format(basename)]
112 |     logging.info(' '.join(hvitecommand))
113 |     proc = subprocess.Popen(hvitecommand,
114 |                             stderr=subprocess.PIPE, stdout=subprocess.PIPE)
115 |     out, err = proc.communicate()
116 |     if proc.returncode == 127:
117 |         return 'HVite binary couldn\'t be found...\n'\
118 |             'It is searched for in {}'.format(hvitebinary)
119 |     logging.info('HVite ran({}):\n\tout: {}\n\terr: {}'.format(
120 |         proc.returncode, out, err))
121 |     if proc.returncode != 0:
122 |         return 'HVite failed with the following error: {}'.format(err)
123 | 
124 |     # Open the output file
125 |     out = 'praat_temp_out'
126 |     ortwords = utterance.replace(",", "").split(' ')
127 |     with open(out, code) as fileio:
128 |         logging.info('Output file selected')
129 |         with open('{}.rec'.format(basename), 'r') as f:
130 |             # Write if necessary the header
131 |             if hdr != 'False':
132 |                 fileio.write('start,end,label,type\n')
133 |                 logging.info('Header written')
134 |             word = None
135 |             for d in itertools.imap(lambda x: x.split(), f):
136 |                 # Parse the time parameters and convert them to seconds
137 |                 start = float(starttime) + int(d[0]) / 1e7
138 |                 end = float(starttime) + int(d[1]) / 1e7
139 | 
140 |                 # Detect word boundaries
141 |                 if d[2] == '<':
142 |                     word = (end, '')
143 |                 # If the end of a word is reached write the word to file
144 |                 elif d[2] == '#':
145 |                     fileio.write('{:f},{:f},{},w\n'.format(
146 |                         word[0], start, word[1]))
147 |                     fileio.write('{:f},{:f},{},c\n'.format(
148 |                         word[0], start, canonical.pop(0)))
149 |                     fileio.write('{:f},{:f},{},o\n'.format(
150 |                         word[0], start, ortwords.pop(0).encode('utf-8')))
151 |                     word = (end, '')
152 |                 # Else add the current phone to the current word
153 |                 else:
154 |                     word = (word[0], word[1] + d[2])
155 |                 # If the length is non zero write the phone to the file
156 |                 if end - start > 0:
157 |                     fileio.write('{:f},{:f},{},p\n'.format(
158 |                         start, end, d[2]))
159 |                     fileio.write('{:f},{:f},{},l\n'.format(
160 |                         start, end, float(d[3])))
161 |             logging.info('Datafile written')
162 |     logging.info('Finished')
163 |     return ''
164 | 
165 | 
166 | def parsesettings(filepath):
167 |     with open(filepath, 'r') as f:
168 |         settings = {k: v.strip() for k, v in (x.split(': ') for x in f)}
169 |     return settings
170 | 
171 | 
172 | if __name__ == '__main__':
173 |     # Load the sett
174 |     sett = parsesettings('isettings')
175 |     sett.update(parsesettings('settings'))
176 | 
177 |     # Initialize the logger
178 |     logging.basicConfig(filename=sett['LOG'], level=20,
179 |                         format='%(created)f: %(message)s')
180 | 
181 |     # Load the phonetizer
182 |     try:
183 |         phone = ph.getphonetizer(
184 |             sett['LAN'], sett['PHO'], sett['DCT'], sett['RUL'])
185 |     except UnicodeError:
186 |         with open('temp.status', 'w') as f:
187 |             f.write('Unicode error. Check if the files are utf-8')
188 |         exit()
189 |     except IOError as e:
190 |         if e.filename == sett['DCT']:
191 |             error = 'Dictionary file couldn\'t be found\n'\
192 |                 'It is searched for in {}'.format(e.filename)
193 |         elif e.filename == sett['RUL']:
194 |             error = 'Ruleset file couldn\'t be found\n'\
195 |                 'It is searched for in {}'.format(e.filename)
196 |         elif e.filename == sett['PHO']:
197 |             error = 'Universal phonetizer file couldn\'t be found\n'\
198 |                 'It is searched for in {}'.format(e.filename)
199 |         else:
200 |             error = 'Some io error: ' + str(e)
201 |         with open('temp.status', 'w') as f:
202 |             f.write(error)
203 |         exit()
204 | 
205 |     p = 'par.{}'.format(sett['MOD'])
206 | 
207 |     if sys.argv[1] == 'tier':
208 |         # Read the data
209 |         with codecs.open(sett['OUT'], 'r', 'utf-8') as f:
210 |             data = f.readlines()
211 | 
212 |         # Setup the code for writing and header sett
213 |         first = 0
214 |         code = 'w'
215 |         hdr = 'True'
216 | 
217 |         # Parse the data
218 |         data = map(lambda x: x.strip().split('\t'), data[1:])
219 | 
220 |         # Setup and align all data
221 |         for i, (start, _, utt, end) in enumerate(data):
222 |             if first == 0:
223 |                 first += 1
224 | 
225 |         # If neccesary unset the header and change write mode to append
226 |             elif first == 1:
227 |                 hdr = 'False'
228 |                 code = 'a'
229 | 
230 |         # Check for empty pre and post annotations for extended bounds option
231 |             if i > 0:
232 |                 start = max(float(start)-float(sett['THR']),
233 |                             float(data[i-1][3]))
234 |             if i < len(data) - 1:
235 |                 end = min(float(end) + float(sett['THR']),
236 |                           float(data[i+1][0]))
237 | 
238 |             dur = str(float(end)-float(start))
239 |         # Do the actual alignment
240 |             if not force(
241 |                     phone, utt, str(start), dur, sett['WAV'],
242 |                     sett['SOX'], sett['HVB'], sett['HCB'], p, 'temp',
243 |                     hdr=hdr, code=code):
244 |                 break
245 |     elif sys.argv[1] == 'annotation':
246 |         force(phone, sett['UTT'], sett['STA'], sett['DUR'], sett['WAV'],
247 |               sett['SOX'], sett['HVB'], sett['HCB'], p, 'temp')
248 | 


--------------------------------------------------------------------------------
/alignannotation.praat:
--------------------------------------------------------------------------------
  1 | include procs.praat
  2 | # Get current selection
  3 | 	selection_start = Get starting point of interval
  4 | 	selection_end = Get end point of interval
  5 | 	utterance$ = Get label of interval
  6 | 
  7 | 	# Settings loads: phonetier_name$, wordtier_name$, cantier_name$, tmpfile$,
  8 | 	# pythonex$, boundary_margin, llhtier_name$
  9 | 	@loadSettings: 
 10 | 
 11 | 	# Load editor and longsound info: sound_file$, sound_object$,
 12 | 	# textgrid_object$, selected_tier, sound_duration, pitch_on,
 13 | 	# intensity_on, spectrum_on, formant_on, pulses_on
 14 | 	@loadFileInfo:
 15 | 
 16 | 	statusfile$ = "temp.status"
 17 | 
 18 | # Unshow pitch, intensity and spectrum if they are enabled
 19 | 	@toggleGUIValues:
 20 | 
 21 | # restore the selection as it might be crappy due to sound selection
 22 | 	Select: selection_start, selection_end
 23 | 
 24 | # Check if the previous or next interval are empty
 25 | 	Select previous interval
 26 | 	interval_label$ = Get label of interval
 27 | 	if interval_label$ = ""
 28 | 		margin_before = boundary_margin
 29 | 	else
 30 | 		margin_before = 0
 31 | 	endif
 32 | 	Select next interval
 33 | 	Select next interval
 34 | 	interval_label$ = Get label of interval
 35 | 	if interval_label$ = ""
 36 | 		margin_after = boundary_margin
 37 | 	else
 38 | 		margin_after = 0
 39 | 	endif
 40 | 	Select previous interval
 41 | 	Zoom to selection
 42 | 	Zoom out
 43 | endeditor
 44 | 
 45 | # Calculate the true start and end times with respect to the extended bounds
 46 | selection_start = max(selection_start - margin_before, 0)
 47 | selection_end = min(selection_end + margin_after, sound_duration)
 48 | selection_duration = selection_end - selection_start
 49 | 
 50 | # Get the index the tiers
 51 | selectObject: textgrid_object$
 52 | 
 53 | phonetier_number = -1
 54 | llhtier_number = -1
 55 | wordtier_number = -1
 56 | cantier_number = -1
 57 | orttier_number = -1
 58 | 
 59 | if phonetier_name$ <> ""
 60 | @indexOfTier: phonetier_name$
 61 | 	phonetier_number = indexOfTier.number
 62 | endif
 63 | 
 64 | if orttier_name$ <> ""
 65 | @indexOfTier: orttier_name$
 66 | 	orttier_number = indexOfTier.number
 67 | 	if indexOfTier.inserted == 1
 68 | 		if phonetier_name$ <> ""
 69 | 			phonetier_number = phonetier_number + 1
 70 | 		endif
 71 | 	endif
 72 | endif
 73 | 
 74 | if llhtier_name$ <> ""
 75 | @indexOfTier: llhtier_name$
 76 | 	llhtier_number = indexOfTier.number
 77 | 	if indexOfTier.inserted == 1
 78 | 		if orttier_name$ <> ""
 79 | 			orttier_number = orttier_number + 1
 80 | 		endif
 81 | 		if phonetier_name$ <> ""
 82 | 			phonetier_number = phonetier_number + 1
 83 | 		endif
 84 | 	endif
 85 | endif
 86 | 
 87 | if wordtier_name$ <> ""
 88 | @indexOfTier: wordtier_name$
 89 | 	wordtier_number = indexOfTier.number
 90 | 	if indexOfTier.inserted == 1
 91 | 		if orttier_name$ <> ""
 92 | 			orttier_number = orttier_number + 1
 93 | 		endif
 94 | 		if phonetier_name$ <> ""
 95 | 			phonetier_number = phonetier_number + 1
 96 | 		endif
 97 | 		if llhtier_name$ <> ""
 98 | 			llhtier_number = llhtier_number + 1
 99 | 		endif
100 | 	endif
101 | endif
102 | 
103 | if cantier_name$ <> ""
104 | @indexOfTier: cantier_name$
105 | 	cantier_number = indexOfTier.number
106 | 	if indexOfTier.inserted == 1
107 | 		if orttier_name$ <> ""
108 | 			orttier_number = orttier_number + 1
109 | 		endif
110 | 		if phonetier_name$ <> ""
111 | 			phonetier_number = phonetier_number + 1
112 | 		endif
113 | 		if llhtier_name$ <> ""
114 | 			llhtier_number = llhtier_number + 1
115 | 		endif
116 | 		if wordtier_name$ <> ""
117 | 			wordtier_number = wordtier_number + 1
118 | 		endif
119 | 	endif
120 | endif
121 | 
122 | # Clean up the phone tier
123 | if orttier_name$ <> ""
124 | 	@cleanAnnotation: textgrid_object$, orttier_number, selection_start, selection_end
125 | endif
126 | if phonetier_name$ <> ""
127 | 	@cleanAnnotation: textgrid_object$, phonetier_number, selection_start, selection_end
128 | endif
129 | if wordtier_name$ <> ""
130 | 	@cleanAnnotation: textgrid_object$, wordtier_number, selection_start, selection_end
131 | endif
132 | if cantier_name$ <> ""
133 | 	@cleanAnnotation: textgrid_object$, cantier_number, selection_start, selection_end
134 | endif
135 | if llhtier_name$ <> ""
136 | 	@cleanAnnotation: textgrid_object$, llhtier_number, selection_start, selection_end
137 | endif
138 | 
139 | # Write the interval specific settings to the settings file
140 | writeFileLine("isettings",
141 | ..."STA: ", selection_start, newline$,
142 | ..."DUR: ", selection_duration, newline$,
143 | ..."UTT: ", utterance$, newline$,
144 | ..."WAV: ", sound_file$)
145 | 
146 | # Do the actual alignment
147 | system 'pythonex$' align.py annotation
148 | 
149 | returnstatus$ = readFile$: statusfile$
150 | if returnstatus$ == ""
151 | 	@insertTableTextGrid: tmpfile$, textgrid_object$, phonetier_name$,
152 | ... orttier_name$, wordtier_name$, cantier_name$, llhtier_name$,
153 | ... phonetier_number, orttier_number, wordtier_number, cantier_number,
154 | ... llhtier_number
155 | else
156 | 	pauseScript: "Error! check the info window for details..."
157 | 	appendInfoLine: returnstatus$
158 | endif
159 | 
160 | # Reset pitch, intensity and spectrum if they were unset before
161 | editor
162 | 	@toggleGUIValues:
163 | 


--------------------------------------------------------------------------------
/aligntier.praat:
--------------------------------------------------------------------------------
  1 | include procs.praat
  2 | 	# Settings loads: phonetier_name$, wordtier_name$, cantier_name$, tmpfile$,
  3 | 	# orttier_name$, pythonex$, boundary_margin
  4 | 	@loadSettings:
  5 | 
  6 | 	# Load editor and longsound info: sound_file$, sound_object$,
  7 | 	# textgrid_object$, selected_tier, sound_duration, pitch_on,
  8 | 	# intensity_on, spectrum_on, formant_on, pulses_on
  9 | 	@loadFileInfo:
 10 | 
 11 | 	# Extract the tier
 12 | 	Extract entire selected tier
 13 | endeditor
 14 | 
 15 | # Select the tier and convert it to a table and write it to a file
 16 | extracted_tier$ = selected$("TextGrid", 1)
 17 | Down to Table: "no", 6, "yes", "no"
 18 | Save as tab-separated file: tmpfile$
 19 | 
 20 | # Remove all the created temporary objects
 21 | Remove
 22 | select TextGrid 'extracted_tier$'
 23 | Remove
 24 | 
 25 | # Write the tier specific settings
 26 | writeFileLine("isettings",
 27 | ..."WAV: ", sound_file$)
 28 | 
 29 | # Remove the tiers if they already exist
 30 | selectObject: textgrid_object$
 31 | number_of_tiers = Get number of tiers
 32 | tiernumber = 1
 33 | while tiernumber < number_of_tiers
 34 |     nametier$ = Get tier name: tiernumber
 35 |     if nametier$ = phonetier_name$ or nametier$ = wordtier_name$ or nametier$ = cantier_name$ or nametier$ = llhtier_name$ or nametier$ = orttier_name$
 36 |         Remove tier: tiernumber
 37 |     else
 38 |         tiernumber = tiernumber + 1
 39 |     endif
 40 |     number_of_tiers = Get number of tiers
 41 | endwhile
 42 | 
 43 | # Get the index the tiers
 44 | selectObject: textgrid_object$
 45 | 
 46 | phonetier_number = -1
 47 | llhtier_number = -1
 48 | wordtier_number = -1
 49 | cantier_number = -1
 50 | orttier_number = -1
 51 | 
 52 | if phonetier_name$ <> ""
 53 | @indexOfTier: phonetier_name$
 54 | 	phonetier_number = indexOfTier.number
 55 | endif
 56 | 
 57 | if orttier_name$ <> ""
 58 | @indexOfTier: orttier_name$
 59 | 	orttier_number = indexOfTier.number
 60 | 	if indexOfTier.inserted == 1
 61 | 		if phonetier_name$ <> ""
 62 | 			phonetier_number = phonetier_number + 1
 63 | 		endif
 64 | 	endif
 65 | endif
 66 | 
 67 | if llhtier_name$ <> ""
 68 | @indexOfTier: llhtier_name$
 69 | 	llhtier_number = indexOfTier.number
 70 | 	if indexOfTier.inserted == 1
 71 | 		if phonetier_name$ <> ""
 72 | 			phonetier_number = phonetier_number + 1
 73 | 		endif
 74 | 		if orttier_name$ <> ""
 75 | 			orttier_number = orttier_number + 1
 76 | 		endif
 77 | 	endif
 78 | endif
 79 | 
 80 | if wordtier_name$ <> ""
 81 | @indexOfTier: wordtier_name$
 82 | 	wordtier_number = indexOfTier.number
 83 | 	if indexOfTier.inserted == 1
 84 | 		if phonetier_name$ <> ""
 85 | 			phonetier_number = phonetier_number + 1
 86 | 		endif
 87 | 		if orttier_name$ <> ""
 88 | 			orttier_number = orttier_number + 1
 89 | 		endif
 90 | 		if llhtier_name$ <> ""
 91 | 			llhtier_number = llhtier_number + 1
 92 | 		endif
 93 | 	endif
 94 | endif
 95 | 
 96 | if cantier_name$ <> ""
 97 | @indexOfTier: cantier_name$
 98 | 	cantier_number = indexOfTier.number
 99 | 	if indexOfTier.inserted == 1
100 | 		if phonetier_name$ <> ""
101 | 			phonetier_number = phonetier_number + 1
102 | 		endif
103 | 		if orttier_name$ <> ""
104 | 			orttier_number = orttier_number + 1
105 | 		endif
106 | 		if llhtier_name$ <> ""
107 | 			llhtier_number = llhtier_number + 1
108 | 		endif
109 | 		if wordtier_name$ <> ""
110 | 			wordtier_number = wordtier_number + 1
111 | 		endif
112 | 	endif
113 | endif
114 | 
115 | # Do the actual alignment
116 | system 'pythonex$' align.py tier
117 | 
118 | # Close the editor for more speed
119 | editor: textgrid_object$
120 |     Close
121 | endeditor
122 | 
123 | @insertTableTextGrid: tmpfile$, textgrid_object$, phonetier_name$,
124 | ... orttier_name$, wordtier_name$, cantier_name$, llhtier_name$,
125 | ... phonetier_number, orttier_number, wordtier_number, cantier_number,
126 | ... llhtier_number
127 | 
128 | # Reselect the TextGrid and re-open editor
129 | selectObject: textgrid_object$
130 | plusObject: sound_object$
131 | Edit
132 | 


--------------------------------------------------------------------------------
/book/.chktexrc:
--------------------------------------------------------------------------------
1 | VerbEnvir { texttt }
2 | WipeArg {
3 | 	\texttt:{}
4 | }
5 | 


--------------------------------------------------------------------------------
/book/.gitignore:
--------------------------------------------------------------------------------
 1 | *.eps
 2 | *.aux
 3 | *.fmt
 4 | *.log
 5 | *.out
 6 | *.pdf
 7 | *.toc
 8 | *.bbl
 9 | *.blg
10 | *.dvi
11 | *.mlog
12 | 


--------------------------------------------------------------------------------
/book/Makefile:
--------------------------------------------------------------------------------
 1 | DOC:=book
 2 | LATEX:=latex
 3 | BIBTEX:=bibtex
 4 | LATEXFLAGS:=-file-line-error -halt-on-error -no-shell-escape
 5 | 
 6 | TEXS:=$(wildcard *.tex)
 7 | 
 8 | .PHONY: all
 9 | .SECONDARY: $(DOC).fmt
10 | 
11 | all: $(DOC).pdf
12 | 
13 | %.fmt: preamble.tex
14 | 	$(LATEX) $(LATEXFLAGS) -ini -jobname="$(basename $@)" "&$(LATEX) $<\dump"
15 | 
16 | %.pdf: %.dvi
17 | 	dvipdfm $<
18 | 
19 | %.dvi: %.tex %.fmt %.bib $(TEXS)
20 | 	$(LATEX) $(LATEXFLAGS) $<
21 | 	grep -q '^\\bibdata{' $(basename $<).aux && $(BIBTEX) $(basename $<) || true
22 | 	$(LATEX) $(LATEXFLAGS) $< | tee $(basename $<).mlog
23 | 	grep -iqF 'rerun' $(basename $<).mlog &&\
24 | 		$(LATEX) $(LATEXFLAGS) $< || true
25 | 
26 | clean:
27 | 	$(RM) $(addprefix $(DOC).,aux log fmt toc bbl blg mlog run.xml out pdf)
28 | 


--------------------------------------------------------------------------------
/book/book.bib:
--------------------------------------------------------------------------------
1 | @inproceedings{schiel1999,
2 | 	address = {San Francisco},
3 | 	title = {Automatic phonetic transcription of non-prompted speech},
4 | 	booktitle = {Proc. of the {ICPhS} 1999},
5 | 	author = {Schiel, Florian},
6 | 	year = {1999},
7 | 	pages = {607--610},
8 | }
9 | 


