pelican-nlp 0.3.7__py3-none-any.whl → 0.3.9__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- pelican_nlp/_version.py +1 -1
- pelican_nlp/praat/eps_conv.praat +193 -0
- pelican_nlp/praat/polytonia.praat +773 -0
- pelican_nlp/praat/prosogram.praat +102 -0
- pelican_nlp/praat/prosomain.praat +3787 -0
- pelican_nlp/praat/prosoplot.praat +1546 -0
- pelican_nlp/praat/segment.praat +1224 -0
- pelican_nlp/praat/setup.praat +1 -0
- pelican_nlp/praat/stylize.praat +2766 -0
- pelican_nlp/praat/util.praat +632 -0
- {pelican_nlp-0.3.7.dist-info → pelican_nlp-0.3.9.dist-info}/METADATA +1 -1
- {pelican_nlp-0.3.7.dist-info → pelican_nlp-0.3.9.dist-info}/RECORD +16 -7
- {pelican_nlp-0.3.7.dist-info → pelican_nlp-0.3.9.dist-info}/WHEEL +0 -0
- {pelican_nlp-0.3.7.dist-info → pelican_nlp-0.3.9.dist-info}/entry_points.txt +0 -0
- {pelican_nlp-0.3.7.dist-info → pelican_nlp-0.3.9.dist-info}/licenses/LICENSE +0 -0
- {pelican_nlp-0.3.7.dist-info → pelican_nlp-0.3.9.dist-info}/top_level.txt +0 -0
|
@@ -0,0 +1,3787 @@
|
|
|
1
|
+
# prosomain.praat -- include file for prosogram
|
|
2
|
+
# This file isn't a stand-alone script; it is included by the main script "prosogram.praat".
|
|
3
|
+
# For documentation, user's guide and tutorial, see http://sites.google.com/site/prosogram/
|
|
4
|
+
# Author: Piet Mertens
|
|
5
|
+
# Requires Praat 6.0.43 or higher.
|
|
6
|
+
# Last modified on 2021-12-17
|
|
7
|
+
|
|
8
|
+
|
|
9
|
+
# For Postscript graphics, modify the next line to match the path to Ghostscript on your computer.
|
|
10
|
+
path_ghostscript$ = "c:\Program Files\gs\gs9.21\bin\gswin64c.exe" ; typical path on Windows-10 64 bit
|
|
11
|
+
# For GIF graphics, modify the next line to match the path to nconvert on your computer.
|
|
12
|
+
path_nconvert$ = "c:\Program Files\XnView\nconvert" ; only used for creating GIF files
|
|
13
|
+
|
|
14
|
+
# Don't change the following lines:
|
|
15
|
+
path_ghostscript$ = replace$ (path_ghostscript$, "\Program Files\", "\Progra~1\", 1)
|
|
16
|
+
path_nconvert$ = replace$ (path_nconvert$, "\Program Files\", "\Progra~1\", 1)
|
|
17
|
+
|
|
18
|
+
# Accepted filename extensions for sound files:
|
|
19
|
+
filename_extensions_sound$ = ":wav:aiff:aifc:nist:flac:sound:s16:mp3:"