--------------------------------------------------------------------------------
/book/book.tex:
--------------------------------------------------------------------------------
   1 | %&book
   2 | \begin{document}
   3 | \cleardoublepage%
   4 | \maketitle
   5 | \setcounter{page}{1}
   6 | \tableofcontents
   7 | \chapter{Introduction}
   8 | \section{Introduction}
   9 | Praatalign is a plug-in for Praat that can be used to do forced phonetic
  10 | alignment on speech signals and in particular free speech. Praatalign combines
  11 | the powerful HTK toolkit and the well trained MAUS models with the
  12 | interactivity and modularity of Praat to create an interactive, intuitive and
  13 | extendible application. Text in \texttt{monospace} means that the word is a
  14 | command, variable or value. Praatalign has the capability to work with
  15 | python-programmed phonetizers, practical orthography, dictionaries and
  16 | pronunciation rules.
  17 | 
  18 | \chapter{Installation}
  19 | \section{Preparation}
  20 | The installation of the program is very straightforward, however installing the
  21 | dependencies might not be on some systems. Some programs are not included in
  22 | the package due to licencing and environment compatibility but they do need to
  23 | be installed in order for the plug-in to work properly. All programs Praatalign
  24 | depends on are free and open source. The following list of programs need to be
  25 | installed with their installation instructions.
  26 | \begin{itemize}
  27 | 	\item \textbf{Praat}\\
  28 | 		Praat is a program that allows you to do phonetic analysis and annotations
  29 | 		with a computer and Praatalign uses Praat to provide an interactive user
  30 | 		interface to the annotated sound files.
  31 | 		\begin{itemize}
  32 | 			\item \textbf{Windows}\\
  33 | 				You can download and install Praat with the instructions on
  34 | 				\url{http://www.fon.hum.uva.nl/praat/download_win.html}.
  35 | 			\item \textbf{Mac}\\
  36 | 				You can download and install Praat with the instructions on
  37 | 				\url{http://www.fon.hum.uva.nl/praat/download_mac.html}.
  38 | 			\item \textbf{Linux and other *NIX}\\
  39 | 				In most of the cases the standard package manager coming with the
  40 | 				Linux distribution also Praat. However this is usually an old version
  41 | 				so use with care. You can download and install the latest Praat with
  42 | 				the instructions on
  43 | 				\url{http://www.fon.hum.uva.nl/praat/download_linux.html}.
  44 | 		\end{itemize}
  45 | 	\item \textbf{Python}\\
  46 | 		Python is used to interpret the scripts that run the core of the aligner.
  47 | 		The script uses specifically Python 2. Python 3 is not supported. All
  48 | 		scripts are tested with Python $2.7.x$, older version might work.
  49 | 		\begin{itemize}
  50 | 			\item \textbf{Windows}\\
  51 | 				You can download and install from the Windows MSI installer found on
  52 | 				this page under the Latest Python 2 Release link
  53 | 				\url{https://www.python.org/downloads/windows/}.
  54 | 			\item \textbf{Mac}\\
  55 | 				Mac OS X 10.8 comes with Python $2.7$ preinstalled. If you run an
  56 | 				obscure version without Python you can download it on 
  57 | 				\url{https://www.python.org/downloads/mac-osx/}.
  58 | 			\item \textbf{Linux and other *NIX}\\
  59 | 				In almost all cases the standard package manager comes with a
  60 | 				sufficiently up to date Python version and most of the distributions
  61 | 				have Python preinstalled. If you work with an obscure system that does
  62 | 				not have a package manager or the Python version in it is obselete you
  63 | 				can download Python on \url{https://www.python.org/downloads/source/}.
  64 | 		\end{itemize}
  65 | 	\item \textbf{SoX}\\
  66 | 		For processing the sound files in a very detailed and controlled way we
  67 | 		use SoX. Although Praat also has sound processing capabilities SoX works
  68 | 		better is some situations, this is because Praat does not allow you to
  69 | 		specify certain options like sampling rate for all formats.
  70 | 		\begin{itemize}
  71 | 			\item \textbf{Windows}\\
  72 | 				You can download SoX on
  73 | 				\url{http://sourceforge.net/projects/sox/files/sox/}. Either download
  74 | 				the executable and run it or unzip the zipfile.
  75 | 			\item \textbf{Mac}\\
  76 | 				You can download SoX on
  77 | 				\url{http://sourceforge.net/projects/sox/files/sox/}. Download the
  78 | 				zipfile and drag the contects to you \texttt{Applications}. Another
  79 | 				option is to use homebrew and type \texttt{brew install sox} in a
  80 | 				terminal.
  81 | 			\item \textbf{Linux and other *NIX}\\
  82 | 				In almost all cases the standard package manager comes with a
  83 | 				sufficiently up to date SoX version. If you work with an obscure system
  84 | 				that does not have a package manager or the SoX version in it is
  85 | 				obselete it on 
  86 | 				\url{http://sourceforge.net/projects/sox/files/sox/14.4.2/}
  87 | 		\end{itemize}
  88 | 	\item \textbf{HCopy \& HVite}\\
  89 | 		HCopy and HVite are programs from the HTK toolkit and due to licencing
  90 | 		issues we can not provide the binaries in a direct way. The program is for
  91 | 		free but you are not allowed to distribute it.
  92 | 		\begin{itemize}
  93 | 			\item \textbf{Windows}\\
  94 | 				You can download a zip file containing precompiled Windows executables
  95 | 				on \url{http://htk.eng.cam.ac.uk/ftp/software/htk-3.3-windows-binary.zip}.
  96 | 				Just unzip them and make sure to add them to \texttt{\%PATH\%} or to
  97 | 				point Praatalign to it in the setup screen.
  98 | 			\item \textbf{Mac, Linux and other *NIX}\\
  99 | 				Installing HCopy and HVite is probably the hardest on Linux and Mac
 100 | 				since you need to compile the binaries yourself. You can find the
 101 | 				latest version on \url{http://htk.eng.cam.ac.uk/ftp/software/}.
 102 | 				Download the zipfile, extract the zipfile and go to the folder with
 103 | 				your terminal. While you can compile the entire toolkit, Praatalign is
 104 | 				only interested in HCopy and HVite. Thus the following commands
 105 | 				suffice:
 106 | 				\begin{lstlisting}[language=bash]
 107 | $ ./configure --disable-hlmtools --disable-hslab
 108 | $ make -j4 htktools
 109 | 				\end{lstlisting}
 110 | 				When the compilation has succeeded you can either add the binary
 111 | 				directory \texttt{HTKTools} to \texttt{\$PATH} or point Praatalign to
 112 | 				the binaries in the setup screen.
 113 | 		\end{itemize}
 114 | \end{itemize}
 115 | 
 116 | \section{Installation}\label{sec:installation}
 117 | The installation of the plug-in is very easy but the method differs for
 118 | different systems.\footnote{The plugin is tested on Windows 7, Windows Server
 119 | 2008 via Windows Terminal Services, Windows 10, Linux and Mac. Other versions
 120 | or other operating systems might also work but are untested.}
 121 | 
 122 | \subsection{Automated installation}
 123 | Run the installation script for your system.
 124 | \begin{itemize}
 125 | 	\item \textbf{Mac, Linux or *NIX}: \texttt{install.sh}\\
 126 | 		Depending on the operating system you either have to run the script from
 127 | 		the terminal or double click it from some explorer like program. Running
 128 | 		the script from the terminal is very easy and preferred. Just start a
 129 | 		terminal program and type the full and exact path in the terminal and press
 130 | 		enter. For example on a Mac this will something like be:\\
 131 | 		\texttt{/Users/frobnicator/Downloads/praatalign/install.sh}.
 132 | 	\item \textbf{Windows}: \texttt{install\_win.bat}\\
 133 | 		This installation script can be double clicked from the explorer and it
 134 | 		will install Praatalign. When the installation is finished you can press
 135 | 		Enter to make the windows disappear.
 136 | \end{itemize}
 137 | 
 138 | \subsection{Manual installation}
 139 | Copy the contents of the root directory to (you have to create the directory if
 140 | it does not already exist):
 141 | \begin{itemize}
 142 | 	\item \textbf{Linux and *NIX}\\
 143 | 		\texttt{\$\{HOME\}/.praat-dir/plugin\_pralign/}.
 144 | 	\item \textbf{Windows}\\
 145 | 		\texttt{\%USERPROFILE\%\textbackslash{} Praat\textbackslash{} plugin\_pralign/}
 146 | 	\item \textbf{Mac}\\
 147 | 		\texttt{\$\{HOME\}/Library/Preferences/Praat Prefs/plugin\_pralign/}
 148 | \end{itemize}
 149 | 
 150 | \chapter{Documentation}
 151 | \section{General information}
 152 | With the Praatalign plugin you can currently align out of the box the data
 153 | using Spanish and Dutch acoustic models from
 154 | MAUS~\footnote{\url{http://www.bas.uni-muenchen.de/Bas/BasMAUS.html}} created
 155 | by Schiel et al.~\cite{schiel:999}. These models are already included in the
 156 | package. %Other models from MAUS can be easily added and will be in the future.
 157 | %If you like to have a language added please contact us.
 158 | %Presets for Australian English, Estonian, German, Hungarian, Italian, New
 159 | %Zealand English Polish and Portuguese are available at minimum.
 160 | 
 161 | Dictionary, ruleset, universal phonetizer all other files are, and should be,
 162 | encoded in \texttt{UTF-8}. To enforce this the plug-in changes the default
 163 | behaviour of Praat every time Praat loads to make sure Praats reading and
 164 | writing preferences are set to \texttt{UTF-8}.
 165 | 
 166 | When the plug-in is successfully installed several menu items are added in the
 167 | \texttt{TextGrid} editor under the \texttt{Interval} menu. The added
 168 | functionality only works when you are editing a \texttt{TextGrid} and a
 169 | \texttt{LongSound} or a \texttt{Sound}\footnote{Sound files are written to disk
 170 | prior to alignment, thus \texttt{Longsound} is preferred}. Currently the plugin
 171 | is only tested on \texttt{WAVE} files. It should however work on all sound
 172 | filetypes SoX can detect from the extension.
 173 | 
 174 | \section{Menu items}
 175 | Almost all menu items will fail when there is no settings file present. The
 176 | settings file has to be created by running \texttt{Set up forced
 177 | alignment\ldots} interactively or by running \texttt{settings\_ni.praat} in a
 178 | script.
 179 | 
 180 | \subsection{\texttt{Generate dictionary from tier}}
 181 | This functions allows the user to generate a dictionary containing all the
 182 | missing or unphonetizable words from the currently selected tier using the
 183 | current settings the plugin is initialized with. The plugin will prompt you
 184 | after pressing the button for a location for the dictionary file. When this
 185 | process is done the user can add the pronunciations after every entry that
 186 | is found in the skeleton dictionary. Note that there is no sanitation
 187 | applied on the words. This means that if the phonetizer removes punctuation
 188 | it can still be present in the dictionary.
 189 | 
 190 | \subsection{\texttt{Clean selection}}
 191 | This function is a helper function to clean up old or wrong alignment. When the
 192 | function runs all annotation data within the selection within the selected tier
 193 | will be removed. 
 194 | 
 195 | \emph{Note that this is not necessary to do before an alignment because this
 196 | function runs by default before any alignment.}
 197 | 
 198 | \subsection{\texttt{Align current interval}}
 199 | This function aligns the current selected interval on the current selected
 200 | tier. When selecting a small interval it should not take much time at all. When
 201 | you select an interval from an output tier (phone, ort, word, canonical or log)
 202 | the function will prompt you to make sure this is what you want to do. This
 203 | functions first clears out the annotations on the output tiers.
 204 | 
 205 | \subsection{\texttt{Align current tier}}
 206 | This function aligns the entire selected tier. Aligning an entire tier can take
 207 | some time, especially when you have a lot of pronunciation variants. This
 208 | functions first clears out the annotations on the output tiers.
 209 | 
 210 | \subsection{\texttt{Set up forced alignment\ldots}}
 211 | This functions spawns two option menus that will create the necessary settings
 212 | file. When you finish the forms a settings file is written to disk.
 213 | 
 214 | \subsubsection{Basic options}	
 215 | The first form contains all the basic settings needed for alignment. It also
 216 | shows the version of the plugin on the first line. The following options can be
 217 | entered. All settings are analogous to the names of the settings in the
 218 | \texttt{settings\_ni.praat} script.
 219 | \begin{itemize}
 220 | 	\item \texttt{new}: %
 221 | 		Name of the output tier storing the phone level alignment
 222 | 
 223 | 		In this option you specify the name of the tier where the phone level
 224 | 		alignment is stored, this can be either an existing tier or a non existing
 225 | 		tier. If the tier does not exist it will be created upon doing the first
 226 | 		alignment. When you leave this field empty no phone level tier will be
 227 | 		created.
 228 | 	\item \texttt{ort}: %
 229 | 		Name of the output tier storing the word level orthographic alignment
 230 | 
 231 | 		In this option you specify the name of the tier where the word level
 232 | 		orthographic alignment is stored, this can be either an existing tier
 233 | 		or a non existing tier. If the tier does not exist it will be created
 234 | 		upon doing the first alignment. When you leave this field empty no
 235 | 		word level orthographic level tier will be created.
 236 | 	\item \texttt{wrd}: %
 237 | 		Name of the output tier storing the word level alignment
 238 | 
 239 | 		In this option you specify the name of the tier where the word level
 240 | 		alignment is stored, this can be either an existing tier or a non existing
 241 | 		tier. If the tier does not exist it will be created upon doing the first
 242 | 		alignment. When you leave this field empty no word level tier will be
 243 | 		created.
 244 | 	\item \texttt{can}: %
 245 | 		Name of the output tier storing the canonical pronunciation
 246 | 
 247 | 		In this option you specify the name of the tier where the canonical
 248 | 		pronunciation of every word is stored, this can be either an existing tier
 249 | 		or a non existing tier. If the tier does not exist it will be created upon
 250 | 		doing the first alignment. When you leave this field empty no canonical
 251 | 		pronunciation tier will be created.
 252 | 	\item \texttt{llh}: %
 253 | 		Name of the output tier storing the log likelihood
 254 | 
 255 | 		In this option you specify the name of the tier where the log likelihood of
 256 | 		every phone is stored, this can be either an existing tier or a non
 257 | 		existing tier. If the tier does not exist it will be created upon doing the
 258 | 		first alignment. When you leave this field empty no log likelihood tier
 259 | 		will be created.
 260 | 	\item \texttt{model}: %
 261 | 		Select model
 262 | 
 263 | 		In this option you specify which acoustic model to use. More info about the
 264 | 		models can be found in Section~\ref{sec:models}.
 265 | 	\item \texttt{lan}: %
 266 | 		Select model
 267 | 
 268 | 		In this option you specify which phonetizer to use. More info about the
 269 | 		models can be found in Section~\ref{sec:phonetizers}.
 270 | 
 271 | 		\emph{Note that if you select the universal phonetizer you will be prompted
 272 | 		to select the universal phonetizer file. More about the universal
 273 | 		phonetizer in Section~\ref{sec:univphonetizer}}
 274 | 	\item \texttt{dic}, \texttt{dictionary}: %
 275 | 		Select a dictionary file
 276 | 
 277 | 		In this option you can specify a dictionary file. When you tick the box you
 278 | 		will be prompted to select a dictionary file. When a dictionary is already
 279 | 		selected the \texttt{dictionary} option is added which contains the path to
 280 | 		the file. When you want to switch to using no dictionary you can clear that
 281 | 		box and leave the \texttt{dic} box unticked.
 282 | 	\item \texttt{rul}, \texttt{ruleset}: %
 283 | 		Select a ruleset file
 284 | 
 285 | 		In this option you can specify a ruleset file. When you tick the box you
 286 | 		will be prompted to select a ruleset file. When a ruleset is already
 287 | 		selected the \texttt{ruleset} option is added which contains the path to
 288 | 		the file. When you want to switch to using no ruleset you can clear that
 289 | 		box and leave the \texttt{rul} box unticked.
 290 | 	\item \texttt{pho}, \texttt{phonetizer}: %
 291 | 		Select a universal phonetizer file
 292 | 
 293 | 		This option will only appear if a phonetizer file has been set
 294 | 		previously. When no phonetizer file has ever been set and the universal
 295 | 		phonetizer is used you will be prompted for it anyways since it is
 296 | 		mandatory. When you tick the box you
 297 | 		will be prompted to select a phonetizer file. The current phonetizer
 298 | 		file is shown in \texttt{phonetize}
 299 | 	\item \texttt{thr}: %
 300 | 		Set the size to add to the annotations
 301 | 
 302 | 		In this option you can specify an extra margin used for every annotation.
 303 | 		When the annotations are placed to close to the real sound the initial
 304 | 		pause can clobber up the beginning of speech and that can reduce the
 305 | 		performance. Setting the \texttt{thr} value to $0.1$ will for example increase
 306 | 		all boundaries from annotations with $100$ms. Note that this does not
 307 | 		change the original annotation and it will only increase the widen the
 308 | 		annotation when there is room to do so, meaning that it will not create
 309 | 		overlap with other annotations.
 310 | \end{itemize}
 311 | 
 312 | \subsubsection{Advanced options}	
 313 | The second form contains all the more advanced settings needed for alignment.
 314 | It shouldn't be necessary to change these options regularly. All settings are
 315 | analogous to the names of the settings in the \texttt{settings\_ni.praat}
 316 | script.
 317 | 
 318 | \begin{itemize}
 319 | 	\item \texttt{log}: %
 320 | 		Set a location for the logfile
 321 | 
 322 | 		In this option you can specify a location to write a debug log to. When you
 323 | 		want to switch to not using a logfile you can redirect the log to either
 324 | 		\texttt{/dev/null} on Linux, Mac and other *NIX systems and \texttt{nul} on
 325 | 		Windows.
 326 | 	\item \texttt{sox}: %
 327 | 		Set a SoX executable
 328 | 
 329 | 		In this option you can specify a SoX executable. When you tick the box you
 330 | 		will be prompted to select a SoX executable. When a SoX executable is
 331 | 		already selected the \texttt{soxex} option is added which contains the path
 332 | 		to the executable. When you want to switch to using the SoX executable in
 333 | 		\texttt{PATH} you can clear that box leave the \texttt{sox} box unticked.
 334 | 
 335 | 		\begin{itemize}
 336 | 			\item \textbf{Windows}\\
 337 | 				If you have installed sox using the MSI you can find \texttt{sox.exe}
 338 | 				in \texttt{C:\textbackslash%
 339 | 				Program Files (x86)\textbackslash{} sox{-}14{-}4{-}1} or
 340 | 				\texttt{C:\textbackslash{} Program Files\textbackslash{}
 341 | 				sox{-}14{-}4{-}1}.
 342 | 				If you just downloaded the zip file you can just point to the location
 343 | 				you extracted the archive and select \texttt{sox.exe}.
 344 | 			\item \textbf{Mac}\\
 345 | 				If you dragged sox to the \texttt{Applications} you can find it there
 346 | 				and you can just point to the sox executable. If you installed sox via
 347 | 				homebrew it is probably already in \texttt{\$PATH}. If this is not the
 348 | 				case you can find the location by typing in a terminal \texttt{which
 349 | 				sox} and pointing Praatalign to that location.
 350 | 			\item \textbf{Linux and other *NIX}\\
 351 | 				If you have installed sox using a package manager it probably already
 352 | 				is in your \texttt{\$PATH}. If this is not the case you can find the
 353 | 				location by typing in a terminal: \texttt{which sox} and pointing
 354 | 				Praatalign to location.
 355 | 		\end{itemize}
 356 | 	\item \texttt{hvite}: %
 357 | 		Set a HVite executable
 358 | 
 359 | 		In this option you can specify a HVite executable. When you tick the box you
 360 | 		will be prompted to select a HVite executable. When a HVite executable is
 361 | 		already selected the \texttt{hviteex} option is added which contains the
 362 | 		path to the executable. When you want to switch to using the HVite
 363 | 		executable in \texttt{PATH} you can clear that box leave the \texttt{hvite}
 364 | 		box unticked.
 365 | 
 366 | 		\begin{itemize}
 367 | 			\item \textbf{Windows}\\
 368 | 				Point to the directory where you unzipped the file
 369 | 				from HTK and select \texttt{HVite.exe}.
 370 | 			\item \textbf{Mac, Linux, and other *NIX}\\
 371 | 				Point to the directory where you compiled the tools from HTK and
 372 | 				select \texttt{HVite}. It resides in the \texttt{HTKTools}
 373 | 				directory.
 374 | 		\end{itemize}
 375 | 	\item \texttt{hcopy}: %
 376 | 		Set a HCopy executable
 377 | 
 378 | 		In this option you can specify a HCopy executable. When you tick the box you
 379 | 		will be prompted to select a HCopy executable. When a HCopy executable is
 380 | 		already selected the \texttt{hcopyex} option is added which contains the
 381 | 		path to the executable. When you want to switch to using the HCopy
 382 | 		executable in \texttt{PATH} you can clear that box leave the \texttt{hcopy}
 383 | 		box unticked.
 384 | 
 385 | 		\begin{itemize}
 386 | 			\item \textbf{Windows}\\
 387 | 				Point to the directory where you unzipped the file
 388 | 				from HTK and select \texttt{HCopy.exe}.
 389 | 			\item \textbf{Mac, Linux, and other *NIX}\\
 390 | 				Point to the directory where you compiled the tools from HTK and
 391 | 				select \texttt{HCopy}. It resides in the \texttt{HTKTools}
 392 | 				directory.
 393 | 		\end{itemize}
 394 | 	\item \texttt{python}: %
 395 | 		Set a Python executable
 396 | 
 397 | 		In this option you can specify a Python executable. When you tick the box
 398 | 		you will be prompted to select a Python executable. When a Python
 399 | 		executable is already selected the \texttt{pythonex} option is added which
 400 | 		contains the path to the executable. When you want to switch to using the
 401 | 		Python executable in \texttt{PATH} you can clear that box leave the
 402 | 		\texttt{python} box unticked.
 403 | 
 404 | 		\begin{itemize}
 405 | 			\item \textbf{Windows}\\
 406 | 				Python can usually be found in \texttt{C:\textbackslash{} Python27}.
 407 | 				From there you can select \texttt{python.exe}
 408 | 			\item \textbf{Mac, Linux, and other *NIX}\\
 409 | 				If you have installed sox using a package manager it probably
 410 | 				already is in your \texttt{\$PATH}. If this is not the case you can
 411 | 				find the location by typing in a terminal: \texttt{which python}
 412 | 				and pointing Praatalign to that location. Note that in some systems
 413 | 				\texttt{python} symlinks to \texttt{python3}, in that case point
 414 | 				Praatalign to \texttt{python2}. If you still can not find the
 415 | 				executable you can try using the search function in the file
 416 | 				manager.
 417 | 		\end{itemize}
 418 | \end{itemize}
 419 | 
 420 | \section{Dictionary}
 421 | To phonetize words Praatalign either uses the provided phonetizer or a
 422 | dictionary. Dictionaries are plain text files that contain words and one or
 423 | more pronunciations. A dictionary file is a \texttt{UTF-8} encoded file
 424 | containing non-empty lines separated by a newline character\footnote{On Mac,
 425 | Linux, and *NIX this is default, on Windows this can cause problems. When using
 426 | Praatalign on windows please refrain to a text editor that has newline
 427 | capabilities like Notepad++}. Lines starting with a \texttt{\#} will be ignored
 428 | and can thus be used as comments. The format of a dictionary entry is a word
 429 | followed by a tab followed by tab separated pronunciations. An example
 430 | dictionary can be found in Listing~\ref{lst:exampledictionary}
 431 | 
 432 | \begin{lstlisting}[caption={Example dictionary},label={lst:exampledictionary}]
 433 | # This is comment
 434 | # This is a word with two possible pronunciations
 435 | ado<TAB>a d o<TAB>a o
 436 | # These are words with one possible pronunciation
 437 | empatar<TAB>e m p a t a r
 438 | empataran<TAB>e m p a t a r a n
 439 | \end{lstlisting}
 440 | 	
 441 | \section{Ruleset}
 442 | Besides generating pronunciation by using the dictionary and phonetization you
 443 | can also use rulesets to define pronunciation variants. Ruleset make you able
 444 | to define general rules applied over all words (phonetized words and dictionary
 445 | words). In this way you can easily define for example deletion rules.
 446 | A ruleset file is a \texttt{UTF-8} encoded file containing non-empty lines
 447 | separated by a newline character. Lines starting with a \texttt{\#} will be
 448 | ignored and can thus be used as comments. There are two ways of defining rules
 449 | for a ruleset.
 450 | \begin{itemize}
 451 | 	\item \textbf{Simple}\\
 452 | 		Simple rules are just find and replace queries. The first column is the
 453 | 		target and the second column is the replace value. For example the deletion
 454 | 		rule \texttt{a d o -> a o} can be written as \texttt{a d o<TAB>a o}.
 455 | 	\item \textbf{Regular}\\
 456 | 		Regular rules are like regular expressions and are therefore much more
 457 | 		expressive but also more complicated to write. Regular rules start with a
 458 | 		single tab character to denote that they are in fact regular expressions.
 459 | 		For example a deletion rule that deletes a \texttt{d} between two vowels
 460 | 		and potentially also across word boundaries can be written like
 461 | 		\verb|<TAB>([aouie] #?) d ([aouie])<TAB>\1\2|
 462 | 		Internally we use the \texttt{re.sub} function from the Python \texttt{re}
 463 | 		library\footnote{\url{https://docs.python.org/2/library/re.html}}. Besides
 464 | 		that there are some extra shortcuts and options:
 465 | 		\begin{itemize}
 466 | 			\item \verb|\v| for vowels (\verb|[aoeiu]|)
 467 | 			\item \verb|\c| for consonants (\verb|[~aoeiu]|)
 468 | 			\item \verb|#| for a inter word silence
 469 | 		\end{itemize}
 470 | \end{itemize}
 471 | 
 472 | 
 473 | \section{Phonetizers}\label{sec:phonetizers}
 474 | \subsection{Spanish}
 475 | The Spanish phonetizer is designed only to work with the spanish models. It
 476 | removes a lot of non speech annotated symbols and does some tricks to get
 477 | exceptions well phonetized. It can be seen as an example of writing an advanced
 478 | phonetizer in Python.
 479 | 
 480 | %\subsection{Tzeltal}
 481 | %The Tzeltal phonetizer is an example of how to use the SAMPA models to align a
 482 | %new language. Thus it only works for the SAMPA models. It removes some non
 483 | %speech annotated symbols and does an almost literal character to character
 484 | %translation.
 485 | 
 486 | \subsection{Universal}\label{sec:univphonetizer}
 487 | When you select the universal phonetizer you will be prompted to point the
 488 | plugin to an universal phonetizer file.
 489 | A universal phonetizer file is a \texttt{UTF-8} encoded file containing
 490 | non-empty lines separated by a newline character. Every line contains a
 491 | translation from practical orthography to phonetic transcription and the order
 492 | of appearance in the file is the order of importance in the phonetizer. When a
 493 | word gets phonetized the phonetizer tries to match the first rule in the
 494 | phonetizer file. For example see the start of an example file listed in
 495 | Listing~\ref{lst:exampleuniversalphonetizer} representing a translation from
 496 | spanish orthography to phonetic transcription. The safest order is always the
 497 | order in which the biggest sections are topmost in the file.
 498 | 
 499 | \begin{lstlisting}[caption={Example universal phonetizer file},
 500 | 	label={lst:exampleuniversalphonetizer},
 501 | 	literate={ü}{{\"u}}1 {ñ}{{\~n}}1 {ç}{{\c{c}}}1 {í}{{\'e}}1]
 502 | gue<TAB>g
 503 | gui<TAB>g
 504 | ch<TAB>t S
 505 | ce<TAB>T
 506 | ci<TAB>T
 507 | cí<TAB>T
 508 | gü<TAB>g u
 509 | ll<TAB>jj
 510 | qu<TAB>k
 511 | ñ<TAB>J
 512 | ç<TAB>T
 513 | j<TAB>x
 514 | c<TAB>k
 515 | v<TAB>b
 516 | w<TAB>b
 517 | z<TAB>T
 518 | y<TAB>j
 519 | q<TAB>k
 520 | ...
 521 | \end{lstlisting}
 522 | 
 523 | \subsection{None}
 524 | The None phonetizer is a dummy phonetizer that does nothing. This means that
 525 | every word should be available in the dictionary.
 526 | 
 527 | \section{Scriptability and batch processing}
 528 | Note that this section is not updated as often as it should. Always check the
 529 | exact format in \texttt{settings\_ni.praat}.
 530 | 
 531 | \subsection{Non interactive settings file creation}
 532 | Although the Praatalign script is inherently interactive it is still possible
 533 | to batch process corpora using simple praat scripts. To facilitate this
 534 | function a file called \texttt{\$DIR/settings\_ni.praat} can be run where
 535 | \texttt{\$DIR} is the location of the plugin files. The location of the plugin
 536 | files for your operating system can be found in Section~\ref{sec:installation}
 537 | in the manual installation section. The \texttt{settings\_ni.praat} is a
 538 | stripped down version of the settings dialog present in the aligner. Since it
 539 | is using a praat form to ask for the user input, in contrary to the pause
 540 | dialog in the normal settings scirpt, it can be run non interactively by
 541 | running the script from a praat script.
 542 | 
 543 | For example if you want to setup the aligner to align a spanish file with all
 544 | custom values on linux under the user frobnicator you can put this in your
 545 | script to setup the aligner:
 546 | 
 547 | \begin{lstlisting}
 548 | runScript: "/home/frobnicator/.praat-dir/plugin_pralign/settings_ni.praat",
 549 | ..."phon", "wrd", "can", "llh", "spa", "spanish", "None", "/path/to/dict",
 550 | ..."/path/to/ruleset", 0, "/some/path/to/logfile",
 551 | ..."/usr/bin/sox", "/usr/bin/HVite", "/usr/bin/HCopy", "python"
 552 | \end{lstlisting}
 553 | 
 554 | Note that due to the lack of interactivity the format is a little bit
 555 | different. The differences are:
 556 | \begin{itemize}
 557 | 	\item \texttt{lan} must be the language code as in the interactive settings.
 558 | 	\item \texttt{mod} must be the model code as in the interactive settings.
 559 | 	\item \texttt{pho} must be the path to the universal phonetizer. When you do
 560 | not want to use a universal phonetizer you must use \texttt{None}.
 561 | 	\item \texttt{dic} and \texttt{rul} must be the full path, when you do not
 562 | want to use a dictionary or ruleset you must use \texttt{None}.
 563 | 	\item \texttt{sox}, \texttt{hvb}, \texttt{hcb}, and \texttt{py} must be the
 564 | full path, when you do not want a custom location you must use respectively
 565 | \texttt{sox}, \texttt{HVite}, \texttt{HCopy} and \texttt{python}
 566 | \end{itemize}
 567 | 
 568 | \subsection{Example}
 569 | When you then open a \texttt{TextGrid} and a \texttt{LongSound} file and do
 570 | \texttt{View \& Edit} to open the editor you can run the alignment from the
 571 | script using the button text as function. For example the script could look
 572 | like the script in Listing~\ref{lis:scriptab}.
 573 | 
 574 | \begin{lstlisting}[caption={Example scriptability},label={lis:scriptab}]
 575 | # We assume the LongSound and TextGrid are selected previously
 576 | 
 577 | # Generate the settings file
 578 | 	runScript: "/home/frobnicator/.praat-dir/plugin_pralign/settings_ni.praat",
 579 | ..."phon", "wrd", "can", "llh", "spa", "spanish", "None", "/path/to/dict",
 580 | ..."/path/to/ruleset", 0, "/some/path/to/logfile",
 581 | ..."/usr/bin/sox", "/usr/bin/HVite", "/usr/bin/HCopy", "python"
 582 | 
 583 | # Spawn the editor
 584 | View & Edit
 585 | 
 586 | # Open the editor
 587 | editor: "TextGrid " + objectname$
 588 | 	# This bit of code is a small snippet to select a specific tier with the
 589 | 	# index: tiernum, tiernum is obtained by querying all tiers outside the
 590 | 	# editor and finding the tier that matches the name
 591 | 	currenttiernum = -1
 592 | 	while currenttiernum <> tiernum
 593 | 		Select next tier
 594 | 		inf$ = Editor info
 595 | 		currenttiernum = extractNumber(inf$, "Selected tier: ")
 596 | 	endwhile
 597 | 	
 598 | 	# Do the actual alignment
 599 | 	Align current tier
 600 | 	# When this is done aligned data can be found in: custom_phone_tier and 
 601 | 	# custom_word_tier.
 602 | endeditor
 603 | \end{lstlisting}
 604 | 
 605 | \section{Troubleshooting}
 606 | \begin{itemize}
 607 | 	\item \textbf{Some words are ignored and thus not aligned}\\
 608 | 		Some phonetizers phonetize unphonetizable words into an empty word to avoid
 609 | 		throwing exceptions. When words are not phonetized it means that it is not
 610 | 		in the dictionary nor phonetizable. To fix this you can edit the phonetizer
 611 | 		or add the word to the dictionary.
 612 | 	\item Pop-up stating: \textbf{Error! check the text window for details\ldots}\\
 613 | 		This means something went wrong in the python script. Check the info
 614 | 		window. Usually it is a missing binary, ruleset etc. It could also be
 615 | 		that you used a phone that does not exist in the model
 616 | 	\item Pop-up stating: \textbf{Unknown IO error}\\
 617 | 		Some IO error occurred of an unknown type. Please check the logfile.
 618 | 	\item Log stating: \texttt{sox FAIL trim: Position 1 is behind the
 619 | 		following position.}
 620 | 		You are aligning an annotation that lies outside the wave file.
 621 | 	\item \textbf{Other errors}\\
 622 | 		When the plugin crashes without any reason you should enable logging in the
 623 | 		settings menu to see where it crashes. If the problem is not solvable
 624 | 		please file a bugreport via
 625 | 		github~\footnote{\url{https://github.com/dopefishh/praatalign/issues}} or
 626 | 		contact us directly via e-mail.
 627 | \end{itemize}
 628 | 
 629 | \chapter{Extending Praatalign}
 630 | \section{Introduction}
 631 | %Extending the aligner with new languages should be very easy for languages that
 632 | %can be mapped on the current SAMPA model or on any other existing model (maus
 633 | %model).
 634 | Adding a language with a new model could be possible but no support will be
 635 | given, however you can always try, you can even try getting help. Adding a
 636 | language requires a couple of components that need to be written or adapted.
 637 | 
 638 | \section{Phonetizer}
 639 | Phonetization of your language is the most elegant solution of translating the
 640 | graphemes to phonemes. Implementing a phonetizer is as easy as implementing one
 641 | function called \texttt{phonetizeword}. A skeleton class can be found in
 642 | \texttt{phonetizer.py}. The function in the skeleton class is accompanied by
 643 | comments. A phonetized utterance is always of the following form:\\
 644 | \texttt{utt=[word1, word2, \ldots, wordn]}, 
 645 | \texttt{word=[pron1, pron2, \ldots, pronn]} and
 646 | \texttt{pron=[phone1, phone2, \ldots, phonen]} and every phone is a string.
 647 | So if you want to use the skeleton class with the \texttt{phonetizeword}
 648 | function you need to return a list of lists of strings where every
 649 | string is a phone from the model. If you also want to do utterance based
 650 | translation you need to return a list of lists of lists of strings.
 651 | 
 652 | \section{Dictionary}
 653 | If you do not want to use a phonetizer you can also suffice with only using a
 654 | dictionary based translation. Dictionary based translation still needs to be
 655 | loaded as a phonetizer though. All phonetizers include also a dictionary based
 656 | lookup. In the \texttt{phonetizer.py} a dictionary phonetizer is already
 657 | present called \texttt{PhonetizerDictionary}. There is also a loopback
 658 | phonetizer that takes the literal annotation as transcription. This phonetizer
 659 | is currently not used but could be used when an exact phonetic translation is
 660 | already available.
 661 | 
 662 | \section{Adding the language to the aligner}
 663 | When you have the translation from grapheme to phoneme the only thing that
 664 | needs to be done is adding it to the script files.
 665 | \begin{itemize}
 666 | 	\item \texttt{phonetizer.py}\\
 667 | 		On the bottom of this file there is a dictionary containing all the
 668 | 		translations from language code to phonetizer and parameters directory. You
 669 | 		need to add your language to that dictionary.
 670 | 	\item \texttt{settings.praat}\\
 671 | 		In this file you need to add stuff on multiple locations, namely within the
 672 | 		second \texttt{if} that relies in the first outer \texttt{if} block you
 673 | 		need to add your language with its appropriate position. When you want your
 674 | 		language on top you need to adapt the other numbers too.
 675 | 
 676 | 		Finally within the \texttt{pause} block you need to add your language code
 677 | 		in the \texttt{optionMenu:} block on the same position as specified
 678 | 		earlier.
 679 | \end{itemize}
 680 | When you have changed these files properly your language should be available in
 681 | the menus and work out of the box.
 682 | 
 683 | \chapter{Appendices}
 684 | \section{How to cite}
 685 | \begin{lstlisting}[caption={Bibtex snippet}]
 686 | @misc{praatalign2.0,
 687 | 	author={Lubbers, Mart and Torreira, Francisco},
 688 | 	title={Praatalign: an interactive Praat plug-in for performing phonetic forced alignment},
 689 | 	howpublished={\url{https://github.com/dopefishh/praatalign}},
 690 | 	year={2013-2016},
 691 | 	note={Version 2.0}
 692 | }
 693 | \end{lstlisting}
 694 | 
 695 | \section{Licence}
 696 | \lstinputlisting{../LICENCE}
 697 | 
 698 | \section{Acoustic model phone specifications}\label{sec:models}
 699 | \subsection{Spanish}\label{sec:phspanish}
 700 | The Spanish mapping is an exact mapping with the spanish SAMPA
 701 | phoneset\footnote{\url{http://www.phon.ucl.ac.uk/home/sampa/spanish.htm}}.
 702 | 
 703 | {\small
 704 | \begin{longtable}{l|l|l|l}
 705 | 	\toprule
 706 | 	& Symbol & Word & Transcription\\
 707 | 	\midrule
 708 | 	\multirow{6}{*}{Plosives} &
 709 | 	p & padre & p a D r e\\
 710 |   & b & vino & b i n o\\
 711 |   & t & tomo & t o m o\\
 712 |   & d & donde & d o n d e\\
 713 |   & k & casa & k a s a\\
 714 |   & g & gata & g a t a\\
 715 | 	\midrule
 716 | 	\multirow{2}{*}{Affricatives} &
 717 | 	tS & mucho & m u tS o\\
 718 | 	& jj & hielo & jj e l o\\
 719 | 	\midrule
 720 | 	\multirow{7}{*}{Fricatives} &
 721 | 	f & fácil & f a T i l\\
 722 | 	& B & cabra & k a B r a\\
 723 | 	& T & cinco & T i n k o\\
 724 | 	& D & nada & n a D a\\
 725 | 	& s & sala & s a l a \\
 726 | 	& x & mujer & m u x e r\\
 727 | 	& G & luego & l w e G o\\
 728 | 	\midrule
 729 | 	\multirow{3}{*}{Nasals} &
 730 | 	m & mismo & m i s m o\\
 731 | 	& n & nunca & n u n k a\\
 732 | 	& J & año & a J o\\
 733 | 	\midrule
 734 | 	\multirow{4}{*}{Liquids} & 
 735 | 	l & lejos & l e x o s\\
 736 | 	& L & caballo & k a b a L o\\
 737 | 	& r & puro & p u r o\\
 738 | 	& rr & torre & t o rr e\\
 739 | 	\midrule
 740 | 	\multirow{4}{*}{Semivowels} &
 741 | 	j & rei & rr e j\\
 742 | 	&  & pie & p j e\\
 743 | 	& w & deuda & d e w D a\\
 744 | 	&  & muy & m w i\\
 745 | 	\midrule
 746 | 	\multirow{5}{*}{Vowels} &
 747 | 	i & pico & p i k o\\
 748 | 	& e & pero & p e r o\\
 749 | 	& a & valle & b a L e\\
 750 | 	& o & toro & t o r o\\
 751 | 	& u & duro & d u r o\\
 752 | 	\midrule
 753 | 	\multirow{4}{*}{Special} &
 754 | 	\textless{} & word initial silence & \\
 755 | 	& \textgreater{} & word final silence & \\
 756 | 	& \# & inter word silence & \\
 757 | 	& \textless{}nib\textgreater{} & non speech sound & \\
 758 | 	\bottomrule
 759 | 	\caption{Spanish phone specification}
 760 | \end{longtable}
 761 | }
 762 | 
 763 | \subsection{Dutch}
 764 | The Dutch mapping is an almost\footnote{Derivations are marked in bold} exact
 765 | mapping with the dutch SAMPA phoneset\footnote{\url{
 766 | http://www.phon.ucl.ac.uk/home/sampa/dutch.htm}}
 767 | 
 768 | {\small
 769 | \begin{longtable}{l|l|l|l}
 770 | 	& Praatalign & Word & Transcription\\
 771 | 	\toprule
 772 | 	\multirow{6}{*}{Plosives} &
 773 | 	p & pak & p A k\\
 774 | 	& b & bak & b A k\\
 775 | 	& t & tak & t A k\\
 776 | 	& d & dak & d A k\\
 777 | 	& k & kap & k A p\\
 778 | 	& g & goal & g o: l\\
 779 | 	\midrule
 780 | 	\multirow{9}{*}{Fricatives} &
 781 | 	f & fel & f E l\\
 782 | 	& v & vel & v E l\\
 783 | 	& s & sein & s E i n\\
 784 | 	& z & zijn & z E i n\\
 785 | 	& x & toch & t o x\\
 786 | 	& G & goed & G u t\\