|
|
20
|
+
|
|
21
|
+
|
|
22
|
+
# Procedure hierarchy
|
|
23
|
+
# main main procedure, in both modes (script form or batch mode)
|
|
24
|
+
# initialization_main
|
|
25
|
+
# process_form process fields in script form (in script mode, not in batch mode)
|
|
26
|
+
# process_multiple_input_files all processing in all tasks, for 1 or more input speech files
|
|
27
|
+
# task_flags set flags depending on task
|
|
28
|
+
# map_segmentation_type mapping between segmentation name and tier requirements
|
|
29
|
+
# initialization_multiple_files
|
|
30
|
+
# process_one_input_file
|
|
31
|
+
# initialization_per_file
|
|
32
|
+
# construct_filenames rule-based filenames
|
|
33
|
+
# read_annotation_file read annotation TextGrid from disk
|
|
34
|
+
# read_parameter_files read parameters from file
|
|
35
|
+
# calculate_parameters calculate parameters
|
|
36
|
+
# calculate_pitch_int calculate pitch (internal command, not batch command)
|
|
37
|
+
# calculate_intensity_bp_filtered
|
|
38
|
+
# corpus_conversion apply corpus conventions: tiernames, tier content, etc.
|
|
39
|
+
# prepare_plotted_textgrid create textgrid used in prosogram drawing
|
|
40
|
+
# load_intermediate_data_files
|
|
41
|
+
# get_segmentation read or calculate segmentation (phonetic alignment, etc.)
|
|
42
|
+
# read_nuclei_file_interactive_mode read internal TextGrid from earlier run
|
|
43
|
+
# copy_tiers_from_annotation_tg
|
|
44
|
+
# make_segmentation calculate segmentation
|
|
45
|
+
# safe_nuclei check boundaries of nuclei: unvoiced fragments, octave jumps, etc.
|
|
46
|
+
# create_table_of_nuclei create table of nuclei with prosodic features
|
|
47
|
+
# initialize_nucldat initialize some columns in table of nuclei
|
|
48
|
+
# speaker_info_get list all speakers in annotation and add speaker id to nucldat table
|
|
49
|
+
# stylize_nuclei pitch contour stylization
|
|
50
|
+
# speakers_prosodic_parms calculate prosodic paramters of each speaker
|
|
51
|
+
# pitchrange_normalized_pitch
|
|
52
|
+
# prosodic_profile
|
|
53
|
+
# prosodic_profile_new
|
|
54
|
+
# detect_hesitations detect hesitations from phonetic labels, acoustic prosodic features or words
|
|
55
|
+
# store_stylization
|
|
56
|
+
# speaker_autorange
|
|
57
|
+
# gr_start_picturewin
|
|
58
|
+
# gr_write_all_prosograms
|
|
59
|
+
# cleanup_current_file
|
|
60
|
+
# cleanup_global
|
|
61
|
+
# store_features write spreadsheet with prosodic features for syllables
|
|
62
|
+
#
|
|
63
|
+
# Commands in batch mode:
|
|
64
|
+
# prosogram (batch mode) calculate prosogram
|
|
65
|
+
# prosogram_variants variants for polytonia, boundary detection, stress detection, etc.
|
|
66
|
+
# polytonia (batch mode) calculate tonal annotation
|
|
67
|
+
# segmentation (batch mode) calculate automatic segmentation
|
|
68
|
+
# calculate_pitch (batch mode) calculate pitch (two-pass algorithm)
|
|
69
|
+
|
|
70
|
+
version$ = "Prosogram v3.01"
|
|
71
|
+
; @logging: "reset debug timed", "_log.txt"
|
|
72
|
+
|
|
73
|
+
|
|
74
|
+
corpus$ = "" ; "" = use default corpus conventions;
|
|
75
|
+
; Known corpora: rhapsodie, cprom, project_frfc, RUHRCAT, Prsir, gvlex, Kotsifas_2015
|
|
76
|
+
|
|
77
|
+
|
|
78
|
+
include prosoplot.praat
|
|
79
|
+
include segment.praat
|
|
80
|
+
include stylize.praat
|
|
81
|
+
include polytonia.praat
|
|
82
|
+
include util.praat
|
|
83
|
+
|
|
84
|
+
;if (boundary_annotation)
|
|
85
|
+
; include rules.praat
|
|
86
|
+
; include prominence.praat
|
|
87
|
+
;include duration.praat
|
|
88
|
+
;include elasticity.praat
|
|
89
|
+
;endif
|
|
90
|
+
|
|
91
|
+
|
|
92
|
+
|
|
93
|
+
; types of tasks
|
|
94
|
+
task_prosogram = 4 ; stylization, prosodic features, prosodic profile, Polytonia
|
|
95
|
+
task_interactive = 5
|
|
96
|
+
task_calc_pitch = 1
|
|
97
|
+
task_calc_loudness = 10
|
|
98
|
+
task_calc_bp_intensity = 13
|
|
99
|
+
task_pitch_plot = 2
|
|
100
|
+
task_segmentation = 3 ; automatic segmentation into syllabic nuclei without phonetic alignment annotation
|
|
101
|
+
task_annotation = 6 ; Draw annotation only
|
|
102
|
+
task_validate_tiers = 12
|
|
103
|
+
task_autoanno = 9 ; only when using Praat form, not in batch mode
|
|
104
|
+
|
|
105
|
+
; types of segmentation
|
|
106
|
+
segm_vnucl = 1
|
|
107
|
+
segm_extern = 2
|
|
108
|
+
segm_aloudness = 3
|
|
109
|
+
segm_anucl = 4
|
|
110
|
+
segm_asyll = 5
|
|
111
|
+
segm_msyllvow = 6
|
|
112
|
+
segm_msyllpeak = 7
|
|
113
|
+
segm_mrhyme = 8 ; syllable rhyme using phoneme and syllable annotation
|
|
114
|
+
segm_voiced = 9 ; voiced portions
|
|
115
|
+
segm_specsim = 10 ; spectral similarity
|
|
116
|
+
segm_pitchchange = 11
|
|
117
|
+
segm_pitchterrace = 12
|
|
118
|
+
|
|
119
|
+
|
|
120
|
+
if (not variableExists ("font"))
|
|
121
|
+
font_family$ = "Times" ; Font used in plot; else use "Helvetica"
|
|
122
|
+
fontsize = 10
|
|
123
|
+
else
|
|
124
|
+
font_family$ = extractWord$(font$, "")
|
|
125
|
+
fontsize = extractNumber(font$, " ")
|
|
126
|
+
endif
|
|
127
|
+
|
|
128
|
+
if (not variableExists ("batch_mode")) ; prosomain script started directly, not from prosogram.praat
|
|
129
|
+
batch_mode = 1
|
|
130
|
+
# variables initialized before batch functions
|
|
131
|
+
output_filename$ = ""
|
|
132
|
+
clip_level = 0
|
|
133
|
+
scale_signal_amplitude = 0
|
|
134
|
+
minimum_pitch = 0
|
|
135
|
+
maximum_pitch = 450
|
|
136
|
+
volatile = 0
|
|
137
|
+
viewsize$ = "wide"
|
|
138
|
+
@msg: "Batch mode..."
|
|
139
|
+
endif
|
|
140
|
+
|
|
141
|
+
|
|
142
|
+
procedure initialization_main
|
|
143
|
+
draw_prosograms = 1 ; draw_prosograms (modified by task_flags and by prosogram_variants)
|
|
144
|
+
save_intermediate_data = 0 ; save intermediate data (modified by task_flags and by prosogram_variants)
|
|
145
|
+
hesitation_annotation = 0 ; save detected hesitations
|
|
146
|
+
hesitation_method$ = "none" ; "annotation" / "phon+pros" / "words" / "all" / "none"
|
|
147
|
+
pause_method$ = "nuclei" ; "nuclei" / "phon_tier" / "syll_tier"
|
|
148
|
+
polytonia_annotation = 0 ; automatic annotation of pitch movements and pitch levels
|
|
149
|
+
calc_prominence = 0 ; compute prominence measures
|
|
150
|
+
boundary_annotation = 0 ; automatic detection of prosodic boundaries
|
|
151
|
+
stress_annotation = 0 ; automatic detection of stress
|
|
152
|
+
needs_loudness = 0 ; compute loudness based on excitation spectrum
|
|
153
|
+
|
|
154
|
+
; Plotting
|
|
155
|
+
viewport_width = 7.5 ; width of viewport used for prosogram (inches)
|
|
156
|
+
clip_to_Y_range = 0 ; clip stylisation to Y range
|
|
157
|
+
; Plotting options:
|
|
158
|
+
greyscale = 0 ; use greyscale instead of colors
|
|
159
|
+
show_f0 = 0 ; show F0 in prosogram (both in rich and light modes)
|
|
160
|
+
show_pitchrange = 0 ;
|
|
161
|
+
show_pauses = 0 ; show detected pauses in prosogram by label "P"
|
|
162
|
+
show_hesitations = 0 ; show hesitations in prosogram by label "Hes"
|
|
163
|
+
show_prominence = 0 ;
|
|
164
|
+
show_pseudosyllables = 0 ; show pseudosyllables in TextGrid
|
|
165
|
+
show_harmonicity = 0 ;
|
|
166
|
+
show_elasticity = 0 ;
|
|
167
|
+
show_localrate = 0 ;
|
|
168
|
+
show_rhythm = 0 ;
|
|
169
|
+
show_length = 0 ;
|
|
170
|
+
|
|
171
|
+
single_fname_graphics_output = 0; when true, all graphic files will be numbered using the same basename
|
|
172
|
+
mindur_pause_anno = 0.2 ; min duration for pause in annotation
|
|
173
|
+
prefilterHP100 = 0 ; Apply HP filtering to avoid problems with low quality audio recordings
|
|
174
|
+
prefilterLP = 0 ; Apply LP filtering when fricatives are too strong (low quality audio recording)
|
|
175
|
+
avoid_insuff_memory = 0 ; Used in calculation of intensity of BP filtered signal
|
|
176
|
+
use_duration_model = 0 ;
|
|
177
|
+
; duration_model_filename$ = "db_phon_dur_rhapsodie.txt"
|
|
178
|
+
; duration_model_filename$ = "db_phon_dur_cprom.txt"
|
|
179
|
+
; duration_model_filename$ = "db_phon_dur_gvlex.txt"
|
|
180
|
+
duration_model_filename$ = ""
|
|
181
|
+
collect_output = 0 ; collect output in a single output file, for feature files, duration data, etc.