 787 | 	& h & hand & h A n t\\
 788 | 	& Z & bagage & b A g a: Z @\\
 789 | 	& S & show & S o: u\\
 790 | 	\midrule
 791 | 	\multirow{7}{*}{Sonorants} &
 792 | 	m & met & m E t\\
 793 | 	& n & net & n E t\\
 794 | 	& N & bang & b A N\\
 795 | 	& l & land & l A n t\\
 796 | 	& r & rand & r A n t\\
 797 | 	& w & wit & w I t\\
 798 | 	& j & ja & j a:\\
 799 | 	\midrule
 800 | 	\multirow{6}{*}{Checked vowels} &
 801 | 	I & pit & p I t\\
 802 | 	& E & pet & p E t\\
 803 | 	& A & pat & p A t\\
 804 | 	& O & pot & p O t\\
 805 | 	& Y & put & p Y t\\
 806 | 	& @ & gemakkelijk & G @ m A k @ l @ k\\
 807 | 	\midrule
 808 | 	\multirow{10}{*}{Free vowels} &
 809 | 	i & vier & v i r\\
 810 | 	& y & vuur & v y r\\
 811 | 	& u & voer & v u r\\
 812 | 	& a: & naam & n a: m\\
 813 | 	& e: & veer & v e: r\\
 814 | 	& \textbf{P2:} & deur & d P2: r\\
 815 | 	& o: & voor & v o: r\\
 816 | 	& EI & fijn & f EI n\\
 817 | 	& \textbf{P9y} &huis & h P9y s\\
 818 | 	& Au & goud & x Au t\\
 819 | 	\midrule
 820 | 	\multirow{6}{*}{Dipthongs} &
 821 | 	a:i & draai & d r a: i\\
 822 | 	& o:i &mooi & m o: i\\
 823 | 	& ui & roeiboot & r ui b o: t\\
 824 | 	& iu & nieuw & n iu\\
 825 | 	& yu & duw & d yu\\
 826 | 	& e:u & sneeuw & s n e: u\\
 827 | 	\midrule
 828 | 	\multirow{3}{*}{Marginal vowels} &
 829 | 	E:\ & cr\'eme & k r E:\ m\\
 830 | 	& \textbf{P9:} & freule & f r P9: l @\\
 831 | 	& O:\ & roze & r O:\ z @\\
 832 | 	\midrule
 833 | 	\multirow{4}{*}{Special} &
 834 | 	\textless{} & word initial silence & \\
 835 | 	& \textgreater{} & word final silence & \\
 836 | 	& \# & inter word silence & \\
 837 | 	& \textless{} nib\textgreater{} & non speech sound & \\
 838 | 	\bottomrule
 839 | 	\caption{Dutch phone specification}
 840 | \end{longtable}
 841 | }
 842 | 
 843 | %\subsection{English}
 844 | %The English mapping is an exact\footnote{Derivations are marked in bold} exact
 845 | %mapping with the english SAMPA phoneset\footnote{\url{
 846 | %https://www.phon.ucl.ac.uk/home/sampa/english.htm}} with some borrowed phones.
 847 | %
 848 | %Also there is conversion script for the English CMU
 849 | %dictionary\footnote{\url{http://www.speech.cs.cmu.edu/cgi-bin/cmudict}} located
 850 | %in \texttt{./par.eng/cmu2praatalign.py} that converts the CMU dictionary to
 851 | %Praatalign format. The scripts is a Python script and should download the
 852 | %dictionary if you haven't done it yourself and will write it to a
 853 | %\texttt{dict.eng} file by default. The usage is: \texttt{python
 854 | %cmu2praatalign.py [inputfile [outputfile]]}.
 855 | %
 856 | %{\small
 857 | %\begin{longtable}{l|l|p{.3\linewidth}|l}
 858 | %	& Symbol & Word & Transcription\\
 859 | %	\toprule
 860 | %	\multirow{6}{*}{Plosives} &
 861 | %	  p & pen &\\
 862 | %	& b & but &\\
 863 | %	& t & two &\\
 864 | %	& d & do &\\
 865 | %	& k & skill &\\
 866 | %	& g & go &\\
 867 | %	\midrule
 868 | %	\multirow{2}{*}{Affricatives} &
 869 | %	tS & chair &\\
 870 | %	& dZ & gin &\\
 871 | %	\midrule
 872 | %	\multirow{9}{*}{Fricatives} &
 873 | %		f & fool &\\
 874 | %	& v & voice &\\
 875 | %	& T & thing &\\
 876 | %	& D & this &\\
 877 | %	& s & sin &\\
 878 | %	& z & zoo &\\
 879 | %	& S & she &\\
 880 | %	& Z & pleasure &\\
 881 | %	& h & ham &\\
 882 | %	\midrule
 883 | %	\multirow{3}{*}{Nasals} &
 884 | %	m & man &\\
 885 | %	& n & no &\\
 886 | %	& N & ring &\\
 887 | %	\midrule
 888 | %	\multirow{2}{*}{Liquids} &
 889 | %	r & perro &\\
 890 | %	& l & left &\\
 891 | %	\midrule
 892 | %	\multirow{2}{*}{Sonorant glides} &
 893 | %	w & we &\\
 894 | %	& j & yes &\\
 895 | %	\midrule
 896 | %	\multirow{6}{*}{Checked vowels} &
 897 | %	I & English city &\\
 898 | %	& e & bear &\\
 899 | %	& \{ & cat &\\
 900 | %	& Q & cough &\\
 901 | %	& V & run &\\
 902 | %	& U & put &\\
 903 | %	\midrule
 904 | %	\multirow{1}{*}{Short vowels} &
 905 | %	@ & about &\\
 906 | %	\midrule
 907 | %	\multirow{13}{*}{Free vowels} &
 908 | %	i: & ease &\\
 909 | %	& eI & raise &\\
 910 | %	& aI & rise &\\
 911 | %	& OI & noise &\\
 912 | %	& u: & lose &\\
 913 | %	& @U & nose &\\
 914 | %	& aU & rouse &\\
 915 | %	& 3: & furs &\\
 916 | %	& A:\ & stars &\\
 917 | %	& O:\ & cause &\\
 918 | %	& I@ & dears &\\
 919 | %	& e@ & stairs &\\
 920 | %	& U@ & cures &\\
 921 | %	\midrule
 922 | %	\multirow{3}{*}{Special} &
 923 | %	\textless{} & word initial silence & \\
 924 | %	& \textgreater{} & word final silence & \\
 925 | %	& \# & inter word silence & \\
 926 | %	& \textless{} nib\textgreater{} & non speech sound & \\
 927 | %	\midrule\midrule
 928 | %	\multirow{22}{*}{Borrowed phones} & Symbol & description &
 929 | %		example (language)\\
 930 | %	\midrule
 931 | %	& a\textasciitilde{} & 
 932 | %		nasalized central open vowel & vent (fra)\\
 933 | %	& E\textasciitilde{} & 
 934 | %		nasalized lengthened front half open unrounded vowel & (deu)\\
 935 | %	& o\textasciitilde{} & 
 936 | %		nasalized back half closed rounded vowel & bon (fra)\\
 937 | %	& 6:\ & lengthened central neutral unrounded vowel & (aus)\\
 938 | %	& \}:\ & lengthened central closed rounded vowel & pool (aus)\\
 939 | %	& e:\ & lengthened front half closed unrounded vowel & mehr (deu)\\
 940 | %	& o:\ & lengthened back half closed rounded vowel & Sohle (deu)\\
 941 | %	& @\} & diphthong &\\
 942 | %	& Ae & diphthong & (aus)\\
 943 | %	& \{I & diphthong & (aus)\\
 944 | %	& \{O & diphthong & (aus)\\
 945 | %	& oI & diphthong & (aus)\\
 946 | %	& O & back half open rounded vowel & law (brit)\\
 947 | %	& i & front closed unrounded vowel & see\\
 948 | %	& u & back closed rounded vowel & soon\\
 949 | %	& o & back half closed rounded vowel & sore (us)\\
 950 | %	& E & front half open unrounded vowel & bed\\
 951 | %	& 6 & central neutral unrounded vowel & besser (deu)\\
 952 | %	&? & glottal stop & Verein (ger)\\
 953 | %	& x & voiceless velar fricative & loch (scot)\\
 954 | %	& C & voiceless palatal fricative & Ich (ger)\\
 955 | %	& W & voiceless labial-velar fricative &\\
 956 | %	\bottomrule
 957 | %	\caption{English phone specification}
 958 | %\end{longtable}
 959 | %}
 960 | %
 961 | %\subsection{SAMPA}
 962 | %{\small
 963 | %\begin{longtable}{l|p{.3\linewidth}|p{.15\linewidth}|l|p{.15\linewidth}|l}
 964 | %	Sym & description & examples & lang & type & location\\
 965 | %	\toprule
 966 | %	i	& front closed unrounded vowel	& English see, Spanish sí, French vite, German mi.e.ten, Italian visto	& xxx	& vowel	& front	\\
 967 | %	I	& front closed unrounded vowel, but somewhat more centralised and relaxed, in Polish: mid closed unrounded	& English city, German mit	& xxx	& vowel	& front	\\
 968 | %	1	& close central unrounded vowel	& Russian mys	& xxx	& vowel	& central	\\
 969 | %	e	& front half closed unrounded vowel	& US English bear, Spanish él, French année, German mehr, Italian rete, Catalan més	& xxx	& vowel	& front	\\
 970 | %	E	& front half open unrounded vowel	& English bed, French même, German Herr, Männer, Italian ferro, Catalan mes, Spanish perro	& xxx	& vowel	& front	\\
 971 | %	\{	& front open unrounded vowel	& English cat	& xxx	& vowel	& front	\\
 972 | %	y	& front closed rounded vowel	& French du, German Tür	& xxx	& vowel	& front	\\
 973 | %	2	& front half closed rounded vowel	& French deux (hence '2'), German Höhle	& xxx	& vowel	& front	\\
 974 | %	9	& front half open rounded vowel	& French neuf (hence '9'), German Hölle	& xxx	& vowel	& front	\\
 975 | %	@	& central neutral unrounded vowel	& English about, winner,German bitte	& xxx	& vowel	& central	\\
 976 | %	P6	& central neutral unrounded vowel	& German besser	& xxx	& vowel	& central\\
 977 | %	3	& front half open unrounded vowel, but somewhat more centralised and relaxed	& English bird,nurse	& xxx	& vowel	& central	\\
 978 | %	a	& central open vowel	& Spanish da, barra, French bateau,lac, German Haar, Italian pazzo	& xxx	& vowel	& front	\\
 979 | %	\}	& central closed rounded vowel	& Scottish English pool, Swedish sju	& xxx	& vowel	& central	\\
 980 | %	8	& central neutral rounded vowel	& Swedish kust	& xxx	& vowel	& central	\\
 981 | %	\&	& front open rounded vowel	& American English that & xxx	& vowel	& front	\\
 982 | %	M	& back closed unrounded vowel	& Japanese fuji, Korean eu & xxx	& vowel	& back	\\
 983 | %	7	& back half closed unrounded vowel	& Korean eo & xxx	& vowel	& back	\\
 984 | %	V	& back half open unrounded vowel	& RP and US English run,enough,strut	& xxx	& vowel	& back	\\
 985 | %	A	& back open unrounded vowel	& English arm, US English law, standard French âme	& xxx	& vowel	& back	\\
 986 | %	u	& back closed rounded vowel	& English soon, Spanish tú, French goût, German Hut, Mutter, Italian azzurro,tutto	& xxx	& vowel	& back	\\
 987 | %	U	& back closed rounded vowel somewhat more centralised and relaxed	& English put, Buddhist	& xxx	& vowel	& back	\\
 988 | %	o	& back half closed rounded vowel	& US English sore, Scottish English boat, Spanish yo, French beau, German Sohle, Italian dove, Catalan ona	& xxx	& vowel	& back	\\
 989 | %	O	& back half open rounded vowel	& British English law, caught, Italian cosa, Catalan dona, Spanish ojo, German Wort	& xxx	& vowel	& back	\\
 990 | %	Q	& back open rounded vowel	& British English not, cough	& xxx	& vowel	& back	\\
 991 | %	Y	& lax [y]	& German hübsch	& xxx	& vowel	& front	\\
 992 | %	p	& voiceless bilabial plosive	& English pen	& xxx	& plosive	& bilabial	\\
 993 | %	b	& voiced bilabial plosive	& English but	& xxx	& plosive	& bilabial	\\
 994 | %	t	& voiceless alveolar plosive	& English two, Spanish toma	& xxx	& plosive	& alveolar	\\
 995 | %	d	& voiced alveolar plosive	& English do, Italian cade	& xxx	& plosive	& alveolar	\\
 996 | %	ts	& voiceless alveolar affricate	& Italian calza, German zeit	& xxx	& affricate	& alveolar	\\
 997 | %	dz	& voiced alveolar affricate	& Italian zona	& xxx	& affricate	& alveolar	\\
 998 | %	tS	& voiceless postalveolar affricate	& English chair, Spanish mucho	& xxx	& affricate	& post-alveolar	\\
 999 | %	dZ	& voiced postalveolar affricate	& English gin, Italian giorno	& xxx	& affricate	& post-alveolar	\\
1000 | %	pf	& voiceless labial affricate	& German Pferd	& xxx	& affricate	& bilabial	\\
1001 | %	c	& voiceless palatal plosive	& Hungarian tyúk `hen'	& xxx	& plosive	& palatal	\\
1002 | %	k	& voiceless velar plosive	& English skill	& xxx	& plosive	& velar	\\
1003 | %	g	& voiced velar plosive	& English go	& xxx	& plosive	& velar	\\
1004 | %	q	& voiceless uvular plosive	& Arabic qof	& xxx	& plosive	& uvular	\\
1005 | %	B	& voiced bilabial fricative	& Catalan roba `clothes'	& xxx	& fricative	& bilabial	\\
1006 | %	f	& voiceless labiodental fricative	& English fool, Spanish and Italian falso	& xxx	& fricative	& labio-dental	\\
1007 | %	v	& voiced labiodental fricative	& English voice, German Welt	& xxx	& fricative	& labio-dental	\\
1008 | %	T	& voiceless dental fricative	& English thing, Castilian Spanish caza	& xxx	& fricative	& dental	\\
1009 | %	D	& voiced dental fricative	& English this	& xxx	& fricative	& dental\\
1010 | %	s	& voiceless alveolar fricative	& English see, Spanish sí	& xxx	& fricative	& alveolar	\\
1011 | %	z	& voiced alveolar fricative	& English zoo, German See	& xxx	& fricative	& alveolar	\\
1012 | %	S	& voiceless postalveolar fricative	& English she, French chemin	& xxx	& fricative	& post-alveolar	\\
1013 | %	Z	& voiced postalveolar fricative	& French jour, English pleasure	& xxx	& fricative	& post-alveolar	\\
1014 | %	C	& voiceless palatal fricative	& German Ich	& xxx	& fricative	& palatal	\\
1015 | %	x	& voiceless velar fricative	& Scots loch, Castilian Spanish ajo	& xxx	& fricative	& velar	\\
1016 | %	G	& voiced velar fricative	& Greek $\gamma\alpha\lambda\alpha$	& xxx	& fricative	& velar	\\
1017 | %	h	& voiceless glottal fricative	& English ham, German Hand	& xxx	& fricative	& glottal	\\
1018 | %	m	& bilabial nasal	& English man	& xxx	& nasal	& bilabial	\\
1019 | %	F	& labiodental nasal	& Spanish infierno, Hungarian kámfor	& xxx	& nasal	& labio-dental	\\
1020 | %	n	& alveolar nasal	& English, Spanish and Italian no	& xxx	& nasal	& alveolar	\\
1021 | %	J	& palatal nasal	& Spanish año, French oignon	& xxx	& nasal	& palatal	\\
1022 | %	N	& velar nasal	& English ring, Italian bianco, Tagalog ngayón	& xxx	& nasal	& velar	\\
1023 | %	l	& alveolar lateral approximant	& English left, Spanish largo	& xxx	& lateral-approximant	& alveolar	\\
1024 | %	L	& palatal lateral approximant	& Italian aglio, Catalan colla,	& xxx	& lateral-approximant	& palatal	\\
1025 | %	5	& velarized dental lateral	& English meal Catalan alga	& xxx	& lateral-approximant	& dental-velar	\\
1026 | %	4	& alveolar tap	& Spanish pero, American English muddy	& use	& tap	& alveolar	\\
1027 | %	r	& alveolar trill	& Spanish perro	& xxx	& trill	& alveolar	\\
1028 | %	R	& uvular trill	& German Reich	& xxx	& trill	& uvular	\\
1029 | %	P	& labiodental approximant	& Dutch Waar	& xxx	& approximant	& labio-dental	\\
1030 | %	w	& labial-velar approximant	& English we, French oui	& xxx	& approximant	& labio-velar	\\
1031 | %	H	& labial-palatal approximant	& French huit	& xxx	& approximant	& labio-palatal	\\
1032 | %	j	& palatal approximant	& English yes, French yeux	& xxx	& approximant	& palatal	\\
1033 | %?	& glottal stop	& German Verein, Danish st\o{}d	& xxx	& plosive	& glottal	\\
1034 | %	W	& voiceless labial-velar fricative	& 	& xxx	& fricative	& labio-velar	\\
1035 | %	D:\	& lengthened voiced dental fricative	& 	& fin	& fricative	& dental	\\
1036 | %	T:\	& lengthened voiceless dental fricative	& 	& fin	& fricative	& dental	\\
1037 | %	i:\	& lengthened front closed unrounded vowel	& mieten	& deu	& vowel	& front	\\
1038 | %	ii	& lengthened front closed unrounded vowel in quantity II	& riisu	& ekk	& vowel	& front	\\
1039 | %	ii:	& lengthened front closed unrounded vowel in quantity III	& 	& ekk	& vowel	& front	\\
1040 | %	e:	& lengthened front half closed unrounded vowel	& mehr	& deu	& vowel	& front	\\
1041 | %	ee	& lengthened front half closed unrounded vowel in quantity II	& keere	& ekk	& vowel	& front	\\
1042 | %	ee:	& lengthened front half closed unrounded vowel in quantity III	& 	& ekk	& vowel	& front	\\
1043 | %	E:\	& lengthened front half open unrounded vowel	& M\"ar	& deu	& vowel	& front	\\
1044 | %	y:	& lengthened front closed rounded vowel	& Tür	& deu	& vowel	& front	\\
1045 | %	Y:\	& lengthened lax [y]	& (Swiss German)	& deu	& vowel	& front	\\
1046 | %	2:	& lengthened front half closed rounded vowel	& Höhle	& deu	& vowel	& front	\\
1047 | %	a:	& lengthened central open vowel	& Haar	& deu	& vowel	& central	\\
1048 | %	u:	& lengthened back closed rounded vowel	& Hut	& deu	& vowel	& back	\\
1049 | %	o:	& lengthened back half closed rounded vowel	& Sohle	& deu	& vowel	& back	\\
1050 | %	3:	& lengthened front half open unrounded vowel	& furs	& aus	& vowel	& front	\\
1051 | %	A:\	& lengthened back open unrounded vowel	& stars	& aus	& vowel	& back	\\
1052 | %	O:\	& lengthened back half open rounded vowel	& cause	& aus	& vowel	& back	\\
1053 | %	P6:	& lengthened central neutral unrounded vowel	& 	& aus	& vowel	& central	\\
1054 | %	\}:	& lengthened central closed rounded vowel	& pool	& aus	& vowel	& central	\\
1055 | %	Q:\	& lengthened open back rounded	& (Swiss German)	& aus	& vowel	& back	\\
1056 | %	9:	& lengthened front half open rounded vowel	& 	& nld	& vowel	& front	\\
1057 | %	\{\{	& lengthened front open unrounded vowel in quantity II	& kääru	& ekk	& vowel	& front	\\
1058 | %	\{:	& lengthened front open unrounded vowel in quantity II	& kääru	& ekk	& vowel	& front	\\
1059 | %	\{\{:	& lengthened front open unrounded vowel in quantity III	& 	& ekk	& vowel	& front	\\
1060 | %	yy	& lengthened front closed rounded vowel (quantitiy II)	& müüri	& ekk	& vowel	& front	\\
1061 | %	yy:	& lengthened front closed rounded vowel (quantitiy III)	& 	& ekk	& vowel	& front	\\
1062 | %	22	& lengthened front half closed rounded vowel	& nööri	& ekk	& vowel	& front	\\
1063 | %	22:	& lengthened front half closed rounded vowel in quantity III	& 	& ekk	& vowel	& front	\\
1064 | %	uu	& lengthened back closed rounded vowel	& kuuri	& ekk	& vowel	& back	\\
1065 | %	uu:	& lengthened back closed rounded vowel in quantity III	& 	& ekk	& vowel	& back	\\
1066 | %	oo	& lengthened back half closed rounded vowel	& poori	& ekk	& vowel	& back	\\
1067 | %	oo:	& lengthened back half closed rounded vowel in quantity III	& 	& ekk	& vowel	& back	\\
1068 | %	77	& back half closed unrounded vowel in quantity II	& sõõre	& ekk	& vowel	& back	\\
1069 | %	7:	& back half closed unrounded vowel in quantity II	& sõõre	& ekk	& vowel	& back	\\
1070 | %	77:	& back half closed unrounded vowel in quantity III	& 	& ekk	& vowel	& back	\\
1071 | %	AA	& lengthened back open unrounded vowel in quantity II	& vaaru	& ekk	& vowel	& back	\\
1072 | %	AA:\ & lengthened back open unrounded vowel in quantity III	& 	& ekk	& vowel	& back	\\
1073 | %	aU	& diphthong	& Haus	& deu	& diphthong	& front\textgreater{} back	\\
1074 | %	aI	& diphthong	& Bein	& deu	& diphthong	& front	\\
1075 | %	ai	& diphthong	& 	& ita	& diphthong	& front	\\
1076 | %	a:i	& diphthong	& 	& nld	& diphthong	& front	\\
1077 | %	Ae	& diphthong	& 	& aus	& diphthong	& back\textgreater{} front	\\
1078 | %	Au	& diphthong	& 	& nld	& diphthong	& back	\\
1079 | %	OY	& diphthong	& heulen	& deu	& diphthong	& back\textgreater{} front	\\
1080 | %	eI	& diphthong	& raise	& aus	& diphthong	& front	\\
1081 | %	e@	& diphthong	& stairs	& aus	& diphthong	& front\textgreater{} central	\\
1082 | %	ei	& diphthong	& 	& ita	& diphthong	& front	\\
1083 | %	e:i	& diphthong	& 	& nld	& diphthong	& front	\\
1084 | %	eU	& diphthong	& 	& por	& diphthong	& front\textgreater{} back	\\
1085 | %	EI	& diphthong	& raise	& eng	& diphthong	& front	\\
1086 | %	Ei	& diphthong	& 	& ita	& diphthong	& front	\\
1087 | %	\{I	& diphthong	& 	& aus	& diphthong	& front	\\
1088 | %	\{O	& diphthong	& 	& aus	& diphthong	& front\textgreater{} back	\\
1089 | %	I@	& diphthong	& dears	& aus	& diphthong	& front\textgreater{} central	\\
1090 | %	Ii:	& diphthong	& accede	& aus	& diphthong	& front	\\
1091 | %	I:\	& lengthened front closed unrounded vowel	& (Swiss German)	& deu	& vowel	& front	\\
1092 | %	i:@	& diphthong	& memorial	& nze	& diphthong	& central\textgreater{} front	\\
1093 | %	io	& diphthong	& 	& ita	& diphthong	& front\textgreater{} back	\\
1094 | %	iu	& diphthong	& 	& nld	& diphthong	& front\textgreater{} back	\\
1095 | %	ja	& diphthong	& 	& ita	& diphthong	& front	\\
1096 | %	jo	& diphthong	& 	& ita	& diphthong	& front\textgreater{} back	\\
1097 | %	ju	& diphthong	& 	& ita	& diphthong	& front\textgreater{} back	\\
1098 | %	oI	& diphthong	& 	& aus	& diphthong	& back\textgreater{} front	\\
1099 | %	oi	& diphthong	& 	& ita	& diphthong	& back\textgreater{} front	\\
1100 | %	o:i	& diphthong	& 	& nld	& diphthong	& back\textgreater{} front	\\
1101 | %	oU	& diphthong	& 	& por	& diphthong	& back	\\
1102 | %	oE	& diphthong	& 	& ita	& diphthong	& back\textgreater{} front	\\
1103 | %	OI	& diphthong	& noise	& aus	& diphthong	& back\textgreater{} front	\\
1104 | %	Oi	& diphthong	& 	& ita	& diphthong	& back\textgreater{} front	\\
1105 | %	ue	& diphthong	& 	& ita	& diphthong	& back\textgreater{} front	\\
1106 | %	ui	& diphthong	& 	& nld	& diphthong	& back\textgreater{} front	\\
1107 | %	U@	& diphthong	& cures	& aus	& diphthong	& back\textgreater{} central	\\
1108 | %	U:\	& lenthened back closed rounded vowel somewhat more centralised and relaxed	& (Swiss German)	& deu	& vowel	& back	\\
1109 | %	wa	& diphthong	& 	& ita	& diphthong	& front	\\
1110 | %	we	& diphthong	& 	& ita	& diphthong	& front	\\
1111 | %	wi	& diphthong	& 	& ita	& diphthong	& front	\\
1112 | %	wO	& diphthong	& 	& ita	& diphthong	& back	\\
1113 | %	yu	& diphthong	& 	& nld	& diphthong	& front\textgreater{} back	\\
1114 | %	@U	& diphthong	& nose	& aus	& diphthong	& central\textgreater{} back	\\
1115 | %	@\}	& diphthong	& 	& aus	& diphthong	& central	\\
1116 | %	@@	& geminate of schwa in quantity II	& 	& ekk	& vowel	& central	\\
1117 | %	@:	& geminate of schwa in quantity II	& 	& ekk	& vowel	& central	\\
1118 | %	@@:	& geminate of schwa in quantity III	& 	& ekk	& vowel	& central	\\
1119 | %	9y	& diphthong	& 	& nld	& diphthong	& back\textgreater{} front	\\
1120 | %	QI	& diphthong	& abide	& aus	& diphthong	& central\textgreater{} front	\\
1121 | %	U\}	& diphthong	& abuse	& aus	& diphthong	& central\textgreater{} back	\\
1122 | %	VU	& diphtong	& acetone	& aus	& diphthong	& central\textgreater{} back	\\
1123 | %	Vi	& diphthong	& abased	& aus	& diphthong	& central\textgreater{} front	\\
1124 | %	\{o	& diphthong	& accounts	& aus	& diphthong	& front\textgreater{} back	\\
1125 | %	2i	& diphthong	& 	& ekk	& diphthong	& back\textgreater{} front	\\
1126 | %	2i:	& diphthong	& 	& ekk	& diphthong	& back\textgreater{} front	\\
1127 | %	7o:	& diphthong	& 	& ekk	& diphthong	& back	\\
1128 | %	7u:	& diphthong	& 	& ekk	& diphthong	& back	\\
1129 | %	7u	& diphthong	& 	& ekk	& diphthong	& back	\\
1130 | %	Ai	& diphthong	& 	& ekk	& diphthong	& back\textgreater{} front	\\
1131 | %	Ai:	& diphthong	& 	& ekk	& diphthong	& back\textgreater{} front	\\
1132 | %	Ao:	& diphthong	& 	& ekk	& diphthong	& back	\\
1133 | %	Au:	& diphthong	& 	& ekk	& diphthong	& back	\\
1134 | %	Ae:	& diphthong	& 	& ekk	& diphthong	& back\textgreater{} front	\\
1135 | %	ei:	& diphthong	& 	& ekk	& diphthong	& front	\\
1136 | %	e:u	& diphthong	& 	& nld	& diphthong	& front	\\
1137 | %	i\textasciitilde{}	& nasalized front closed unrounded vowel	& 	& xxx	& vowel	& front	\\
1138 | %	e\textasciitilde{}	& nasalized front half closed unrounded vowel	& vin	& fra	& vowel	& front	\\
1139 | %	a\textasciitilde{}	& nasalized central open vowel	& vent	& fra	& vowel	& front	\\
1140 | %	o\textasciitilde{}	& nasalized back half closed rounded vowel	& bon	& fra	& vowel	& back	\\
1141 | %	9\textasciitilde{}	& nasalized front half open rounded vowel	& brun,neuv	& fra	& vowel	& front	\\
1142 | %	E\textasciitilde{}	& nasalized lengthened front half open unrounded vowel	& 	& deu	& vowel	& front	\\
1143 | %	O\textasciitilde{}	& nasalized back half open rounded vowel	& 	& deu	& vowel	& back	\\
1144 | %	u\textasciitilde{}	& nasalized back closed rounded vowel	& 	& xxx	& vowel	& back	\\
1145 | %	a:\textasciitilde{}	& nasalized lengthened central open vowel	& 	& deu	& vowel	& central	\\
1146 | %	E:\textasciitilde{}	& nasalized lengthened front half open unrounded vowel	& 	& deu	& vowel	& front	\\
1147 | %	o:\textasciitilde{}	& nasalized lengthened back half closed rounded vowel	& 	& deu	& vowel	& back	\\
1148 | %	O:\textasciitilde{}	& nasalized lengthened back half open rounded vowel	& 	& nld	& vowel	& back	\\
1149 | %	ts\_j	& palatalized voiceless alveolar affricate	& c'ma	& pol	& affricate	& post-alveolar	\\
1150 | %	dz\_j	& palatalized voiced alveolar affricate	& dz'wig	& pol	& affricate	& alveolar	\\
1151 | %	s\_j	& palatalized voiceless alveolar fricative	& syk	& pol	& fricative	& alveolar	\\
1152 | %	s\_js	& palatalized voiceless alveolar fricative in quantity II	& kassi	& ekk	& fricative	& alveolar	\\
1153 | %	s\_j:s	& palatalized voiceless alveolar fricative in quantity III	& 	& ekk	& fricative	& alveolar	\\
1154 | %	z\_j	& palatalized voiced alveolar fricative	& zbir	& pol	& fricative	& alveolar	\\
1155 | %	n\_j	& palatalized alveolar nasal	& kon'	& pol	& nasal	& alveolar	\\
1156 | %	n\_jn	& palatalized alveolar nasal in quantity II	& panni	& ekk	& nasal	& alveolar	\\
1157 | %	n\_j:n	& palatalized alveolar nasal in quantity III	& 	& ekk	& nasal	& alveolar	\\
1158 | %	l\_j	& palatalized alveolar lateral approximant	& pali	& ekk	& lateral-approximant	& alveolar	\\
1159 | %	l\_jl	& palatalized alveolar lateral approximant in quantity II	& palli	& ekk	& lateral-approximant	& alveolar	\\
1160 | %	l\_j:l	& palatalized alveolar lateral approximant in quantity III	& 	& ekk	& lateral-approximant	& alveolar	\\
1161 | %	t\_j	& palatalized voiceless alveolar plosive	& padi	& ekk	& plosive	& alveolar	\\
1162 | %	t\_jt	& palatalized voiceless alveolar plosive in quantity II	& pati	& ekk	& plosive	& alveolar	\\
1163 | %	t\_j:t	& palatalized voiceless alveolar plosive in quantity III	& 	& ekk	& plosive	& alveolar	\\
1164 | %	d\_j	& palatalized voiced alveolar plosive	& gyár	& hun	& plosive	& alveolar	\\
1165 | %	dd\_j	& geminate of d'	& egy	& hun	& plosive	& alveolar	\\
1166 | %	g\_j	& palatalized voiced velar plosive	& Gienek	& pol	& plosive	& velar	\\
1167 | %	x\_j	& palatalized voiceless velar fricative	& hiacynt	& pol	& fricative	& velar	\\
1168 | %	k\_j	& palatalized voiceless velar plosive	& kierowca	& pol	& plosive	& velar	\\
1169 | %	p\_j	& palatalized voiceless bilabial plosive	& piasek	& pol	& plosive	& bilabial	\\
1170 | %	xx\_j	& geminate of x'	& 	& hun	& fricative	& velar	\\
1171 | %	p\_h	& aspirated voiceless bilabial plosive	& 	& spa	& plosive	& bilabial	\\
1172 | %	t\_h	& aspirated voiceless alveolar plosive	& 	& spa	& plosive	& alveolar	\\
1173 | %	k\_h	& aspirated voiceless velar plosive	& 	& spa	& plosive	& velar	\\
1174 | %	tt	& geminate of t	& fatto	& ita	& fricative	& bilabial	\\
1175 | %	t:	& lengthened t	& että	& fin	& plosive	& alveolar	\\
1176 | %	t:t	& geminate of t in quantity III	& 	& ekk	& plosive	& alveolar	\\
1177 | %	pp	& geminate of p	& 	& ita	& plosive	& bilabial	\\
1178 | %	p:	& lengthened p	& 	& fin	& plosive	& bilabial	\\
1179 | %	p:p	& geminate of p in quantity III	& 	& ekk	& plosive	& bilabial	\\
1180 | %	kk	& geminate of k	& 	& ita	& plosive	& velar	\\
1181 | %	k:	& lengthened k	& takkinsa	& fin	& plosive	& velar	\\
1182 | %	k:k	& geminate of k in quantity III	& 	& ekk	& plosive	& velar	\\
1183 | %	dd	& geminate of d	& 	& ita	& plosive	& alveolar	\\
1184 | %	gg	& geminate of g	& 	& ita	& plosive	& velar	\\
1185 | %	g:	& lengthened g	& 	& fin	& plosive	& velar	\\
1186 | %	bb	& geminate of b	& 	& ita	& plosive	& bilabial	\\
1187 | %	b:	& lengthened b	& 	& fin	& plosive	& bilabial	\\
1188 | %	ttS	& geminate of tS	& 	& ita	& affricate	& post-alveolar	\\
1189 | %	tts	& geminate of ts	& 	& ita	& affricate	& alveolar	\\
1190 | %	ddZ	& geminate of dZ	& 	& ita	& affricate	& post-alveolar	\\
1191 | %	ddz	& geminate of dz	& zona	& ita	& affricate	& alveolar	\\
1192 | %	vv	& geminate of v	& 	& ita	& fricative	& labio-dental	\\
1193 | %	v:	& lengthened v	& 	& fin	& fricative	& labio-dental	\\
1194 | %	v:v	& geminate of v in quantity III	& 	& ekk	& fricative	& labio-dental	\\
1195 | %	ss	& geminate of s	& 	& ita	& fricative	& alveolar	\\
1196 | %	s:	& lengthened s	& 	& fin	& fricative	& alveolar	\\
1197 | %	s:s	& geminate of s in quantity III	& 	& ekk	& fricative	& alveolar	\\
1198 | %	zz	& geminate of z	& 	& hun	& fricative	& alveolar	\\
1199 | %	SS	& geminate of S	& 	& ita	& fricative	& post-alveolar	\\
1200 | %	S:\	& lengthened S	& 	& fin	& fricative	& post-alveolar	\\
1201 | %	S:S	& geminate of S in quantity III	& 	& ekk	& fricative	& post-alveolar	\\
1202 | %	ZZ	& geminate of Z	& 	& hun	& fricative	& post-alveolar	\\
1203 | %	xx	& geminate of x	& 	& hun	& fricative	& velar	\\
1204 | %	rr	& geminate of r	& 	& ita	& trill	& alveolar	\\
1205 | %	r:	& lengthened r	& 	& fin	& trill	& alveolar	\\
1206 | %	r:r	& geminate of r in quantity III	& 	& ekk	& trill	& alveolar	\\
1207 | %	RR	& geminate of R in quantity II	& 	& ekk	& trill	& uvular	\\
1208 | %	nn	& geminate of n	& 	& ita	& nasal	& alveolar	\\
1209 | %	n:	& lengthened n	& 	& fin	& nasal	& alveolar	\\
1210 | %	n:n	& geminate of n in quantity III	& 	& ekk	& nasal	& alveolar	\\
1211 | %	NN	& geminate of N	& (Swiss German)	& deu	& nasal	& velar	\\
1212 | %	N:\	& lengthened N	& 	& fin	& nasal	& velar	\\
1213 | %	N\_j	& palatalized velar nasal	& 	& ekk	& nasal	& velar	\\
1214 | %	ww	& geminate of w	& (Swiss German)	& deu	& approximant	& labio-velar	\\
1215 | %	mm	& geminate of m	& 	& ita	& nasal	& bilabial	\\
1216 | %	m:	& lengthened m	& hommasta	& fin	& nasal	& bilabial	\\
1217 | %	m:m	& geminate of m in quantity III	& 	& ekk	& nasal	& bilabial	\\
1218 | %	FF	& geminate of F	& 	& hun	& nasal	& labio-dental	\\
1219 | %	LL	& geminate of L	& 	& ita	& lateral-approximant	& palatal	\\
1220 | %	ll	& geminate of l	& 	& ita	& lateral-approximant	& alveolar	\\
1221 | %	l:	& lengthened l	& jolla	& fin	& lateral-approximant	& alveolar	\\
1222 | %	l:l	& geminate of l in quantity III	& 	& ekk	& approximant	& alveolar	\\
1223 | %	JJ	& geminate of J	& 	& ita	& nasal	& palatal	\\
1224 | %	jj	& geminate of j	& 	& ekk	& approximant	& palatal	\\
1225 | %	j:	& lengthened j	& 	& fin	& approximant	& palatal	\\
1226 | %	j:j	& geminate of j in quantity III	& 	& ekk	& approximant	& palatal	\\
1227 | %	ff	& geminate of f	& 	& ita	& fricative	& bilabial	\\
1228 | %	f:	& lengthened f	& 	& fin	& fricative	& bilabial	\\
1229 | %	f:f	& geminate of f in quantity III	& 	& ekk	& fricative	& bilabial	\\
1230 | %	hh	& geminate of h	& 	& ekk	& fricative	& glottal	\\
1231 | %	h:	& lengthened h	& 	& fin	& fricative	& glottal	\\
1232 | %	h:h	& geminate of h in quantity III	& 	& ekk	& fricative	& glottal	\\
1233 | %	ttS\_cl	& closure of ttS	& 	& ita	& affricate	& post-alveolar	\\
1234 | %	ttS\_rl	& release of ttS	& 	& ita	& affricate	& post-alveolar	\\
1235 | %	tts\_cl	& closure of tts	& 	& ita	& affricate	& alveolar	\\
1236 | %	tts\_rl	& release of tts	& 	& ita	& affricate	& alveolar	\\
1237 | %	ddZ\_cl	& closure of ddZ	& 	& ita	& affricate	& post-alveolar	\\
1238 | %	ddZ\_rl	& release of ddZ	& 	& ita	& affricate	& post-alveolar	\\
1239 | %	ddz\_cl	& closure of ddz	& zona	& ita	& affricate	& alveolar	\\
1240 | %	ddz\_rl	& release of ddz	& zona	& ita	& affricate	& alveolar	\\
1241 | %	tS\_cl	& closure of tS	& 	& ita	& affricate	& post-alveolar	\\
1242 | %	tS\_rl	& release of tS	& 	& ita	& affricate	& post-alveolar	\\
1243 | %	ts\_cl	& closure of ts	& 	& ita	& affricate	& alveolar	\\
1244 | %	ts\_rl	& release of ts	& 	& ita	& affricate	& alveolar	\\
1245 | %	dZ\_cl	& closure of dZ	& 	& ita	& affricate	& post-alveolar	\\
1246 | %	dZ\_rl	& release of dZ	& 	& ita	& affricate	& post-alveolar	\\
1247 | %	dz\_cl	& closure of dz	& 	& ita	& affricate	& alveolar	\\
1248 | %	dz\_rl	& release of dz	& 	& ita	& affricate	& alveolar	\\
1249 | %	tt\_rl	& release of tt	& 	& ita	& plosive	& alveolar	\\
1250 | %	tt\_cl	& closure of tt	& 	& ita	& plosive	& alveolar	\\
1251 | %	pp\_cl	& closure of pp	& 	& ita	& plosive	& bilabial	\\
1252 | %	pp\_rl	& release of pp	& 	& ita	& plosive	& bilabial	\\
1253 | %	kk\_cl	& closure of kk	& 	& ita	& plosive	& velar	\\
1254 | %	kk\_rl	& release of kk	& 	& ita	& plosive	& velar	\\
1255 | %	dd\_cl	& release of dd	& 	& ita	& plosive	& alveolar	\\
1256 | %	dd\_rl	& release of dd	& 	& ita	& plosive	& alveolar	\\
1257 | %	gg\_cl	& release of gg	& 	& ita	& plosive	& velar	\\
1258 | %	gg\_rl	& release of gg	& 	& ita	& plosive	& velar	\\
1259 | %	bb\_cl	& release of bb	& 	& ita	& plosive	& bilabial	\\
1260 | %	bb\_rl	& release of bb	& 	& ita	& plosive	& bilabial	\\
1261 | %	t\_cl	& closure of t	& 	& ita	& plosive	& alveolar	\\
1262 | %	t\_rl	& release of t	& 	& ita	& plosive	& alveolar	\\
1263 | %	p\_cl	& closure of p	& 	& ita	& plosive	& bilabial	\\
1264 | %	p\_rl	& release of p	& 	& ita	& plosive	& bilabial	\\
1265 | %	k\_cl	& closure of k	& 	& ita	& plosive	& velar	\\
1266 | %	k\_rl	& release of k	& 	& ita	& plosive	& velar	\\
1267 | %	g\_cl	& closure of g	& 	& ita	& plosive	& velar	\\
1268 | %	g\_rl	& release of g	& 	& ita	& plosive	& velar	\\
1269 | %	d\_cl	& closure of d	& 	& ita	& plosive	& alveolar	\\
1270 | %	d\_rl	& release of d	& 	& ita	& plosive	& alveolar	\\
1271 | %	b\_cl	& closure of b	& 	& ita	& plosive	& bilabial	\\
1272 | %	b\_rl	& release of b	& 	& ita	& plosive	& bilabial	\\
1273 | %	\textless{}	& recording initial silence	& 	& xxx	& silence	& silence	\\
1274 | %	\textgreater{}	& recording trailing silence	& 	& xxx	& silence	& silence	\\
1275 | %	\#	& inter-word silence	& 	& xxx	& silence	& silence	\\
1276 | %	\textless{} nib\textgreater{}	& noise, non-human	& 	& xxx	& noise	& noise	\\
1277 | %	\textless{} p:\textgreater{}	& silence interval	& 	& xxx	& silence	& silence	\\
1278 | %	\textless{} usb\textgreater{}	& human noise, garbage	& 	& xxx	& noise	& noise	\\
1279 | %	p\_\textgreater{}	& voiceless bilabial ejective	& Georgian p\'{}erangi	& xxx	& ejective	& bilabial\\	
1280 | %	t\_\textgreater{}	& voiceless alveolar ejective	& Georgian zedat\'{}ani	& xxx	& ejective	& alveolar\\
1281 | %	ts\_\textgreater{}	& voiceless alveolar affricate ejective	& Georgian ts'indebi	& xxx	& ejective	& alveolar	\\
1282 | %	ts\`	& retroflex voiceless affricate	& cyk	& pol	& affricate	& retroflex	\\
1283 | %	dz\`	& retroflex voiced affricate	& dzwon	& pol	& affricate	& retroflex	\\
1284 | %	3`	& retroflex front half open unrounded vowel	& furs	& use	& vowel	& front	\\
1285 | %	tS\_\textgreater{}	& voiceless postalveolar affricate ejective	& Georgian k'uch'is	& xxx	& ejective	& post-alveolar	\\
1286 | %	c\_\textgreater{}	& voiceless palatal ejective plosive	& Georgian ch' & xxx	& ejective	& palatal	\\
1287 | %	J-	& voiced palatal plosive	& Hungarian egy `one'	& xxx	& plosive	& palatal	\\
1288 | %	k\_\textgreater{}	& voiceless velar ejective	& Georgian uk'an	& xxx	& ejective	& velar	\\
1289 | %	q\_\textgreater{}	& voiceless uvular ejective	& Georgian saavadmq'opo	& xxx	& ejective	& uvular\\
1290 | %	p-	& voiceless bilabial fricative	& Japanese fu	& xxx	& fricative	& bilabial	\\
1291 | %	s\`	& retroflex voiceless fricative	& szyk	& pol	& fricative	& retroflex	\\
1292 | %	z\`	& retroflex voiced fricative	& żyto	& pol	& fricative	& retroflex	\\
1293 | %	j-	& voiced palatal fricative	& Spanish ayuda	& xxx	& fricative	& palatal	\\
1294 | %	M-	& velar approximant	& Spanish algo	& xxx	& approximant	& velar	\\
1295 | %	X-	& voiceless pharyngeal fricative	& Arabic h.\^a	& xxx	& fricative	& pharyngeal	\\
1296 | %?{-}	& voiced pharyngeal fricative	& Arabic \'ayn	& xxx	& fricative	& pharyngeal	\\
1297 | %	h-	& voiced glottal fricative	& English	& use	& fricative	& glottal	\\
1298 | %	r-	& alveolar approximant	& English run	& xxx	& approximant	& alveolar	\\
1299 | %	N=	& syllabic velar nasal	& English walking	& use	& nasal	& velar	\\
1300 | %	n=	& syllabic alveolar nasal	& English Boston	& use	& nasal	& alveolar	\\
1301 | %	m=	& syllabic bilabial nasal	& English bottom	& use	& nasal	& bilabial	\\
1302 | %	l=	& syllabic lateral	& English bridal	& use	& lateral-approximant	& alveolar	\\
1303 | %\end{longtable}
1304 | %}
1305 | 
1306 | \newpage
1307 | \section{Version history}
1308 | \begin{longtable}{|p{0.22\linewidth}p{0.8\linewidth}|}
1309 | 	\toprule
1310 | 	2.0 (2016{-}12{-}28) & \tabitem{} Add orthographic word level alignment.\\
1311 | 		& \tabitem{} Clean up book source code and makefile.\\
1312 | 	\midrule
1313 | 	1.9 (2016{-}01{-}19) & \tabitem{} Better errors in textwindow.\\
1314 | 		& \tabitem{} Phonetizer file selector in settings.\\
1315 | 		& \tabitem{} Added the errors to align tier.\\
1316 | 		& \tabitem{} Temp files are cleaned up before alignment, therefore no
1317 | 			ghost annotations when the alignment has failed.\\
1318 | 	\midrule
1319 | 	1.8 (2015{-}11{-}19) & \tabitem{} Added support for \texttt{Sound}.\\
1320 | 		& \tabitem{} Changed email address of author.\\
1321 | 		& \tabitem{} Elaborated the installation instructions.\\
1322 | 		& \tabitem{} Splitted up the model and phonetizer selection.\\
1323 | 		& \tabitem{} Cleaned up the phone specifications.\\
1324 | 		& \tabitem{} Added an universal phonetizer.\\
1325 | 		& \tabitem{} Big book update with structural changes.\\
1326 | 	\midrule
1327 | 	1.7a (2015{-}10{-}13) & \tabitem{} Fixed critical tier creation bug.\\
1328 | 	& \tabitem{} Added version to settings window.\\
1329 | 	\midrule
1330 | 	1.7 (2015{-}10{-}08) & \tabitem{} More robuust tier creation.\\
1331 | 		& \tabitem{} Some typos fixed in the manual.\\
1332 | 		& \tabitem{} Updated the bibtex snippet.\\
1333 | 		&	\tabitem{} Better tier alignment when the tier is empty.\\
1334 | 		&	\tabitem{} Corrected some spelling errors.\\
1335 | 		&	\tabitem{} Tidied the makefile for the book.\\
1336 | 	\midrule
1337 | 	1.6a (2015{-}09{-}29) & \tabitem{} Fixed the dictionary generator for English.\\
1338 | 		& \tabitem{} Correctly added websites to the authors of the models.\\
1339 | 	\midrule
1340 | 	1.60 (2015{-}09{-}07) & \tabitem{} Fixed menu name from \texttt{force} to
1341 | \texttt{forced}.\\
1342 | 		& \tabitem{} Fixed silence problem in experimental models.\\
1343 | 		& \tabitem{} Added version to setup dialog.\\
1344 | 		& \tabitem{} Added language for using the general sampa model without
1345 | 			phonetizer.\\
1346 | 		& \tabitem{} Created a better error when loading rulesets or dictionaries in
1347 | 			non supported encodings or non existent files.\\
1348 | 		& \tabitem{} Changed the licence to MIT.\\
1349 | 	\midrule
1350 | 	1.50 (2015{-}07{-}07) & \tabitem{} Spanish models merged in original master.\\
1351 | 		& \tabitem{} Procedures in different file, thus cleaner code.\\
1352 | 		& \tabitem{} Reinitialized repo.\\
1353 | 		& \tabitem{} Faster book compilation.\\
1354 | 	\midrule
1355 | 	1.40 (2015{-}07{-}01) & \tabitem{} Added branch for own spanish models.\\
1356 | 		& \tabitem{} Updated book about ruleset.\\
1357 | 		& \tabitem{} Changed name in menu from \texttt{Setup\ldots} to 
1358 | 			\texttt{Set up\dots}.\\
1359 | 	\midrule
1360 | 	1.39 (2015{-}04{-}01) & \tabitem{} Fixed tier alignment.\\
1361 | 	\midrule
1362 | 	1.3 (2015{-}03{-}25) & \tabitem{} Added option for custom python path.\\
1363 | 	 & \tabitem{} Greatly simplified rulesets.\\
1364 | 	\midrule
1365 | 	1.2 (2015{-}02{-}18) & \tabitem{} Simplified installation scripts.\\
1366 | 		& \tabitem{} Fixed a unicode bug in generating dictionaries.\\
1367 | 	\midrule
1368 | 	1.1 (2015{-}01{-}30) & \tabitem{} Implemented better error messaging.\\
1369 | 		& \tabitem{} Simplified the python code into one file.\\
1370 | 		& \tabitem{} Fixed a bug that leaded to a messed up view.\\
1371 | 		& \tabitem{} Fixed small phonetization errors.\\
1372 | 	\midrule
1373 | 	1.0 (2015{-}01{-}09) & \tabitem{} Converted the readme to a pdf.\\
1374 | 		& \tabitem{} Speeded up the process by disabling pitch, intensity,
1375 | spectrum, pulses and formants while aligning, the settings do get restored
1376 | afterwards.\\
1377 | 		& \tabitem{} Fixed a bug that leaded to a messed up view.\\
1378 | 	\midrule
1379 | 	0.9a (2014{-}12{-}02) & \tabitem{} Small bugfix in dictionary generation fixed.\\
1380 | 	\midrule
1381 | 	0.9 & \tabitem{} Cleaned up some parts of the readme.\\
1382 | 		& \tabitem{} Added language specific information.\\
1383 | 		& \tabitem{} Added english as language. Although there is no phonetizing
1384 | implemented.\\
1385 | 		& \tabitem{} README.html better with light background for code blocks.\\
1386 | 		& \tabitem{} Updated citing method with bibtex.\\
1387 | 	\midrule
1388 | 	0.8 (2014{-}10{-}31) & \tabitem{} Removed all the binary folders.\\
1389 | 		& \tabitem{} Made the binary finding interactive.\\
1390 | 		& \tabitem{} Made all the file chooser dialogs interactive.\\
1391 | 	\midrule
1392 | 	0.7 (2014{-}10{-}29) & \tabitem{} Added windows support.\\
1393 | 		&	\tabitem{} Cleaned up documentation.\\
1394 | 		& \tabitem{} Removed binaries due htk licence.\\
1395 | 	\midrule
1396 | 	0.6 (2014{-}10{-}22) & \tabitem{} Refactored and cleaned up the source.\\
1397 | 	\midrule
1398 | 	0.5a (2014{-}09{-}08) & \tabitem{} Added comments to source code (praat).\\
1399 | 		& \tabitem{} Cleaned up source.\\
1400 | 	\midrule
1401 | 	0.5 (2014{-}09{-}04) & \tabitem{} Fixed acronyms in spanish.\\
1402 | 		& \tabitem{} Fixed cleaning with extended boundaries.\\
1403 | 		& \tabitem{} Added rudimentary ruleset implementation.\\
1404 | 	\midrule
1405 | 	0.4 (2014{-}08{-}29) & \tabitem{} Added option for enlarging the boundaries
1406 | automatically.\\
1407 | 	\midrule
1408 | 	0.21 (2014{-}08{-}13) & \tabitem{} Settings split in non interactive and
1409 | interactive so that the interactive one reflects the current settings.\\
1410 | 	\midrule
1411 | 	0.2 (2014{-}08{-}11) & \tabitem{} Better mac compatibility.\\
1412 | 	\midrule
1413 | 	0.1a (2014{-}06{-}30) & \tabitem{} Tier alignment fixed.\\
1414 | 		& \tabitem{} Readme for dutch.\\
1415 | 	\midrule
1416 | 	0.08 (2014{-}04{-}29) & \tabitem{} Cleaned up some stuff.\\
1417 | 		& \tabitem{} Added dutch.\\
1418 | 		& \tabitem{} Readme for spanish and sampa.\\
1419 | 	\midrule
1420 | 	0.07 (2014{-}04{-}28) & \tabitem{} Non interactive alignment implemented.\\
1421 | 		& \tabitem{} Table of contents in readme.\\
1422 | 	\midrule
1423 | 	0.06 (2014{-}04{-}25) & \tabitem{} Conversion to editor scripts.\\
1424 | 	\midrule
1425 | 	0.05 (2014{-}04{-}03) & \tabitem{} Better readme.\\
1426 | 		& \tabitem{} Functional program for linux.\\
1427 | 	\midrule
1428 | 	0.04 (2014{-}04{-}03) & \tabitem{} Pronunciation variants implemented.\\
1429 | 	\midrule
1430 | 	0.03 (2014{-}03{-}31) & \tabitem{} Aligner works in python.\\
1431 | 	\midrule
1432 | 	0.02 (2014{-}03{-}27) & \tabitem{} Python script around aligner started.\\
1433 | 		& \tabitem{} Phonetizer skeleton done.\\
1434 | 	\bottomrule
1435 | 	\caption{Version history}
1436 | \end{longtable}
1437 | 
1438 | \bibliographystyle{ieeetr}
1439 | \bibliography{book}
1440 | 
1441 | \end{document}
1442 | 