|
|
182
|
+
long_feature_table = 0 ; use long format feature table, along standard format
|
|
183
|
+
rich_format_stylization = 0 ; enable rich format stylization output file (one of the intermediate data files)
|
|
184
|
+
do_cleanup = 1 ; cleanup all loaded or computed data at end of process (disable for debugging)
|
|
185
|
+
save_BP = 0 ; only used in batch_mode segmentation
|
|
186
|
+
globalsheet_available = 0 ; multiple inputfile prosodic profile
|
|
187
|
+
endproc
|
|
188
|
+
|
|
189
|
+
|
|
190
|
+
procedure main
|
|
191
|
+
@debug_msg: "main: entry"
|
|
192
|
+
clearinfo ; clear the info window
|
|
193
|
+
@debug_msg: "main: Praat version = 'praatVersion'"
|
|
194
|
+
if (praatVersion < 6043)
|
|
195
|
+
@fatal_error: "Requires Praat version 6.0.43 or higher. Please update your Praat application."
|
|
196
|
+
endif
|
|
197
|
+
|
|
198
|
+
@initialization_main
|
|
199
|
+
@process_form
|
|
200
|
+
@process_multiple_input_files: task, input_files$, anal_t1, anal_t2
|
|
201
|
+
|
|
202
|
+
@msg: "Ready"
|
|
203
|
+
endproc
|
|
204
|
+
|
|
205
|
+
|
|
206
|
+
procedure process_multiple_input_files: .task, .filespec$, pmf_t1, pmf_t2
|
|
207
|
+
# <pmf_t1> starttime as used in call to multiple file batch processing
|
|
208
|
+
# <pmf_t2> endtime as used in call to multiple file batch processing
|
|
209
|
+
@debug_msg: "process_multiple_input_files: entry, filespec=<'.filespec$'>"
|
|
210
|
+
@msg: "'newline$'Script version 'version$'..."
|
|
211
|
+
task = .task ; needed in batch_mode, because global variable <task> is defined in process_form
|
|
212
|
+
.nfiles = 0
|
|
213
|
+
if (index (.filespec$, "*")) ; wildcard found
|
|
214
|
+
filelistID = Create Strings as file list: "filelist", .filespec$
|
|
215
|
+
.nfiles = Get number of strings
|
|
216
|
+
.wildcard_found = 1
|
|
217
|
+
else
|
|
218
|
+
.wildcard_found = 0
|
|
219
|
+
if (fileReadable (.filespec$))
|
|
220
|
+
.nfiles = 1
|
|
221
|
+
endif
|
|
222
|
+
endif
|
|
223
|
+
if (.nfiles == 0)
|
|
224
|
+
@fatal_error: "No input files found for <'.filespec$'>"
|
|
225
|
+
endif
|
|
226
|
+
; filename returned by 'Get string' does not include path
|
|
227
|
+
@fname_parts: .filespec$
|
|
228
|
+
indir$ = result4$
|
|
229
|
+
@task_flags
|
|
230
|
+
@initialization_multiple_files
|
|
231
|
+
if (task == task_interactive and .nfiles > 1)
|
|
232
|
+
@msg: "Interactive mode uses a single speech file, not many."
|
|
233
|
+
.nfiles = 1
|
|
234
|
+
endif
|
|
235
|
+
for .ifile to .nfiles
|
|
236
|
+
if (not .wildcard_found)
|
|
237
|
+
.filespec$ = replace_regex$ (.filespec$, "\*", "",1) ; remove wildcard '*' (Kleene star)
|
|
238
|
+
@fname_parts: .filespec$
|
|
239
|
+
.fname$ = result1$
|
|
240
|
+
else
|
|
241
|
+
select Strings filelist
|
|
242
|
+
.fname$ = Get string: .ifile
|
|
243
|
+
; filename returned by 'Get string' does not include path
|
|
244
|
+
endif
|
|
245
|
+
@process_one_input_file: pmf_t1, pmf_t2, .ifile, .nfiles, "'indir$''.fname$'"
|
|
246
|
+
endfor
|
|
247
|
+
@cleanup_global
|
|
248
|
+
endproc
|
|
249
|
+
|
|
250
|
+
|
|
251
|
+
procedure process_one_input_file: anal_t1, anal_t2, .ifile, .nfiles, .fullname$
|
|
252
|
+
@debug_msg: "process_one_input_file: entry, save_intermediate_data='save_intermediate_data'"
|
|
253
|
+
@initialization_per_file
|
|
254
|
+
@construct_filenames: .fullname$
|
|
255
|
+
@fname_parts: .fullname$
|
|
256
|
+
.fname$ = result2$
|
|
257
|
+
@msg: "'newline$'Processing input file <'.fname$'>..."