--------------------------------------------------------------------------------
/book/preamble.tex:
--------------------------------------------------------------------------------
 1 | \documentclass[a4paper]{book}
 2 | 
 3 | \usepackage{graphicx}
 4 | \usepackage[dvipdfm]{hyperref}
 5 | \usepackage[utf8]{inputenc}
 6 | \usepackage{authblk}
 7 | \usepackage{float}
 8 | \usepackage{booktabs}
 9 | \usepackage{longtable}
10 | \usepackage{multirow}
11 | \usepackage{listings}
12 | \usepackage{geometry}
13 | 
14 | \author[1]{Mart Lubbers}
15 | \author[2]{Francisco Torreira}
16 | \affil[1]{\url{mart@martlubbers.net}}
17 | \affil[2]{\url{francisco.torreira@mpi.nl}}
18 | \title{Praatalign: an interactive Praat plug-in for performing phonetic forced
19 | alignment\\\large A detailed manual for version $2.0$}
20 | \date{\today} 
21 | \hypersetup{%
22 | 	pdftitle={Praatalign detailed manual},
23 | 	pdfauthor={Mart Lubbers and Francisco Torreira},
24 | 	pdfsubject={Praatalign},
25 | 	pdfproducer={Mart Lubbers},
26 | 	pdfkeywords={praat,htk,phonetic forced alignment},
27 | 	hidelinks
28 | }
29 | 
30 | \lstset{%
31 | 	basicstyle=\footnotesize,
32 | 	breaklines=true,
33 | 	captionpos=b,
34 | 	frame=L,
35 | 	numbers=left
36 | }
37 | 
38 | \newcommand{\tabitem}{~~\llap{\textbullet}~~}
39 | 
40 | \graphicspath{{./img/}}
41 | 


--------------------------------------------------------------------------------
/cleaninterval.praat:
--------------------------------------------------------------------------------
 1 | include procs.praat
 2 | 	selection_start = Get start of selection
 3 | 	selection_end = Get end of selection
 4 | 	editor_info$ = TextGrid info
 5 | 	textgrid_object$ = "TextGrid " + extractLine$(editor_info$, "Object name: ")
 6 | 	editor_info$ = Editor info
 7 | 	selected_tier_number = extractNumber(editor_info$, "Selected tier:")
 8 | endeditor
 9 | 	@cleanAnnotation: textgrid_object$, selected_tier_number, selection_start, selection_end
10 | 


--------------------------------------------------------------------------------
/generatedict.praat:
--------------------------------------------------------------------------------
 1 | # Extract settings
 2 |     if not fileReadable("settings")
 3 |         exitScript("No settings file found, please run the setup first")
 4 |     endif
 5 |     settings$ = readFile$("settings")
 6 |     tmpfile$ = extractLine$(settings$, "OUT: ")
 7 |     editor_info$ = Editor info
 8 |     curtier = extractNumber(editor_info$, "Selected tier:")
 9 | 
10 | # Extract tier
11 |     Extract entire selected tier
12 | endeditor
13 | textgrid_object$ = selected$("TextGrid", 1)
14 | 
15 | # Create table and save
16 | Down to Table... "no" 6 "yes" "no"
17 | Save as tab-separated file... 'tmpfile$'
18 | 
19 | # Clean up
20 | Remove
21 | select TextGrid 'textgrid_object$'
22 | Remove
23 | 
24 | # Ask for location and write the dictionary
25 | dict_filepath$ = chooseWriteFile$("Pick the location for the dictionary",
26 | ..."missing.txt")
27 | system python generatedict.py 'dict_filepath$'
28 | pause "Dictionary successfully generated"
29 | 