|
|
258
|
+
|
|
259
|
+
if (task == task_prosogram and .ifile == 1)
|
|
260
|
+
; delete global report file created in previous run of Prosogram script
|
|
261
|
+
deleteFile: globalfile$
|
|
262
|
+
endif
|
|
263
|
+
|
|
264
|
+
if (needs_segm_type)
|
|
265
|
+
if (task == task_segmentation and not batch_mode)
|
|
266
|
+
segmentation_name$ = "asyll"
|
|
267
|
+
endif
|
|
268
|
+
if (task = task_interactive) ; peek segmentation_name$
|
|
269
|
+
@peek_settings
|
|
270
|
+
endif
|
|
271
|
+
if ((task = task_prosogram or task = task_interactive) and segmentation_name$ = "optimal")
|
|
272
|
+
@select_optimal_segmentation_method
|
|
273
|
+
@msg: "Selected optimal segmentation method: 'segmentation_name$'"
|
|
274
|
+
endif
|
|
275
|
+
@map_segmentation_type: segmentation_name$
|
|
276
|
+
endif
|
|
277
|
+
|
|
278
|
+
if (needs_segm_tg and not fileReadable (segfile$))
|
|
279
|
+
@error_msg: "Cannot find annotation file <'segfile$'>"
|
|
280
|
+
success = 0
|
|
281
|
+
goto file_done
|
|
282
|
+
endif
|
|
283
|
+
|
|
284
|
+
if (task == task_validate_tiers)
|
|
285
|
+
@validate_syllable_tier: segfile$
|
|
286
|
+
goto file_done
|
|
287
|
+
elsif (task = task_calc_pitch)
|
|
288
|
+
@calculate_pitch_int
|
|
289
|
+
removeObject: pitchID
|
|
290
|
+
pitch_available = 0
|
|
291
|
+
goto file_done
|
|
292
|
+
elsif (task = task_calc_bp_intensity)
|
|
293
|
+
@calculate_intensity_bp_filtered
|
|
294
|
+
removeObject: intbpID
|
|
295
|
+
intbp_available = 0
|
|
296
|
+
goto file_done
|
|
297
|
+
elsif (task = task_calc_loudness)
|
|
298
|
+
@calculate_loudness
|
|
299
|
+
removeObject: loudnessID
|
|
300
|
+
loudness_available = 0
|
|
301
|
+
goto file_done
|
|
302
|
+
endif
|
|
303
|
+
|
|
304
|
+
if (task == task_interactive) ; Read segmentation_name from nucleus file
|
|
305
|
+
@peek_settings
|
|
306
|
+
@gr_start_demowin: basename$
|
|
307
|
+
@gr_printline: "Interactive mode: Processing input file '.fname$'..."
|
|
308
|
+
endif
|
|
309
|
+
|
|
310
|
+
@peek_signal
|
|
311
|
+
anal_t1 = max(anal_t1, signal_start)
|
|
312
|
+
if (anal_t2 == 0)
|
|
313
|
+
anal_t2 = signal_finish
|
|
314
|
+
elsif (anal_t2 > 0)
|
|
315
|
+
anal_t2 = min(anal_t2, signal_finish)
|
|
316
|
+
endif
|
|
317
|
+
if (anal_t1 >= signal_finish)
|
|
318
|
+
@fatal_error: "The start time you supplied ('anal_t1:3') is outside the time range of the sound ('signal_start:3' - 'signal_finish:3')"
|
|
319
|
+
endif
|
|
320
|
+
if (anal_t1 >= anal_t2)
|
|
321
|
+
@fatal_error: "The start time you supplied ('anal_t1:3') is greater than the used end time ('anal_t2:3')"
|
|
322
|
+
endif
|
|
323
|
+
|
|
324
|
+
if (task == task_prosogram or task == task_interactive)
|
|
325
|
+
if (anal_t1 == signal_start and (abs(anal_t2 - signal_finish) < time_step))
|
|
326
|
+
needs_pitchrange = 1 ; will calculate pitchrange (for prosodic profile or for interactive mode)
|
|
327
|
+
; needs_pitchrange is also set for prosodic profile in task_flags
|
|
328
|
+
; if (needs_pitchrange = 1) then full signal is analyzed
|
|
329
|
+
endif
|
|
330
|
+
endif
|
|
331
|
+
|
|
332
|
+
.minutes = (anal_t2 - anal_t1) div 60
|
|
333
|
+
.sec = (anal_t2 - anal_t1) mod 60
|
|
334
|
+
@msg: "Analysis time range: 'anal_t1:3' - 'anal_t2:3' s ('.minutes' min, '.sec:3' s)"
|
|
335
|
+
|
|
336
|
+
if (needs_parameters)
|
|
337
|
+
if (not volatile)
|
|
338
|
+
@gr_printline: "Reading parameter files..."
|
|
339
|
+
@read_parameter_files
|
|
340
|
+
endif
|
|
341
|
+
@gr_printline: "Calculating parameters..."
|
|
342
|
+
@calculate_parameters
|
|
343
|
+
endif
|
|
344
|
+
|
|
345
|
+
if (task != task_interactive and signal_available) ; Free memory if possible
|
|
346
|
+
removeObject: soundID
|
|
347
|
+
signal_available = 0
|
|
348
|
+
endif
|
|
349
|
+
|
|
350
|
+
if (needs_segm_tg or draw_prosograms or needs_pitchrange)
|
|
351
|
+
@read_annotation_file
|
|
352
|
+
endif
|
|
353
|
+
if (segfile_available) ; apply conversion to segmentation textgrid before making textgrid for plotting
|
|
354
|
+
selectObject: segmentationID
|
|
355
|
+
Convert to backslash trigraphs
|
|
356
|
+
@corpus_conversion: 1, "segmentationID", corpus$
|
|
357
|
+
endif
|
|
358
|
+
|
|
359
|
+
if (draw_prosograms)
|
|
360
|
+
@gr_printline: "Preparing TextGrid plotted in prosograms..."