--------------------------------------------------------------------------------
/generatedict.py:
--------------------------------------------------------------------------------
 1 | #!/bin/env python
 2 | # -*- coding: utf-8 -*-
 3 | 
 4 | import codecs
 5 | import phonetizer as ph
 6 | import sys
 7 | import os
 8 | 
 9 | # Load the settings
10 | with open('settings', 'r') as f:
11 |     sett = {k: v.strip() for k, v in map(lambda x: x.split(': '), f)}
12 | 
13 | try:
14 |     os.remove('temp.status')
15 | except:
16 |     pass
17 | 
18 | # Load the phonetizer
19 | try:
20 |     phone = ph.getphonetizer(
21 |         sett['LAN'], sett['PHO'], sett['DCT'], sett['RUL'])
22 | except UnicodeError:
23 |     with open('temp.status', 'w') as f:
24 |         f.write('Unicode error. Check if the files are utf-8')
25 |     exit()
26 | except IOError as e:
27 |     if e.filename == sett['DCT']:
28 |         error = 'Dictionary file couldn\'t be found\n'\
29 |             'It is searched for in {}'.format(e.filename)
30 |     elif e.filename == sett['RUL']:
31 |         error = 'Ruleset file couldn\'t be found\n'\
32 |             'It is searched for in {}'.format(e.filename)
33 |     elif e.filename == sett['PHO']:
34 |         error = 'Universal phonetizer file couldn\'t be found\n'\
35 |             'It is searched for in {}'.format(e.filename)
36 |     else:
37 |         error = 'Some io error: ' + str(e)
38 |     with open('temp.status', 'w') as f:
39 |         f.write(error)
40 |     exit()
41 | 
42 | # Open the output file and put all missing words in the set
43 | mis = set()
44 | with codecs.open(sett['OUT'], 'r', 'utf-8') as i:
45 |     for line in i:
46 |         words = line.split('\t')[2]
47 |         for word in words.split():
48 |             if phone.phonetizeword(word) is None:
49 |                 mis.add(word)
50 | 
51 | # Write the set to the specified file
52 | with codecs.open(' '.join(sys.argv[1:]), 'w', 'utf-8') as o:
53 |     for word in mis:
54 |         o.write(u'{}\t\n'.format(word))
55 | 


--------------------------------------------------------------------------------
/install.sh:
--------------------------------------------------------------------------------
 1 | #!/bin/bash
 2 | set -e
 3 | if [ "$(id -u)" == "0" ]; then
 4 | 	echo "This script should not need root rights..." 1>&2
 5 | 	echo "Please press CRTL+C and rerun the script as a normal user"
 6 | 	echo "If you still wish to continue type PLEASE"
 7 | 	read line
 8 | 	if [ "$line" -ne "PLEASE" ]; then
 9 | 		exit
10 | 	fi
11 | fi
12 | case "$OSTYPE" in
13 | 	solaris* | *bsd* | linux*)
14 | 		echo "Detected LINUX/SOLARIS/BSD"
15 | 		dir="$HOME/.praat-dir/plugin_pralign/"
16 | 		;;
17 | 	darwin*)
18 | 		echo "Detected MAC"
19 | 		dir="$HOME/Library/Preferences/Praat Prefs/plugin_pralign"
20 | 		;;
21 | 	msys* | win32*)
22 | 		echo "You are running windows? Please run install_win.bat"
23 | 		exit 2
24 | 		;;
25 | 	*)
26 | 		echo "Unknown operating system..."
27 | 		exit 2
28 | esac
29 | scriptdir="$(cd "$(dirname "$0")" && pwd)"
30 | 
31 | set -x
32 | mkdir -p "$dir"
33 | rm -fr "$dir"/* || true
34 | cp -fR "$scriptdir/"*.{py,praat} "$dir"
35 | cp -fR "$scriptdir/"par.* "$dir"
36 | set +x
37 | echo "Installing complete"
38 | 


--------------------------------------------------------------------------------
/install_win.bat:
--------------------------------------------------------------------------------
 1 | setlocal enabledelayedexpansion
 2 | @ECHO OFF
 3 | SET _plugindir=%USERPROFILE%\Praat\plugin_pralign
 4 | RD /Q /S "%_plugindir%"
 5 | MD "%_plugindir%"
 6 | COPY %~dp0*.py "%_plugindir%"
 7 | COPY %~dp0*.praat "%_plugindir%"
 8 | MD "%_plugindir%\par.spanish"
 9 | COPY par.spanish\* "%_plugindir%\par.spanish"
10 | MD "%_plugindir%\par.dutch"
11 | COPY par.dutch\* "%_plugindir%\par.dutch"
12 | echo Installing completed, please press enter to close this window...
13 | pause > nul
14 | 


--------------------------------------------------------------------------------
/mkrelease.sh:
--------------------------------------------------------------------------------
1 | #!/bin/bash
2 | VERSION=2.0a
3 | zip -9rv praatalign_$VERSION.zip *.{praat,py,md} {install,par,LICENCE}*
4 | tar -cvf praatalign_$VERSION.tar *.{praat,py,md} {install,par,LICENCE}*
5 | xz -ve praatalign_$VERSION.tar
6 | 


--------------------------------------------------------------------------------
/par.dutch/DICT:
--------------------------------------------------------------------------------
 1 | <usb> <nib>
 2 | <nib> <nib>
 3 | <p:> <p:>
 4 | a:~ a:
 5 | o:~ o:
 6 | E:~ E:
 7 | O:~ O
 8 | a:i aI
 9 | o:i OY
10 | e:u OY
11 | P2: P2:
12 | e: e:
13 | E~ E
14 | a: a:
15 | o: o:
16 | E: E:
17 | P9: P9
18 | O: O
19 | EI aI
20 | @U aU
21 | Ei aI
22 | P9y OY
23 | Au aU
24 | ui OY 
25 | iu u:
26 | yu u:
27 | O~ O
28 | p p
29 | b b
30 | t t
31 | T s
32 | d d
33 | D v
34 | k k
35 | g g
36 | G C
37 | f f
38 | v v
39 | w v
40 | s s
41 | z z
42 | S S
43 | Z S
44 | j j
45 | x x
46 | h h
47 | m m
48 | n n
49 | N N
50 | l l
51 | r r
52 | I I
53 | i i:
54 | E E
55 | A O
56 | O O
57 | u u:
58 | Y Y
59 | y y:
60 | @ @
61 | ? ?
62 | < <p:>
63 | > <p:>
64 | # <p:>
65 | 


--------------------------------------------------------------------------------
/par.dutch/HVITECONF:
--------------------------------------------------------------------------------
1 | ALLOWCXTEXP=F
2 | ALLOWXWRDEXP=F
3 | FORCEOUT=T
4 | 


--------------------------------------------------------------------------------
/par.dutch/MONOPHONES:
--------------------------------------------------------------------------------
 1 | P2:
 2 | @
 3 | C
 4 | E
 5 | E:
 6 | I
 7 | N
 8 | O
 9 | OY
10 | ?
11 | S
12 | U
13 | Z
14 | P6
15 | P9
16 | a
17 | a:
18 | aI
19 | aU
20 | b
21 | d
22 | e
23 | e:
24 | f
25 | g
26 | h
27 | i:
28 | j
29 | k
30 | l
31 | m
32 | n
33 | <nib>
34 | o
35 | o:
36 | p
37 | <p:>
38 | r
39 | s
40 | t
41 | u:
42 | v
43 | x
44 | y:
45 | z
46 | Y
47 | <
48 | >
49 | 


--------------------------------------------------------------------------------
/par.dutch/PRECONFIGNIST:
--------------------------------------------------------------------------------
 1 | SUOURCEKIND = WAVEFORM
 2 | TARGETKIND = MFCC_E_D_A
 3 | TARGETFORMAT = HTK
 4 | SOURCEFORMAT = NIST
 5 | SOURCERATE = 625
 6 | TARGETRATE = 100000
 7 | WINDOWSIZE = 200000
 8 | NUMCHANS = 26
 9 | CEPLIFTER = 22
10 | NUMCEPS = 12
11 | 


--------------------------------------------------------------------------------
/par.spanish/DICT:
--------------------------------------------------------------------------------
 1 | <usb> <nib>
 2 | <nib> <nib>
 3 | <p:> <p:>
 4 | p_h p
 5 | t_h t
 6 | k_h k
 7 | P2: P2:
 8 | aI aI
 9 | aU aU
10 | OI OY
11 | eI aI
12 | eU OY
13 | @U aU
14 | rr r
15 | jj C
16 | e~ e
17 | a~ a
18 | o~ o:
19 | E~ E
20 | P9 P9
21 | P6 P6
22 | y: y:
23 | tt t
24 | pp p
25 | kk k
26 | p p
27 | b b
28 | t t
29 | T s
30 | d d
31 | D v
32 | k k
33 | g g
34 | f f
35 | B v
36 | s s
37 | z z
38 | S S
39 | w v
40 | x x
41 | h h
42 | G C
43 | v m
44 | n n
45 | m m
46 | J N
47 | N N
48 | F m
49 | l l
50 | L j
51 | j j
52 | r r
53 | i i:
54 | e e
55 | E E
56 | a a
57 | o o:
58 | u u:
59 | @ @
60 | Y Y
61 | < <
62 | > >
63 | # <p:>
64 | 


--------------------------------------------------------------------------------
/par.spanish/HVITECONF:
--------------------------------------------------------------------------------
1 | ALLOWCXTEXP=F
2 | ALLOWXWRDEXP=F
3 | FORCEOUT=T
4 | 


--------------------------------------------------------------------------------
/par.spanish/MONOPHONES:
--------------------------------------------------------------------------------
 1 | P2:
 2 | @
 3 | C
 4 | E
 5 | E:
 6 | I
 7 | N
 8 | O
 9 | OY
10 | S
11 | U
12 | Z
13 | P6
14 | P9
15 | a
16 | a:
17 | aI
18 | aU
19 | b
20 | d
21 | e
22 | e:
23 | f
24 | g
25 | h
26 | i:
27 | j
28 | k
29 | l
30 | m
31 | n
32 | <nib>
33 | o
34 | o:
35 | p
36 | <p:>
37 | r
38 | s
39 | t
40 | u:
41 | v
42 | x
43 | y:
44 | z
45 | Y
46 | <
47 | >
48 | 


--------------------------------------------------------------------------------
/par.spanish/PRECONFIGNIST:
--------------------------------------------------------------------------------
 1 | SOURCEKIND = WAVEFORM
 2 | TARGETKIND = MFCC_E_D_A
 3 | TARGETFORMAT = HTK
 4 | SOURCEFORMAT = NIST
 5 | SOURCERATE = 625
 6 | TARGETRATE = 100000
 7 | WINDOWSIZE = 200000
 8 | NUMCHANS = 26
 9 | CEPLIFTER = 22
10 | NUMCEPS = 12
11 | 