|
|
361
|
+
@prepare_plotted_textgrid: tiers_to_show$
|
|
362
|
+
@gr_printline: "Preparing TextGrid plotted in prosograms... Ready"
|
|
363
|
+
endif
|
|
364
|
+
|
|
365
|
+
if (needs_segm_tg) ; apply conversion to segmentation textgrid after making textgrid for plotting
|
|
366
|
+
@corpus_conversion: 2, "segmentationID", corpus$
|
|
367
|
+
endif
|
|
368
|
+
|
|
369
|
+
if (needs_stylization or task == task_segmentation)
|
|
370
|
+
if (needs_pitchrange) ; analyse entire signal
|
|
371
|
+
t1s = signal_start
|
|
372
|
+
t2s = signal_finish
|
|
373
|
+
else ; analyse only part of the speech signal
|
|
374
|
+
t1s = anal_t1
|
|
375
|
+
t2s = anal_t2
|
|
376
|
+
endif
|
|
377
|
+
if (not task == task_interactive)
|
|
378
|
+
@gr_printline: "Segmentation into syllabic nuclei. Method='segmentation_name$', Time range='t1s:3'-'t2s:3'"
|
|
379
|
+
endif
|
|
380
|
+
@get_segmentation: segm_type
|
|
381
|
+
; also does @safe_nuclei
|
|
382
|
+
endif
|
|
383
|
+
|
|
384
|
+
if (task == task_segmentation) ; post-processing segmentation results and write to file
|
|
385
|
+
selectObject: nucleiID
|
|
386
|
+
.tmp1 = Extract tier: nucleus_tier
|
|
387
|
+
.tmpN = Into TextGrid
|
|
388
|
+
selectObject: nucleiID
|
|
389
|
+
.tmp2 = Extract tier: syllable_tier
|
|
390
|
+
.tmpS = Into TextGrid
|
|
391
|
+
selectObject: nucleiID
|
|
392
|
+
.tmp3 = Extract tier: dip_tier
|
|
393
|
+
.tmpD = Into TextGrid
|
|
394
|
+
removeObject: .tmp1, .tmp2, .tmp3
|
|
395
|
+
if (segm_type == segm_pitchchange or segm_type == segm_pitchterrace) ; only keep syllable tier and name it "segm"
|
|
396
|
+
selectObject: .tmpS, .tmpD
|
|
397
|
+
.tmp = Merge
|
|
398
|
+
Set tier name: 1, "segm"
|
|
399
|
+
Set tier name: 2, "state"
|
|
400
|
+
else ; segm_asyll, segm_specsim ... ; keep nucleus tier and syllable tier
|
|
401
|
+
selectObject: .tmpN, .tmpS
|
|
402
|
+
.tmp = Merge
|
|
403
|
+
@nucleus_tier_postproc: .tmp, 1, 2
|
|
404
|
+
if (batch_mode)
|
|
405
|
+
if (trimleft > 0 or trimright > 0)
|
|
406
|
+
@tier_trim_nucleus: .tmp, 1, trimleft, trimright
|
|
407
|
+
endif
|
|
408
|
+
endif
|
|
409
|
+
endif
|
|
410
|
+
selectObject: .tmp
|
|
411
|
+
Write to text file: autosegfile$
|
|
412
|
+
@msg: "Segmentation written to <'autosegfile$'>"
|
|
413
|
+
removeObject: .tmpN, .tmpS, .tmpD, .tmp
|
|
414
|
+
if (batch_mode and save_BP)
|
|
415
|
+
selectObject: intbpID
|
|
416
|
+
Write to text file: intbpfile$
|
|
417
|
+
@msg: "Intensity of BP-filtered signal written to <'intbpfile$'>"
|
|
418
|
+
endif
|
|
419
|
+
goto file_done
|
|
420
|
+
endif
|
|
421
|
+
|
|
422
|
+
if (needs_stylization)
|
|
423
|
+
if (nrof_nuclei_analysed < 1)
|
|
424
|
+
@error_msg: "No syllabic nuclei found in speech signal."
|
|
425
|
+
@error_msg: "If you are using segmentation from annotation tier (phon... or syll...), check tier name and tier content."
|
|
426
|
+
if (save_intermediate_data)
|
|
427
|
+
selectObject: nucleiID
|
|
428
|
+
Write to text file: nuclfile$
|
|
429
|
+
endif
|
|
430
|
+
success = 0
|
|
431
|
+
goto file_done
|
|
432
|
+
endif
|
|
433
|
+
@create_table_of_nuclei
|
|
434
|
+
endif
|
|
435
|
+
|
|
436
|
+
if (task == task_interactive and reuse_nucl)
|
|
437
|
+
; stylization cannot be used by boundary_annotation, because procedure stylize computes data in nucldatID
|
|
438
|
+
@load_intermediate_data_files
|
|
439
|
+
endif
|
|
440
|
+
|
|
441
|
+
if (needs_stylization)
|
|
442
|
+
@initialize_nucldat: t1s, t2s
|
|
443
|
+
|
|
444
|
+
if (task <> task_interactive and hesitation_method$ <> "none")
|
|
445
|
+
; hesitations are used in duration data, in Polytonia, in detection of boundaries and prominence
|
|
446
|
+
@msg: "Detecting hesitations (method='hesitation_method$')..."
|
|
447
|
+
@detect_hesitations: anal_t1, anal_t2, hesitation_method$
|
|
448
|
+
if (hesitation_annotation)
|
|
449
|
+
selectObject: nucleiID
|
|
450
|
+
.tmpID = Extract one tier: hesitation_tier
|
|
451
|
+
Write to text file: hesitfile$
|
|
452
|
+
@msg: "Hesitation annotation saved to file... <'hesitfile$'>"
|
|
453
|
+
removeObject: .tmpID
|
|
454
|
+
endif
|
|
455
|
+
endif
|
|
456
|
+
|
|
457
|
+
@debug_msg: "process_one_input_file: needs_stylization"
|
|
458
|
+
if (not stylization_available)
|
|
459
|
+
# Create pitch tier object for stylization
|
|
460
|
+
if (adaptive_glissando)
|
|
461
|
+
s$ = "G(adapt)='glissando_low'-'glissando'/T^2"
|
|
462
|
+
else
|
|
463
|
+
s$ = "G='glissando'/T^2"
|
|
464
|
+
endif
|
|
465
|
+
s$ = "'segmentation_name$', " + s$ + ", DG='diffgt', dmin='mindur_ts:3'"
|
|
466
|
+
@gr_printline: "Calculating stylization... ('s$')"
|
|
467
|
+
stylID = Create PitchTier: "stylization", signal_start, signal_finish
|
|
468
|
+
@stylize_nuclei: t1s, t2s
|
|
469
|
+
stylization_available = 1
|
|
470
|
+
@gr_printline: "Calculating stylization... Ready"
|
|
471
|
+
endif
|
|
472
|
+
; next line comes after stylize_nuclei, because it uses values in nucldatID
|
|
473
|
+
|
|
474
|
+
; @debug_msg: "process_one_input_file: profile_available='profile_available' needs_pitchrange='needs_pitchrange'"
|
|
475
|
+
if (not profile_available) ; in interactive mode profile may have been read from file
|
|
476
|
+
@speaker_info_get
|
|
477
|
+
@profile_table_create
|
|
478
|
+
@profile_table_prepare
|
|
479
|
+
profile_available = 1
|
|
480
|
+
; Table is available, but needs to be filled by @speakers_prosodic_parms
|
|
481
|
+
if (needs_pitchrange)
|
|
482
|
+
;if (variableExists ("corpus$") and corpus$ = "Proust" and variableExists ("corpus_speaker_range_1$"))
|
|
483
|
+
; @msg: "Reading pitch range from corpus_speaker_range..."
|
|
484
|
+
; speaker_range_1$ = corpus_speaker_range_1$
|
|
485
|
+
@gr_printline: "Calculating pitch range of each speaker..."
|
|
486
|
+
@speakers_prosodic_parms
|
|
487
|
+
; values stored in prosodic profile table
|
|
488
|
+
if (task <> task_interactive)
|
|
489
|
+
@msg: "Calculating pitch range normalized pitch..."
|
|
490
|
+
@pitchrange_normalized_pitch
|
|
491
|
+
endif
|
|
492
|
+
endif
|
|
493
|
+
endif
|
|
494
|
+
|
|
495
|
+
if (needs_prosodic_profile and (task <> task_interactive) and (not polytonia_annotation))
|
|
496
|
+
; Write prosodic profile report file
|
|
497
|
+
@prosodic_profile
|
|
498
|
+
@msg: "Writing prosodic profile of current input file to: 'profile_file$'"
|
|
499
|
+
selectObject: profileID
|
|
500
|
+
Save as headerless spreadsheet file: profile_file$
|
|
501
|
+
@prosodic_profile_new
|
|
502
|
+
@msg: "Writing global prosodic profile of all input files to: 'globalfile$'"
|
|
503
|
+
@update_global_report
|
|
504
|
+
endif
|
|
505
|
+
|
|
506
|
+
if (calc_prominence)
|
|
507
|
+
@msg: "Calculating prominence..."
|
|
508
|
+
@calc_prom_dynwidth: nucldatID, j_nucldur, j_promL2D_nucldur, 2, 0, 1
|
|
509
|
+
; @calculate_prominence_measures
|
|
510
|
+
endif
|
|
511
|
+
if (show_pseudosyllables and draw_prosograms)
|
|
512
|
+
@grid_append_tier: nucleiID, syllable_tier, "newgridID"
|
|
513
|
+
endif
|
|
514
|
+
|
|
515
|
+
if (polytonia_annotation)
|
|
516
|
+
@msg: "Tonal annotation..."
|
|
517
|
+
@polytonia_main: anal_t1, anal_t2
|
|
518
|
+
if (draw_prosograms)
|
|
519
|
+
; In prosograms, only the extended Polytonia is displayed
|
|
520
|
+
@grid_append_tier: nucleiID, polytonia2_tier, "newgridID"
|
|
521
|
+
if (not segfile_available) ; no other tiers available
|
|
522
|
+
selectObject: newgridID
|
|
523
|
+
Remove tier: 1 ; dummy tier
|
|
524
|
+
endif
|
|
525
|
+
endif
|
|
526
|
+
selectObject: nucleiID
|
|
527
|
+
.tmp1 = Extract one tier: polytonia_tier
|
|
528
|
+
selectObject: nucleiID
|
|
529
|
+
.tmp2 = Extract one tier: polytonia2_tier
|
|
530
|
+
selectObject: .tmp1, .tmp2
|
|
531
|
+
.tmp3 = Merge
|
|
532
|
+
Write to text file: polytonia_file$
|
|
533
|
+
@msg: "Tonal annotation saved to file... 'polytonia_file$'"
|
|
534
|
+
removeObject: .tmp1, .tmp2, .tmp3
|
|
535
|
+
endif
|
|
536
|
+
|
|
537
|
+
if (boundary_annotation)
|
|
538
|
+
@msg: "Boundary annotation..."
|
|
539
|
+
show_prominence = 0 ; plot prominence measures
|
|
540
|
+
boundary_annotation_verbose = 1
|
|
541
|
+
save_intermediate_data = 1
|
|
542
|
+
needs_prosodic_profile = 0
|
|
543
|
+
if (not (segm_type == segm_msyllvow or segm_type == segm_mrhyme or segm_type == segm_msyllpeak))
|
|
544
|
+
@error_msg: "Boundary annotation requires Syllabic segmentation method"
|
|
545
|
+
segm_type = segm_msyllvow
|
|
546
|
+
endif
|
|
547
|
+
if (pause_method$ == "syll_tier")
|
|
548
|
+
boundary_use_nuclei = 0 ; use pauses from annotation
|
|
549
|
+
elsif (pause_method$ == "nuclei"
|
|
550
|
+
boundary_use_nuclei = 1 ; use gaps between nuclei
|
|
551
|
+
endif
|
|
552
|
+
@msg: "Calculating boundaries (boundary_use_nuclei='boundary_use_nuclei')..."