--------------------------------------------------------------------------------
/phonetizer.py:
--------------------------------------------------------------------------------
  1 | # -*- coding: utf-8 -*-
  2 | 
  3 | import codecs
  4 | import re
  5 | import unicodedata
  6 | import itertools
  7 | 
  8 | 
  9 | class tdict(dict):
 10 |     """Dictionary with a missing value that is the normalized version
 11 |     of the input
 12 |     """
 13 |     def __missing__(self, key):
 14 |         char = unicodedata.normalize('NFKD', unicode(key))
 15 |         return str(char.encode('ascii', 'ignore')).lower()
 16 | 
 17 | 
 18 | class Phonetizer:
 19 |     """Skeleton class for a phonetizer for the praat align program"""
 20 |     reps = [(r'\v', '[aoeiu]'), (r'\c', '[^aoeui]'), ('\n', '')]
 21 | 
 22 |     def __init__(self, univ_phon=None, dictpath=None, ruleset=None):
 23 |         """Constructor with an optional dictionary
 24 | 
 25 |         filters   -- A regular expression of things to filter out of the
 26 |                      utterances
 27 |         dictpath  -- Path to an optional dictonary
 28 |         ruleset   -- Path to an optional ruleset file
 29 |         univ_phon -- Path to an optional universal phonetizer file
 30 |         """
 31 |         self.r = []
 32 |         self.dictionary = dict()
 33 |         self.univ_phon = []
 34 |         if dictpath is not None:
 35 |             with codecs.open(dictpath, 'r', 'utf-8') as f:
 36 |                 for l in f:
 37 |                     if l and l[0] != '#':
 38 |                         l = [s.strip() for s in l.split('\t')]
 39 |                         self.dictionary[l[0]] = [var.split() for var in
 40 |                                                  map(str, l[1:])]
 41 |         if ruleset is not None:
 42 |             with codecs.open(ruleset, 'r', 'utf-8') as f:
 43 |                 for l in itertools.ifilter(lambda x: x[0] != '#', f):
 44 |                     if not l.startswith('\t'):
 45 |                         l = re.sub(r'([.^$*+?{}[\]\|()])', r'\\\1', l)
 46 |                     else:
 47 |                         for target, repl in self.reps:
 48 |                             l = l.replace(target, repl)
 49 |                     self.r.append(l.strip().split('\t'))
 50 |         if univ_phon is not None:
 51 |             with codecs.open(univ_phon, 'r', 'utf-8') as f:
 52 |                 for line in f:
 53 |                     line = [l.strip() for l in line.split('\t')]
 54 |                     if line:
 55 |                         self.univ_phon.append((line[0], line[1].split(' ')))
 56 | 
 57 |     def todawg(self, pron):
 58 |         """Converts the pronunciation variants and rules to a graph
 59 |         representation.
 60 | 
 61 |         pron     -- Pronunciation
 62 |         """
 63 |         # Make a translation for all multi character phones
 64 |         c2 = [ch for wd in pron for var in wd for ch in var if len(ch) > 1]
 65 |         c2 = dict(zip(c2, map(chr, range(1, len(c2)+1))))
 66 | 
 67 |         # Create all possible combinations of pronunciation variants
 68 |         pron = ((' '.join(varnt + ['#'])
 69 |                  for varnt in word) for word in pron)
 70 | 
 71 |         # Make strings of these variants
 72 |         pron = (' '.join(x) for x in itertools.product(*pron))
 73 |         # Create combinations of all rulesets
 74 |         rcs = [()]
 75 |         for i in range(len(self.r)):
 76 |             rcs += itertools.combinations(self.r, i+1)
 77 |         all_prons = set()
 78 |         # For every variant
 79 |         for variant in pron:
 80 |             # For every combination of rules
 81 |             for combo in rcs:
 82 |                 all_prons.add(variant)
 83 |                 # For every rule combination
 84 |                 for pat, target in combo:
 85 |                     app = sorted(x.span() for x in re.finditer(pat, variant))
 86 |                     apporders = [()]
 87 |                     for i in range(len(app)):
 88 |                         apporders += itertools.permutations(app, i+1)
 89 |                     # For every appliance order
 90 |                     for appc in reversed(apporders):
 91 |                         wordapp = variant
 92 |                         for st, en in appc:
 93 |                             try:
 94 |                                 wordapp = '{}{}{}'.format(
 95 |                                     wordapp[:st],
 96 |                                     re.sub(pat, target, wordapp[st:en]),
 97 |                                     wordapp[en:])
 98 |                             except:
 99 |                                 pass
100 |                         all_prons.add(wordapp)
101 | 
102 |         # Replace all the multichar phones with their appropriate byte
103 |         for f, t in c2.iteritems():
104 |             all_prons = {x.replace(f, t) for x in all_prons}
105 |         all_prons = {x.replace(' ', '') for x in all_prons}
106 |         c2 = {v: k for k, v in c2.iteritems()}
107 | 
108 |         # Add all the words to the DAWG
109 |         import pyDAWG
110 |         pyd = pyDAWG.DAWG()
111 |         for word in sorted(all_prons):
112 |             pyd.add_word(word)
113 | 
114 |         # Create a slf with breath first search and use a translation list to
115 |         # map the unnumbered nodes to their numbered variants
116 |         nodenum = 0
117 |         final_nodes, nodes, edges, translation = [], [], [], []
118 |         queue = [(0, pyd.q0)]
119 |         visited = set()
120 |         if pyd.q0.final:
121 |             final_nodes.append(nodenum)
122 |         else:
123 |             nodes.append(nodenum)
124 |         nodenum += 1
125 |         # While we still have nodes to visit
126 |         while queue:
127 |             # If we have not visited the newest node
128 |             current = queue.pop()
129 |             if not current[0] in visited:
130 |                 # Visit the node
131 |                 visited.add(current[0])
132 |                 # Go throught all the children
133 |                 for char, child in current[1].children.iteritems():
134 |                     # Find translation for the child
135 |                     matches = [c for c in translation if c[0] == child]
136 |                     curnum = -1
137 |                     # Translate if there is a translation
138 |                     if matches:
139 |                         curnum = matches[-1][1]
140 |                     # Otherwise create a translation
141 |                     else:
142 |                         translation.append((child, nodenum))
143 |                         curnum = nodenum
144 |                         nodenum += 1
145 |                     # If the node is final, mark is as such
146 |                     if child.final:
147 |                         final_nodes.append(curnum)
148 |                     else:
149 |                         nodes.append(curnum)
150 |                     # Append the edge to the list and append the child to the
151 |                     # queue
152 |                     edges.append((current[0], char, curnum))
153 |                     queue.append((curnum, child))
154 |         # Convert the edges to a dictionary with for every number a set of
155 |         # connections and convert the nodes to a dictionary with for every
156 |         # number the appropriate character. Also convert the multichar phones
157 |         # back to their original form.
158 |         nodes += final_nodes
159 |         nnodes = {0: '<'}
160 |         nedges = {}
161 |         for fr, ch, to in edges:
162 |             nnodes[to] = c2[ch] if ch in c2 else ch
163 |             if fr not in nedges:
164 |                 nedges[fr] = set()
165 |             nedges[fr].add(to)
166 |         # Find the last node and remember the position to connect the end of
167 |         # the words to it
168 |         finalnode = len(nnodes)
169 |         nnodes[finalnode] = '>'
170 |         for final in final_nodes:
171 |             if final not in nedges:
172 |                 nedges[final] = set()
173 |             nedges[final].add(finalnode)
174 |         return nedges, nnodes
175 | 
176 |     def todot(self, nedges, nnodes):
177 |         dot = 'digraph {rankdir=LR;'
178 |         dot += ''.join('q{} [label="{}"];'.format(*x) for x in nnodes.items())
179 |         for fr, to in nedges.iteritems():
180 |             for c in to:
181 |                 dot += 'q{} -> q{};'.format(fr, c)
182 |         return dot + '}'
183 | 
184 |     def toslf(self, nedges, nnodes):
185 |         slf = 'N={} L={}\n'.format(
186 |             len(nnodes), sum(len(to) for _, to in nedges.iteritems()))
187 |         # Write all the nodes
188 |         slf += ''.join(
189 |             'I={} W={}\n'.format(k, v) for k, v in nnodes.iteritems())
190 |         # Write all the edges
191 |         i = 0
192 |         for fr, to in nedges.iteritems():
193 |             for c in to:
194 |                 slf += 'J={} S={} E={}\n'.format(i, fr, c)
195 |                 i += 1
196 |         return slf
197 | 
198 |     def permute(self, items, output):
199 |         """Helper function that creates all possible combinations of a list
200 | 
201 |         items  -- Possible items
202 |         output -- Results list
203 |         """
204 |         # Append the current combination
205 |         output.append(items)
206 |         # For all nodes left remove 1 and put the resulting list back in
207 |         for i, item in enumerate(items):
208 |             items2 = items[:]
209 |             items2.remove(item)
210 |             self.permute(items2, output)
211 | 
212 |     def applyrules(self, phon):
213 |         """Apply the rules specified in the ruleset
214 | 
215 |         phon - phonetization of the utterance
216 |         """
217 |         if phon is not None and len(phon) == 1:
218 |             word = ''.join(phon[0])
219 |             for rule in self.r:
220 |                 mo = rule.search(word)
221 |                 if mo is not None:
222 |                     wordafter = word[:mo.end('fr')]+word[mo.start('to'):]
223 |                     phon.append(list(wordafter))
224 |         return phon
225 | 
226 |     def phonetize(self, utterance):
227 |         """Phonetizes one utterance
228 | 
229 |         utterance -- The utterance to phonetize
230 |         """
231 |         try:
232 |             pron = [self.phonetizeword(unicode(word, 'utf-8')) for
233 |                     word in utterance.split()]
234 |         except TypeError:
235 |             pron = [self.phonetizeword(unicode(word)) for
236 |                     word in utterance.split()]
237 |         pron = filter(None, pron)
238 |         return pron
239 | 
240 |     def phonetizeword(self, word):
241 |         """Returns a list of phones generated from the utterance and should
242 |         return None when unable to phonetize
243 | 
244 |         word -- the word to phonetize
245 |         """
246 |         if word in self.dictionary:
247 |             return self.dictionary[word]
248 |         else:
249 |             return None
250 | 
251 | 
252 | class PhonetizerTzeltal(Phonetizer):
253 |     """Phonetizer for the tzeltal language"""
254 | 
255 |     trans = tdict({'j': 'x', 'w': 'b', 'x': 'S', '\'': '?', 'y': 'j'})
256 | 
257 |     def phonetizeword(self, word):
258 |         word = re.sub('[.,\-]', '', word.lower())
259 |         if word in self.dictionary:
260 |             return self.dictionary[word]
261 |         phonemap = list()
262 |         it = iter(enumerate(word))
263 |         for i, character in it:
264 |             if character == 'c' and i+1 < len(word) and word[i+1] == 'h':
265 |                 if i+2 < len(word) and word[i+2] == '\'':
266 |                     next(it, None)
267 |                 phonemap.append('ts_j')
268 |                 next(it, None)
269 |             elif character == 't' and i+1 < len(word) and word[i+1] == '\'':
270 |                 phonemap.append('t')
271 |                 next(it, None)
272 |             elif character == 't' and i+1 < len(word) and word[i+1] == 'z':
273 |                 if i+2 < len(word) and word[i+2] == '\'':
274 |                     next(it, None)
275 |                 phonemap.append('ts_j')
276 |                 next(it, None)
277 |             else:
278 |                 phonemap.append(self.trans[character])
279 |         self.dictionary[word] = [phonemap]
280 |         return [phonemap]
281 | 
282 | 
283 | class PhonetizerSpanish(Phonetizer):
284 |     """Phonetizer for the spanish language"""
285 | 
286 |     acronymmap = tdict({
287 |         'B': ['b', 'e'], 'C': ['T', 'e'], 'D': ['d', 'e'],
288 |         'F': ['e', 'f', 'e'], 'G': ['x', 'e'], 'H': ['a', 't', 'S', 'e'],
289 |         'J': ['x', 'o', 't', 'a'], 'K': ['k', 'a'], 'L': ['e', 'l', 'e'],
290 |         'M': ['e', 'm', 'e'], 'N': ['e', 'n', 'e'], 'Q': ['k', 'u'],
291 |         'R': ['e', 'r', 'r', 'e'], 'S': ['e', 's', 'e'], 'T': ['t', 'e'],
292 |         'V': ['u', 'b', 'u'], 'W': ['u', 'b', 'u', 'd', 'o', 'b', 'l', 'e'],
293 |         'X': ['e', 'k', 'i', 's'], 'Y': ['i', 'g', 'r', 'i', 'e', 'g', 'a'],
294 |         'Z': ['T', 'e', 't', 'a']
295 |         })
296 | 
297 |     trans = tdict({'ñ'.decode('utf-8'): 'J', 'ç'.decode('utf-8'): 'T',
298 |                    'j': 'x', 'c': 'k', 'v': 'b', 'w': 'b', 'z': 'T', 'y': 'j'})
299 | 
300 |     def phonetizeword(self, word):
301 |         word = ''.join(ch for ch in word if
302 |                        unicodedata.category(ch).startswith('L') or
303 |                        ch in '[]<>()&')
304 |         word = re.sub('[<>]', '&', re.sub('\(.*\)', '', word))
305 |         uppercases = len([ch for ch in word if
306 |                          unicodedata.category(ch).startswith('Lu')])
307 |         if word in self.dictionary:
308 |             return self.dictionary[word]
309 |         # Check sounds and foreign language.
310 |         if word and '&' == word[0] or '[lang' in word:
311 |             return None
312 |         # Remove optional laughter and breathing and try again to look it up
313 |         if '[' in word or ']' in word:
314 |             return self.phonetizeword(re.sub('\[.*\]', '', word))
315 |         # Allocate the map
316 |         phonemap = list()
317 |         if uppercases == len(word):
318 |             return [[a for b in [self.acronymmap[i] for i in word] for a in b]]
319 |         else:
320 |             lw = word.lower()
321 |             it = iter(enumerate(lw))
322 |             for i, ch in it:
323 |                 if ch == 'c' and i+1 < len(lw) and\
324 |                         lw[i+1] in 'eií'.decode('utf-8'):
325 |                     phonemap.append('T')
326 |                 elif ch == 'c' and i+1 < len(lw) and lw[i+1] == 'h':
327 |                     phonemap += ['t', 'S']
328 |                     next(it, None)
329 |                 elif ch == 'g' and i+1 < len(lw) and\
330 |                         lw[i+1] == 'ü'.decode('utf-8'):
331 |                     phonemap += ['g', 'u']
332 |                     next(it, None)
333 |                 elif ch == 'g' and i+2 < len(lw) and lw[i+1] == 'u' and\
334 |                         lw[i+2] in 'ei':
335 |                     phonemap.append('g')
336 |                     next(it, None)
337 |                 elif ch == 'g' and i+1 < len(lw) and lw[i+1] in 'ei':
338 |                     phonemap.append('x')
339 |                 elif ch == 'l' and i+1 < len(lw) and lw[i+1] == 'l':
340 |                     phonemap.append('jj')
341 |                     next(it, None)
342 |                 elif ch == 'q':
343 |                     phonemap.append('k')
344 |                     if i+1 < len(lw) and lw[i+1] == 'u':
345 |                         next(it, None)
346 |                 elif ch == 'r' and ((i is 0 or lw[i-1] in 'nlsm') or
347 |                                     (i+1 < len(lw) and lw[i+1] == 'r')):
348 |                     if i+1 < len(lw) and lw[i+1] == 'r':
349 |                         next(it, None)
350 |                     phonemap += ['rr']
351 |                 elif ch == 'y' and i+1 >= len(lw):
352 |                     phonemap.append('i')
353 |                 elif ch == 'x':
354 |                     phonemap += ['k', 's']
355 |                 elif ch == 'h':
356 |                     continue
357 |                 else:
358 |                     phonemap.append(self.trans[ch])
359 |         self.dictionary[word] = [phonemap]
360 |         return [phonemap]
361 | 
362 | 
363 | class PhonetizerDictionary(Phonetizer):
364 |     """Dummy phonetizer for dictionary only phonetizers"""
365 | 
366 |     def phonetizeword(self, word):
367 |         if word in self.dictionary:
368 |             return self.dictionary[word]
369 |         else:
370 |             return None
371 | 
372 | 
373 | class PhonetizerLoopback(Phonetizer):
374 |     """Dummy phonetizer that uses direct loopback to generate phonetic
375 |     transcription, therefore allowing the user to use his own phonetic
376 |     transcription"""
377 | 
378 |     def phonetize(self, utterance):
379 |         """Phonetizes one utterance
380 | 
381 |         utterance -- The utterance to phonetize
382 |         """
383 |         return [[[a]] for a in utterance.split(' ')]
384 | 
385 | 
386 | class PhonetizerUniversal(Phonetizer):
387 |     """Universal phonetizer that uses a po dictionary"""
388 | 
389 |     def phonetizeword(self, word):
390 |         """Returns a list of phones generated from the utterance and should
391 |         return None when unable to phonetize
392 | 
393 |         word -- the word to phonetize
394 |         """
395 |         if word in self.dictionary:
396 |             return self.dictionary[word]
397 |         else:
398 |             i = 0
399 |             phones = []
400 |             while i <= len(word):
401 |                 replaces = [t for t in self.univ_phon if
402 |                             word[i:].startswith(t[0])]
403 |                 if replaces:
404 |                     phones += replaces[0][1]
405 |                     i += len(replaces[0][0])
406 |                 else:
407 |                     i += 1
408 |         return None if not phones else [phones]
409 | 
410 | 
411 | class PhonetizerSkeleton(Phonetizer):
412 |     """Skeleton to create your own phonetizer"""
413 | 
414 |     def phonetizeword(self, word):
415 |         #  Substitute all transcription specific useless markings and lowercase
416 |         # word
417 |         # word = re.sub('[.,\-]', '', word.lower())
418 |         # If the word already exists in the dictionary return it immediatly
419 |         # if word in self.dictionary:
420 |         #     return self.dictionary[word]
421 |         # If some condition is not met and the word is unphonetizable you
422 |         # should return None
423 |         # if some_condition:
424 |         #     return None
425 |         # Do the magic phonetizing here
426 |         # Add the phontization to the dictionary
427 |         # self.dictionary[word] = [phonemap]
428 |         # Return the map
429 |         # return [phonemap]
430 |         pass
431 | 
432 | phonetizerdict = {
433 |     'none': PhonetizerDictionary,
434 |     'universal': PhonetizerUniversal,
435 |     'spanish': PhonetizerSpanish,
436 |     'tzeltal': PhonetizerTzeltal
437 |     }
438 | 
439 | 
440 | def getphonetizer(lang, univ_phon=None, dictpath=None, ruleset=None):
441 |     """
442 |     Gives a phonetizer by code
443 | 
444 |     lang - language code, has to be present in phonetizerdict
445 |     dictpath  - optional dictionary file
446 |     ruleset   - optional ruleset file
447 |     univ_phon - optional universal phonetizer file
448 |     """
449 |     dictpath = None if dictpath == "None" else dictpath
450 |     ruleset = None if ruleset == "None" else ruleset
451 |     univ_phon = None if univ_phon == "None" else univ_phon
452 |     return phonetizerdict[lang](univ_phon, dictpath, ruleset)
453 | 


--------------------------------------------------------------------------------
/procs.praat:
--------------------------------------------------------------------------------
  1 | procedure loadSettings: 
  2 | 	if not fileReadable("settings")
  3 | 		exitScript("No settings file found, please run the setup first")
  4 | 	endif
  5 | 	settings$ = readFile$("settings")
  6 | 	phonetier_name$ = extractLine$(settings$, "NEW: ")
  7 | 	wordtier_name$ = extractLine$(settings$, "WRD: ")
  8 | 	cantier_name$ = extractLine$(settings$, "CAN: ")
  9 | 	llhtier_name$ = extractLine$(settings$, "LLH: ")
 10 | 	orttier_name$ = extractLine$(settings$, "ORT: ")
 11 | 	tmpfile$ = extractLine$(settings$, "OUT: ")
 12 | 	pythonex$ = extractLine$(settings$, "PY2: ")
 13 | 	boundary_margin = extractNumber(settings$, "THR: ")
 14 | endproc
 15 | 
 16 | procedure loadFileInfo:
 17 | # Try longsound first
 18 | 	longsound_info$ = ""
 19 | 	longsound_info$ = nocheck LongSound info
 20 | 	if longsound_info$ <> ""
 21 | 		sound_file$ = extractLine$(longsound_info$, "File name: ")
 22 | 		sound_object$ = "LongSound " + extractLine$(longsound_info$, "Object name: ")
 23 | 		sound_duration = extractNumber(longsound_info$, "Duration: ")
 24 | 	else
 25 | 		sound_info$ = Sound info
 26 | 		sound_object$ = "Sound " + extractLine$(sound_info$, "Object name: ")
 27 | 		sound_duration = extractNumber(sound_info$, "   Total duration: ")
 28 | 		Select: 0, sound_duration
 29 | 		Save selected sound as WAV file: "longsound_sound.wav"
 30 | 		sound_file$ = "longsound_sound.wav"
 31 | 	endif
 32 | 
 33 | 	textgrid_info$ = TextGrid info
 34 | 	textgrid_object$ = "TextGrid " + extractLine$(textgrid_info$, "Object name: ")
 35 | 
 36 | 	editorinfo$ = Editor info
 37 | 	selected_tier = extractNumber(editorinfo$, "Selected tier:")
 38 | 	pitch_on = if extractWord$(editorinfo$, "Pitch show: ") == "yes"
 39 | ... then 1 else 0 fi
 40 | 	intensity_on = if extractWord$(editorinfo$, "Intensity show: ") == "yes"
 41 | ... then 1 else 0 fi
 42 | 	spectrum_on = if extractWord$(editorinfo$, "Spectrogram show: ") == "yes"
 43 | ... then 1 else 0 fi
 44 | 	formant_on = if extractWord$(editorinfo$, "Formant show: ") == "yes"
 45 | ... then 1 else 0 fi
 46 | 	pulses_on = if extractWord$(editorinfo$, "Pulses show: ") == "yes"
 47 | ... then 1 else 0 fi
 48 | endproc
 49 | 
 50 | procedure indexOfTier: .name$
 51 | 	.inserted = 0
 52 | 	if .name$ <> ""
 53 | 		.number = -1
 54 | 		.ntiers = Get number of tiers
 55 | 		for .i to .ntiers
 56 | 			.tiername$ = Get tier name: .i
 57 | 			if .tiername$ = .name$
 58 | 				.number = .i
 59 | 			endif
 60 | 		endfor
 61 | 		if .number = -1
 62 | 			selected_tier = selected_tier + 1
 63 | 			Insert interval tier: 1, .name$
 64 | 			.number = 1
 65 | 			.inserted = 1
 66 | 		endif
 67 | 	endif
 68 | endproc
 69 | 
 70 | procedure cleanAnnotation: .obj$, .num, .start, .end
 71 | 	editor: .obj$
 72 | 		.selected_tier_num = -1
 73 | 		repeat
 74 | 			Select next tier
 75 | 			.editor_info$ = Editor info
 76 | 			.selected_tier_num = extractNumber(.editor_info$, "Selected tier:")
 77 | 		until .num = .selected_tier_num
 78 | 		Select: .start, .end
 79 | 
 80 | 		Move cursor to: .start
 81 | 		.interval_end = Get end point of interval
 82 | 
 83 | 		.to_remove_len = 0
 84 | 		while .interval_end < .end
 85 | 			.to_remove_len = .to_remove_len + 1
 86 | 			Select next interval
 87 | 			.to_remove[.to_remove_len] = Get starting point of interval
 88 | 			.interval_end = Get end point of interval
 89 | 		endwhile
 90 | 	endeditor
 91 | 
 92 | 	selectObject: .obj$
 93 | 	for .i to .to_remove_len
 94 | 		nocheck Remove boundary at time: .selected_tier_num, .to_remove[.i]
 95 | 	endfor
 96 | 
 97 | 	.tiernum = Get interval at time: .selected_tier_num, .start-0.01
 98 | 	nocheck Set interval text: .selected_tier_num, .tiernum, ""
 99 | endproc
100 | 
101 | procedure toggleGUIValues:
102 | 	if pitch_on == 1
103 | 		Show pitch
104 | 	endif
105 | 	if intensity_on == 1
106 | 		Show intensity
107 | 	endif
108 | 	if spectrum_on == 1
109 | 		Show spectrogram
110 | 	endif
111 | 	if formant_on == 1
112 | 		Show formants
113 | 	endif
114 | 	if pulses_on == 1
115 | 		Show pulses
116 | 	endif
117 | endproc
118 | 
119 | procedure getBinary: message$, path$, default$, .toask
120 | 	.ex$ = ""
121 | 	if .toask == 1
122 | 		beginPause: "Instructions"
123 | 			comment: "Please point me to the " + default$ + " executable."
124 | 			comment: "When " + default$ + " is in your " + path$ + " you can cancel this."
125 | 			comment: "It can be found " + message$
126 | 		.clicked = endPause: "Cancel", "Continue", 2, 1
127 | 		if .clicked = 2
128 | 			.ex$ = chooseReadFile$("Point me to the " + default$ + " executable")
129 | 		endif
130 | 		if .ex$ = ""
131 | 			.ex$ = default$
132 | 		endif
133 | 	endif
134 | endproc
135 | 
136 | procedure insertTableTextGrid: .tablefile$, .obj$, .phon$, .ort$, .wrd$, .can$, .llh$,
137 | ... .phonn, .ortn, .wrdn, .cann, .llhn
138 | 	nocheck Read Table from comma-separated file: .tablefile$
139 | 	if extractWord$(selected$(), "") == "Table"
140 | 		.number_rows = Get number of rows
141 | 		# Put the results in the textgrid
142 | 		for .i to .number_rows
143 | 			#Extract the values
144 | 			selectObject: "Table praat_temp_out"
145 | 			.current_start$ = Get value: .i, "start"
146 | 			.current_start = number(.current_start$)
147 | 			.current_end$ = Get value: .i, "end"
148 | 			.current_end = number(.current_end$)
149 | 			.current_value$ = Get value: .i, "label"
150 | 			.current_type$ = Get value: .i, "type"
151 | 			selectObject: .obj$
152 | 		
153 | 			if .current_type$ = "p" and .phon$ <> ""
154 | 				nocheck Insert boundary: .phonn, .current_start
155 | 				nocheck Insert boundary: .phonn, .current_end
156 | 				.intnum = Get interval at time: .phonn, .current_start+0.0001
157 | 				nocheck Set interval text: .phonn, .intnum, .current_value$
158 | 			elif .current_type$ = "w" and .wrd$ <> ""
159 | 				nocheck Insert boundary: .wrdn, .current_start
160 | 				nocheck Insert boundary: .wrdn, .current_end
161 | 				.intnum = Get interval at time: .wrdn, .current_start+0.0001
162 | 				nocheck Set interval text: .wrdn, .intnum, .current_value$
163 | 			elif .current_type$ = "o" and .wrd$ <> ""
164 | 				nocheck Insert boundary: .ortn, .current_start
165 | 				nocheck Insert boundary: .ortn, .current_end
166 | 				.intnum = Get interval at time: .ortn, .current_start+0.0001
167 | 				nocheck Set interval text: .ortn, .intnum, .current_value$
168 | 			elif .current_type$ = "c" and .can$ <> ""
169 | 				nocheck Insert boundary: .cann, .current_start
170 | 				nocheck Insert boundary: .cann, .current_end
171 | 				.intnum = Get interval at time: .cann, .current_start+0.0001
172 | 				nocheck Set interval text: .cann, .intnum, .current_value$
173 | 			elif .current_type$ = "l" and .llh$ <> ""
174 | 				nocheck Insert boundary: .llhn, .current_start
175 | 				nocheck Insert boundary: .llhn, .current_end
176 | 				.intnum = Get interval at time: .llhn, .current_start+0.0001
177 | 				nocheck Set interval text: .llhn, .intnum, .current_value$
178 | 			endif
179 | 		endfor
180 | 		
181 | 		# Remove temporary table file
182 | 		selectObject: "Table praat_temp_out"
183 | 		Remove
184 | 	endif
185 | endproc
186 | 