|
|
553
|
+
boundary_skip_hesit = 0
|
|
554
|
+
@boundary_analysis: anal_t1, anal_t2, boundary_use_nuclei, boundary_skip_hesit
|
|
555
|
+
selectObject: nucleiID
|
|
556
|
+
tmpID = Extract one tier: boundary_tier
|
|
557
|
+
Write to text file: boundaryfile$
|
|
558
|
+
if (draw_prosograms)
|
|
559
|
+
@grid_append_tier: nucleiID, boundary_tier, "newgridID"
|
|
560
|
+
@grid_append_tier: nucleiID, boundary_tier, "segmentationID"
|
|
561
|
+
endif
|
|
562
|
+
@boundary_pass2: anal_t1, anal_t2
|
|
563
|
+
if (draw_prosograms)
|
|
564
|
+
@grid_append_tier: nucleiID, boundary2_tier, "newgridID"
|
|
565
|
+
endif
|
|
566
|
+
call tier_get segmentationID "^boundary-manu$" ref_tier "Cannot find boundary reference tier in segmentation grid" 0
|
|
567
|
+
call create_evaluation_data segmentationID ref_tier 'evalfile$'
|
|
568
|
+
@msg: "Boundaries ready"
|
|
569
|
+
endif
|
|
570
|
+
|
|
571
|
+
if (stress_annotation)
|
|
572
|
+
@stress_analysis: anal_t1, anal_t2, pause_use_nuclei
|
|
573
|
+
@grid_append_tier: nucleiID, stress_tier, "newgridID"
|
|
574
|
+
endif
|
|
575
|
+
|
|
576
|
+
if (save_intermediate_data)
|
|
577
|
+
if (not reuse_nucl)
|
|
578
|
+
selectObject: nucleiID
|
|
579
|
+
Write to text file: nuclfile$
|
|
580
|
+
endif
|
|
581
|
+
if (not reuse_styl)
|
|
582
|
+
selectObject: stylID
|
|
583
|
+
Write to text file: stylfile$
|
|
584
|
+
endif
|
|
585
|
+
if (rich_format_stylization)
|
|
586
|
+
@store_stylization: stylID, nucleiID, targetsfile$
|
|
587
|
+
endif
|
|
588
|
+
; call store_stylization_resampled stylID nucleiID time_step 'stylpitchfile$'
|
|
589
|
+
;.outfname$ = prefix$ + "_dur.txt" ; uses basename
|
|
590
|
+
;.outfname$ = indir$ + "_dur_syllprom.txt" ; uses single output file
|
|
591
|
+
;call duration_data_from_textgrid segmentationID .ifile "syllprom" "'.outfname$'" 'basename$'
|
|
592
|
+
endif
|
|
593
|
+
|
|
594
|
+
selectObject: stylID ; stylization pitch tier in Hz
|
|
595
|
+
stylSTID = Copy: "styl_ST"
|
|
596
|
+
@convert_Hz_ST: stylSTID
|
|
597
|
+
|
|
598
|
+
# Initialization for plotting results of current input file
|
|
599
|
+
if (draw_prosograms and auto_pitchrange)
|
|
600
|
+
; GLOBAL automatic pitch range selection for entire corpus to be analysed
|
|
601
|
+
@speaker_autorange: anal_t1, anal_t2
|
|
602
|
+
ySTmax = ymax
|
|
603
|
+
ySTmin = ymin
|
|
604
|
+
endif
|
|
605
|
+
else ; no stylization needed
|
|
606
|
+
if (pitch_available)
|
|
607
|
+
selectObject: pitchID
|
|
608
|
+
.y = Get mean: 0, 0, "semitones re 1 Hz"
|
|
609
|
+
ySTmin = .y - 12
|
|
610
|
+
ySTmax = .y + 12
|
|
611
|
+
endif
|
|
612
|
+
endif ; needs_stylization
|
|
613
|
+
|
|
614
|
+
# Draw all prosograms for current input file
|
|
615
|
+
if (draw_prosograms)
|
|
616
|
+
if (task == task_interactive)
|
|
617
|
+
@gr_start_demowin: basename$
|
|
618
|
+
@gr_run_demowin: anal_t1, anal_t2, timeincr, ySTmin, ySTmax
|
|
619
|
+
@cleanup_current_file
|
|
620
|
+
@cleanup_global
|
|
621
|
+
exit
|
|
622
|
+
else
|
|
623
|
+
@msg: "Drawing and saving all prosograms..."
|
|
624
|
+
@gr_start_picturewin
|
|
625
|
+
@gr_write_all_prosograms: anal_t1, anal_t2, timeincr, .ifile, .nfiles
|
|
626
|
+
endif
|
|
627
|
+
endif ; draw_prosograms
|
|
628
|
+
|
|
629
|
+
if (task == task_annotation)
|
|
630
|
+
@gr_start_picturewin
|
|
631
|
+
if (nrof_pages == 0)
|
|
632
|
+
@gr_first_viewport_of_page
|
|
633
|
+
nrof_pages = 1
|
|
634
|
+
@gr_write_all_annotation: anal_t1, anal_t2, timeincr, .ifile, .nfiles
|
|
635
|
+
endif
|
|
636
|
+
endif
|
|
637
|
+
|
|
638
|
+
if (save_intermediate_data) ; only for some tasks
|
|
639
|
+
; this step must follow plotting, because store_features removes columns
|
|
640
|
+
if (not collect_output)
|
|
641
|
+
@store_features: 1, 0, sheetfile$
|
|
642
|
+
if (long_feature_table)
|
|
643
|
+
@store_features: 1, 1, tablefile$
|
|
644
|
+
endif
|
|
645
|
+
else
|
|
646
|
+
@store_features: .ifile, 0, collectfile$
|
|
647
|
+
endif
|
|
648
|
+
endif
|
|
649
|
+
|
|
650
|
+
label file_done
|
|
651
|
+
|
|
652
|
+
# Delete temporary objects for current input file
|
|
653
|
+
@cleanup_current_file
|
|
654
|
+
endproc
|
|
655
|
+
|
|
656
|
+
|
|
657
|
+
procedure select_optimal_segmentation_method
|
|
658
|
+
if (not fileReadable (segfile$))
|
|
659
|
+
segmentation_name$ = "asyll"
|
|
660
|
+
else
|
|
661
|
+
@debug_msg: "select_optimal_segmentation_method: task='task', segmentation file found"
|
|
662
|
+
.tmpID = Read from file: segfile$
|
|
663
|
+
@tier_number_by_name: .tmpID, "^phon"
|
|
664
|
+
.phon_tier = result
|
|
665
|
+
@tier_number_by_name: .tmpID, "^syll"
|
|
666
|
+
.syll_tier = result
|
|
667
|
+
removeObject: .tmpID
|
|
668
|
+
@debug_msg: "select_optimal_segmentation_method: phon_tier='.phon_tier', syll_tier='.syll_tier'"
|
|
669
|
+
if (.phon_tier and .syll_tier)
|
|
670
|
+
segmentation_name$ = "rhyme"
|
|
671
|
+
elsif (.phon_tier)
|
|
672
|
+
segmentation_name$ = "vow-nucl"
|
|
673
|
+
elsif (.syll_tier)
|
|
674
|
+
segmentation_name$ = "syll"
|
|
675
|
+
else
|
|
676
|
+
segmentation_name$ = "asyll"