--------------------------------------------------------------------------------
/pyDAWG.py:
--------------------------------------------------------------------------------
  1 | #!/bin/env python
  2 | # -*- coding: utf-8 -*-
  3 | 
  4 | """
  5 |     This is part of pydawg Python module.
  6 | 
  7 |     Pure python implementation.
  8 | 
  9 |     Author    : Wojciech Muła, wojciech_mula@poczta.onet.pl
 10 |     WWW       : http://0x80.pl/proj/pydawg/
 11 |     License   : Public domain
 12 |     Date      : $Date$
 13 | 
 14 |     $Id$
 15 | """
 16 | 
 17 | 
 18 | class DAWGNode:
 19 |     __slots__ = ["children", "final", "number"]
 20 | 
 21 |     def __init__(self, char):
 22 |         self.children = {}
 23 |         self.final = False
 24 |         self.number = None
 25 | 
 26 |     def get_next(self, char):
 27 |         try:
 28 |             return self.children[char]
 29 |         except KeyError:
 30 |             return None
 31 | 
 32 |     def set_next(self, char, child):
 33 |         self.children[char] = child
 34 | 
 35 |     def has_transition(self, char):
 36 |         return char in self.children
 37 | 
 38 |     def __str__(self):
 39 |         return "<" + "".join(self.children.keys()) + ">"
 40 | 
 41 | 
 42 | def equivalence(p, q):
 43 |     "check if states p and q are equivalent"
 44 | 
 45 |     if p.final != q.final:
 46 |         return False
 47 | 
 48 |     if len(p.children) != len(q.children):
 49 |         return False
 50 | 
 51 |     s = set(p.children)
 52 |     if s != set(q.children):
 53 |         return False
 54 | 
 55 |     """
 56 |     # exact definition of equivalence
 57 |     for c in s:
 58 |         if not equivalence(p.children[c], q.children[c]):
 59 |                 return False
 60 |     """
 61 |     # pratical implementation - constraints make
 62 |     # this much simpler and faster
 63 |     for c in s:
 64 |         if p.children[c] != q.children[c]:
 65 |             return False
 66 | 
 67 |     return True
 68 | 
 69 | 
 70 | class DAWG:
 71 |     def __init__(self):
 72 |         self._numbers_valid = False
 73 |         self.register = set()
 74 |         self.q0 = DAWGNode(None)
 75 |         self.wp = ''
 76 | 
 77 |     def add_word(self, word):
 78 |         assert word > self.wp
 79 |         return self.add_word_unchecked(word)
 80 | 
 81 |     def add_word_unchecked(self, word):
 82 |         # 1. skip existing
 83 |         i = 0
 84 |         s = self.q0
 85 |         while i < len(word) and s.has_transition(word[i]):
 86 |             s = s.get_next(word[i])
 87 |             i = i + 1
 88 | 
 89 |         assert s is not None
 90 | 
 91 |         # 2. minimize
 92 |         if i < len(self.wp):
 93 |             self._replace_or_register(s, self.wp[i:])
 94 | 
 95 |         # 3. add suffix
 96 |         while i < len(word):
 97 |             n = DAWGNode(word[i])
 98 |             s.set_next(word[i], n)
 99 |             assert n == s.get_next(word[i])
100 |             s = n
101 |             i = i + 1
102 | 
103 |         s.final = True
104 |         self.wp = word
105 |         self._numbers_valid = False
106 | 
107 |     def _replace_or_register(self, state, suffix):
108 |         stack = []
109 |         while suffix:
110 |             letter = suffix[0]
111 |             next = state.get_next(letter)
112 |             stack.append((state, letter, next))
113 | 
114 |             state = next
115 |             suffix = suffix[1:]
116 | 
117 |         while stack:
118 |             parent, letter, state = stack.pop()
119 | 
120 |             found = False
121 |             for r in self.register:
122 |                 if equivalence(state, r):
123 |                     assert(parent.children[letter] == state)
124 |                     parent.children[letter] = r
125 | 
126 |                     found = True
127 |                     break
128 | 
129 |             if not found:
130 |                 self.register.add(state)
131 | 
132 |     def freeze(self):
133 |         self._replace_or_register(self.q0, self.wp)
134 |         self._numbers_valid = False
135 | 
136 |     close = freeze
137 | 
138 |     def _num_nodes(self):
139 |         def clear_aux(node):
140 |             node.number = None
141 |             for child in node.children.values():
142 |                 clear_aux(child)
143 | 
144 |         def num_aux(node):
145 |             if node.number is None:
146 |                 n = int(node.final)
147 |                 for child in node.children.values():
148 |                     n += num_aux(child)
149 | 
150 |                 node.number = n
151 | 
152 |             return node.number
153 | 
154 |         if not self._numbers_valid:
155 |             clear_aux(self.q0)
156 |             num_aux(self.q0)
157 |             self._numbers_valid = True
158 | 
159 |     def word2index(self, word):
160 |         self._num_nodes()
161 | 
162 |         state = self.q0
163 |         index = 0
164 |         for c in word:
165 |             try:
166 |                 next = state.children[c]
167 |             except KeyError:
168 |                 return None
169 | 
170 |             for C in sorted(state.children):
171 |                 if C < c:
172 |                     index += state.children[C].number
173 |                 else:
174 |                     break
175 | 
176 |             state = next
177 |             if state.final:
178 |                 index = index + 1
179 | 
180 |         return index
181 | 
182 |     def index2word(self, index):
183 |         self._num_nodes()
184 | 
185 |         state = self.q0
186 |         count = index
187 |         output_word = ""
188 |         while True:
189 |             for c in sorted(state.children):
190 |                 tmp = state.get_next(c)
191 |                 if tmp.number < count:
192 |                     count -= tmp.number
193 |                 else:
194 |                     output_word += c
195 |                     state = tmp
196 |                     if state.final:
197 |                         count -= 1
198 | 
199 |                     break
200 |             if count <= 0:
201 |                 break
202 | 
203 |         return output_word
204 | 
205 |     def words(self):
206 |         L = []
207 | 
208 |         def aux(node, word):
209 |             if node.final:
210 |                 L.append(word)
211 | 
212 |             for letter, child in node.children.items():
213 |                 aux(child, word + letter)
214 | 
215 |         aux(self.q0, '')
216 |         return L
217 | 
218 |     def __iter__(self):
219 |         return iter(self.words())
220 | 


--------------------------------------------------------------------------------
/settings.praat:
--------------------------------------------------------------------------------
  1 | include procs.praat
  2 | 
  3 | pythonexecutablename$ = if windows then "python.exe" else "python" fi
  4 | hcopyexecutablename$ = if windows then "HCopy.exe" else "HCopy" fi
  5 | hviteexecutablename$ = if windows then "HVite.exe" else "HVite" fi
  6 | soxexecutablename$ = if windows then "sox.exe" else "sox" fi
  7 | # If the settings is already present extract all the entries
  8 | if fileReadable("settings")
  9 | 	settingsData$ = readFile$("settings")
 10 | 	dictionary$ = extractLine$(settingsData$, "DCT: ")
 11 | 	if dictionary$ = "None"
 12 | 		dictionary$ = ""
 13 | 	endif
 14 | 	ruleset$ = extractLine$(settingsData$, "RUL: ")
 15 | 	if ruleset$ = "None"
 16 | 		ruleset$ = ""
 17 | 	endif
 18 | 	new$ = extractLine$(settingsData$, "NEW: ")
 19 | 	ort$ = extractLine$(settingsData$, "ORT: ")
 20 | 	wrd$ = extractLine$(settingsData$, "WRD: ")
 21 | 	can$ = extractLine$(settingsData$, "CAN: ")
 22 | 	llh$ = extractLine$(settingsData$, "LLH: ")
 23 | 	phonetizer$ = extractLine$(settingsData$, "PHO: ")
 24 | 	if phonetizer$ = "None"
 25 | 		phonetizer$ = ""
 26 | 	endif
 27 | 	pho = 0
 28 | 	lan$ = extractLine$(settingsData$, "LAN: ")
 29 | 	if lan$ = "spanish"
 30 | 		lan = 1
 31 | 	elif lan$ = "tzeltal"
 32 | 		lan = 2
 33 | 	elif lan$ = "universal"
 34 | 		lan = 3
 35 | 	elif lan$ = "none"
 36 | 		lan = 4
 37 | 	else
 38 | 		lan = 1
 39 | 	endif
 40 | 	model$ = extractLine$(settingsData$, "MOD: ")
 41 | 	if model$ = "spanish"
 42 | 		model = 1
 43 | 	elif model$ = "dutch"
 44 | 		model = 2
 45 | 	else
 46 | 		model = 1
 47 | 	endif
 48 | 	pho = 0
 49 | 	log$ = extractLine$(settingsData$, "LOG: ")
 50 | 	soxex$ = extractLine$(settingsData$, "SOX: ")
 51 | 	hviteex$ = extractLine$(settingsData$, "HVB: ")
 52 | 	hcopyex$ = extractLine$(settingsData$, "HCB: ")
 53 | 	pythonex$ = extractLine$(settingsData$, "PY2: ")
 54 | 	thr = extractNumber(settingsData$, "THR: ")
 55 | # If the settings isn't present initialize everything
 56 | else
 57 | 	dictionary$ = ""
 58 | 	ruleset$ = ""
 59 | 	new$ = "phon"
 60 | 	ort$ = ""
 61 | 	wrd$ = ""
 62 | 	can$ = ""
 63 | 	llh$ = ""
 64 | 	lan = 1
 65 | 	phonetizer$ = ""
 66 | 	pho = 0
 67 | 	model = 1
 68 | 	if windows
 69 | 		log$ = "nul"
 70 | 	else
 71 | 		log$ = "/dev/null"
 72 | 	endif
 73 | 	soxex$ = soxexecutablename$
 74 | 	hviteex$ = hviteexecutablename$
 75 | 	hcopyex$ = hcopyexecutablename$
 76 | 	pythonex$ = pythonexecutablename$
 77 | 	thr = 0
 78 | endif
 79 | 
 80 | # Spawn the option window for the user
 81 | beginPause: "Basic options"
 82 | 	comment: "Praatalign version 2.0a"
 83 | 	comment: "Name for the output tier(may already exist)"
 84 | 	sentence: "new", new$
 85 | 
 86 | 	comment: "Name for the output tier used for orthographic word level alignment"
 87 | 	sentence: "ort", ort$
 88 | 	
 89 | 	comment: "Name for the output tier used for word level alignment"
 90 | 	sentence: "wrd", wrd$
 91 | 
 92 | 	comment: "Name for the output tier used for canonical pronunciation"
 93 | 	sentence: "can", can$
 94 | 
 95 | 	comment: "Name for the output tier used for log likelyhood values"
 96 | 	sentence: "llh", llh$
 97 | 
 98 | 	comment: "Select model"
 99 | 	optionMenu: "model", model
100 | 		option: "spanish"
101 | 		option: "dutch"
102 | 
103 | 	comment: "Select phonetizer"
104 | 	optionMenu: "lan", lan
105 | 		option: "spanish"
106 | 		option: "tzeltal"
107 | 		option: "universal"
108 | 		option: "none"
109 | 
110 | 	comment: "Select a dictionary when pressing apply"
111 | 	boolean: "dic", 0
112 | 	if dictionary$ <> ""
113 | 		sentence: "dictionary", dictionary$
114 | 	endif
115 | 
116 | 	comment: "Select a ruleset file when pressing apply"
117 | 	boolean: "rul", 0
118 | 	if ruleset$ <> ""
119 | 		sentence: "ruleset", ruleset$
120 | 	endif
121 | 
122 | 	if phonetizer$ <> ""
123 | 		comment: "Select a phonetizer file when pressing apply"
124 | 		comment: "This always happens if you use the universal phonetizer"
125 | 		boolean: "pho", 0
126 | 		sentence: "phonetizer", phonetizer$
127 | 	endif
128 | 
129 | 	comment: "Set the length of the added length to the annotations"
130 | 	real: "thr", thr
131 | endPause: "Apply", 1
132 | 
133 | beginPause: "Advanced options"
134 | 	comment: "Developer/debug options"
135 | 	sentence: "log", log$
136 | 
137 | 	comment: "Select sox executable location when pressing apply"
138 | 	boolean: "sox", 0
139 | 	if soxex$ <> soxexecutablename$
140 | 		sentence: "soxex", soxex$
141 | 	endif
142 | 
143 | 	comment: "Select HVite executable location when pressing apply"
144 | 	boolean: "hvite", 0
145 | 	if hviteex$ <> hviteexecutablename$
146 | 		sentence: "hviteex", hviteex$
147 | 	endif
148 | 	
149 | 	comment: "Select HCopy executable location when pressing apply"
150 | 	boolean: "hcopy", 0
151 | 	if hcopyex$ <> hcopyexecutablename$
152 | 		sentence: "hcopyex", hcopyex$
153 | 	endif
154 | 
155 | 	comment: "Select python executable location when pressing apply"
156 | 	boolean: "python", 0
157 | 	if pythonex$ <> pythonexecutablename$
158 | 		sentence: "pythonex", pythonex$
159 | 	endif
160 | endPause: "Apply", 1
161 | 
162 | if (phonetizer$ == "" and lan$ == "universal") or pho
163 | 	beginPause: "Instructions"
164 | 		comment: "Please point me to the universal phonetizer file."
165 | 		comment: "The format for such file can be found in the manual"
166 | 	clicked = endPause: "Cancel", "Continue", 2, 1
167 | 	if clicked = 2
168 | 		phonetizer$ = chooseReadFile$("Point me to the phonetizer file")
169 | 	endif
170 | 	if phonetizer$ = ""
171 | 		pause The plugin will probably malfunction since you didn't select a file.
172 | 	endif
173 | endif
174 | if phonetizer$ = ""
175 | 	phonetizer$ = "None"
176 | endif
177 | 
178 | # Ask for the dictionary
179 | if dic
180 | 	beginPause: "Instructions"
181 | 		comment: "Please point me to the dictionary file."
182 | 		comment: "The format for such file can be found in the manual"
183 | 		comment: "Dictionary skeletons can be generated with Generate dictionary from tier..."
184 | 	clicked = endPause: "Cancel", "Continue", 2, 1
185 | 	if clicked = 2
186 | 		dictionary$ = chooseReadFile$("Point me to the dictionary file")
187 | 	endif
188 | endif
189 | if dictionary$ = ""
190 | 	dictionary$ = "None"
191 | endif
192 | 
193 | # Ask for the ruleset
194 | if rul
195 | 	beginPause: "Instructions"
196 | 		comment: "Please point me to the ruleset file."
197 | 		comment: "The format for such file can be found in the manual"
198 | 	clicked = endPause: "Cancel", "Continue", 2, 1
199 | 	if clicked = 2
200 | 		ruleset$ = chooseReadFile$("Point me to the ruleset file")
201 | 	endif
202 | endif
203 | if ruleset$ = ""
204 | 	ruleset$ = "None"
205 | endif
206 | 
207 | pathstring$ = if windows then "%PATH%" else "$PATH" fi
208 | messagesox$ = if windows then "C:\Program Files" else "/usr/bin" fi
209 | messagepy$ = if windows then "C:\Python27\" else "/usr/bin" fi
210 | messagehtk$ = if windows then
211 | ... "where you unzipped the file pointed out in README.html" else
212 | ... "where you downloaded or compiled the executables" fi
213 | 
214 | @getBinary: messagesox$, pathstring$, soxexecutablename$, sox
215 | soxex$ = if getBinary.ex$ <> "" then getBinary.ex$ else soxex$ fi
216 | 
217 | @getBinary: messagehtk$, pathstring$, hviteexecutablename$, hvite
218 | hviteex$ = if getBinary.ex$ <> "" then getBinary.ex$ else hviteex$ fi
219 | 
220 | @getBinary: messagehtk$, pathstring$, hcopyexecutablename$, hcopy
221 | hcopyex$ = if getBinary.ex$ <> "" then getBinary.ex$ else hcopyex$ fi
222 | 
223 | @getBinary: messagepy$, pathstring$, pythonexecutablename$, python
224 | pythonex$ = if getBinary.ex$ <> "" then getBinary.ex$ else pythonex$ fi
225 | 
226 | # Delete the original settings file and write the new one
227 | deleteFile("settings")
228 | writeFileLine("settings",
229 | ..."DCT: ", dictionary$, newline$,
230 | ..."HCB: ", hcopyex$, newline$,
231 | ..."HVB: ", hviteex$, newline$,
232 | ..."LAN: ", lan$, newline$,
233 | ..."MOD: ", model$, newline$,
234 | ..."PHO: ", phonetizer$, newline$,
235 | ..."LOG: ", log$, newline$,
236 | ..."LLH: ", llh$, newline$,
237 | ..."NEW: ", new$, newline$,
238 | ..."CAN: ", can$, newline$,
239 | ..."ORT: ", ort$, newline$,
240 | ..."OUT: ", "praat_temp_out", newline$,
241 | ..."RUL: ", ruleset$, newline$,
242 | ..."SOX: ", soxex$, newline$,
243 | ..."THR: ", thr, newline$,
244 | ..."PY2: ", pythonex$, newline$,
245 | ..."WRD: ", wrd$)
246 | 


--------------------------------------------------------------------------------
/settings_ni.praat:
--------------------------------------------------------------------------------
 1 | # The form for non interactive setup
 2 | form Set the variables
 3 | 	sentence new align
 4 | 	sentence ort align_o
 5 | 	sentence wrd align_w
 6 | 	sentence can align_c
 7 | 	sentence llh align_l
 8 | 	sentence lan
 9 | 	sentence model
10 | 	sentence pho
11 | 	sentence dic
12 | 	sentence rul
13 | 	real thr
14 | 	sentence log /dev/null
15 | 	sentence sox sox
16 | 	sentence hvb hvb
17 | 	sentence hcb hcb
18 | 	sentence py py
19 | endform
20 | 
21 | # Write the settings file
22 | writeFileLine("settings",
23 | ..."DCT: ", dic$, newline$,
24 | ..."HVB: ", hvb$, newline$,
25 | ..."HCB: ", hcb$, newline$,
26 | ..."LAN: ", lan$, newline$,
27 | ..."MOD: ", model$, newline$,
28 | ..."PHO: ", pho$, newline$,
29 | ..."LOG: ", log$, newline$,
30 | ..."LLH: ", llh$, newline$,
31 | ..."NEW: ", new$, newline$,
32 | ..."CAN: ", can$, newline$,
33 | ..."ORT: ", ort$, newline$,
34 | ..."OUT: ", "praat_temp_out", newline$,
35 | ..."RUL: ", rul$, newline$,
36 | ..."SOX: ", sox$, newline$,
37 | ..."THR: ", thr, newline$,
38 | ..."PY2: ", py$, newline$,
39 | ..."WRD: ", wrd$)
40 | 


--------------------------------------------------------------------------------
/setup.praat:
--------------------------------------------------------------------------------
 1 | # Set the write settings to UTF-8 and read settings too
 2 | Text writing preferences: "UTF-8"
 3 | Text reading preferences: "try UTF-8, then ISO Latin-1"
 4 | 
 5 | # Add the menu commands within the TextGrid editor
 6 | Add menu command... "TextGridEditor" "Interval" "Generate dictionary from tier..." "" 0 generatedict.praat
 7 | Add menu command... "TextGridEditor" "Interval" "Clean selection" "" 0 cleaninterval.praat
 8 | Add menu command... "TextGridEditor" "Interval" "Align current interval" "" 0 alignannotation.praat
 9 | Add menu command... "TextGridEditor" "Interval" "Align current tier" "" 0 aligntier.praat
10 | Add menu command... "TextGridEditor" "Interval" "Set up forced alignment..." "" 0 settings.praat
11 | 


--------------------------------------------------------------------------------