|
|
677
|
+
endif
|
|
678
|
+
endif
|
|
679
|
+
endproc
|
|
680
|
+
|
|
681
|
+
|
|
682
|
+
procedure map_segmentation_type: .method$
|
|
683
|
+
# Maps segmentation name to a type and sets some flags.
|
|
684
|
+
@debug_msg: "map_segmentation_type: entry, method='.method$'"
|
|
685
|
+
needs_segm_tg = 1 ; selected segmentation type needs segmentation TextGrid
|
|
686
|
+
needs_phon_tier = 0 ; selected segmentation type needs phoneme tier
|
|
687
|
+
needs_syll_tier = 0 ; selected segmentation type needs syllable tier
|
|
688
|
+
segmentation_name$ = .method$
|
|
689
|
+
if (.method$ = "vow-nucl")
|
|
690
|
+
segm_type = segm_vnucl
|
|
691
|
+
needs_phon_tier = 1
|
|
692
|
+
elsif (.method$ = "extern")
|
|
693
|
+
segm_type = segm_extern
|
|
694
|
+
elsif (.method$ = "loudness") ; automatic, loudness peaks
|
|
695
|
+
segm_type = segm_aloudness
|
|
696
|
+
needs_segm_tg = 0
|
|
697
|
+
elsif (.method$ = "int-BP") ; automatic, peaks in bandpass filters speech
|
|
698
|
+
segm_type = segm_anucl
|
|
699
|
+
needs_intbp = 1
|
|
700
|
+
needs_segm_tg = 0
|
|
701
|
+
elsif (.method$ = "asyll")
|
|
702
|
+
segm_type = segm_asyll
|
|
703
|
+
needs_intbp = 1
|
|
704
|
+
needs_segm_tg = 0
|
|
705
|
+
elsif (.method$ = "rhyme")
|
|
706
|
+
segm_type = segm_mrhyme
|
|
707
|
+
needs_phon_tier = 1
|
|
708
|
+
needs_syll_tier = 1
|
|
709
|
+
elsif (.method$ = "syll+vow")
|
|
710
|
+
segm_type = segm_msyllvow
|
|
711
|
+
needs_phon_tier = 1
|
|
712
|
+
needs_syll_tier = 1
|
|
713
|
+
elsif (.method$ = "syll")
|
|
714
|
+
segm_type = segm_msyllpeak
|
|
715
|
+
needs_syll_tier = 1
|
|
716
|
+
elsif (.method$ = "voiced")
|
|
717
|
+
segm_type = segm_voiced
|
|
718
|
+
needs_segm_tg = 0
|
|
719
|
+
elsif (.method$ = "specsim")
|
|
720
|
+
segm_type = segm_specsim
|
|
721
|
+
needs_segm_tg = 0
|
|
722
|
+
elsif (.method$ = "pitchchange")
|
|
723
|
+
segm_type = segm_pitchchange
|
|
724
|
+
needs_segm_tg = 0
|
|
725
|
+
needs_intbp = 1
|
|
726
|
+
elsif (.method$ = "pitchterrace")
|
|
727
|
+
segm_type = segm_pitchterrace
|
|
728
|
+
needs_segm_tg = 0
|
|
729
|
+
else
|
|
730
|
+
segmentation_name$ = "unknown"
|
|
731
|
+
@fatal_error: "Unknown segmentation type: '.method$'"
|
|
732
|
+
endif
|
|
733
|
+
if (boundary_annotation)
|
|
734
|
+
@debug_msg: "map_segmentation_type: boundary_annotation, segmentation name='segmentation_name$' ('segm_type')"
|
|
735
|
+
if (not (segm_type == segm_msyllvow or segm_type == segm_mrhyme or segm_type == segm_msyllpeak))
|
|
736
|
+
@error_msg: "Boundary annotation requires Syllabic segmentation method"
|
|
737
|
+
segm_type = segm_msyllvow
|
|
738
|
+
endif
|
|
739
|
+
endif
|
|
740
|
+
@debug_msg: "map_segmentation_type: exit"
|
|
741
|
+
endproc
|
|
742
|
+
|
|
743
|
+
|
|
744
|
+
procedure task_flags
|
|
745
|
+
# set some flags depending on task type
|
|
746
|
+
@debug_msg: "task_flags: entry"
|
|
747
|
+
needs_parameters = 1
|
|
748
|
+
needs_stylization = 1
|
|
749
|
+
needs_loudness = 0
|
|
750
|
+
needs_intbp = 0
|
|
751
|
+
needs_pitchrange = 0
|
|
752
|
+
needs_prosodic_profile = 0
|
|
753
|
+
needs_segm_tg = 0
|
|
754
|
+
needs_segm_type = 0
|
|
755
|
+
needs_picture_win = 0
|
|
756
|
+
nrofplottedtiers = 0
|
|
757
|
+
show_settings = 1
|
|
758
|
+
show_portee = 1
|
|
759
|
+
show_tg_bound = 1 ; plot vertical lines for interval boundaries in TextGrid
|
|
760
|
+
show_y_scale = 1 ; plot calibration in ST on vertical axis
|
|
761
|
+
show_y_scale_r = 1 ; plot calibration in Hz on vertical axis at right side
|
|
762
|
+
show_x_scale = 1
|
|
763
|
+
show_vuv = 1 ; plot V/UV parameter
|
|
764
|
+
show_intensity = 1 ; plot intensity parameter
|
|
765
|
+
show_intbp = 1 ; plot int-BP parameter
|
|
766
|
+
show_tiernames = 1 ; plot tier names to the right of tier
|
|
767
|
+
show_trajectories = 0 ; show intrasyllab up/down and intersyllab pitch intervals
|
|
768
|
+
if (task == task_calc_pitch or task == task_calc_bp_intensity or task == task_calc_loudness)
|
|
769
|
+
needs_parameters = 0
|
|
770
|
+
volatile = 0
|
|
771
|
+
needs_stylization = 0
|
|
772
|
+
draw_prosograms = 0
|
|
773
|
+
if (task == task_calc_loudness)
|
|
774
|
+
needs_loudness = 1
|
|
775
|
+
endif
|
|
776
|
+
elsif (task == task_annotation)
|
|
777
|
+
needs_parameters = 0
|
|
778
|
+
volatile = 0
|
|
779
|
+
needs_stylization = 0
|
|
780
|
+
needs_picture_win = 1
|
|
781
|
+
draw_prosograms = 0
|
|
782
|
+
show_y_scale = 0
|
|
783
|
+
elsif (task == task_prosogram)
|
|
784
|
+
needs_segm_type = 1 ; may be modified later by procedure: map_segmentation_type
|
|
785
|
+
if (calc_prominence)
|
|
786
|
+
; needs_loudness = 1
|
|
787
|
+
endif
|
|
788
|
+
needs_picture_win = 1
|
|
789
|
+
if (viewsize$ = "compact")
|
|
790
|
+
show_tiernames = 0
|
|
791
|
+
endif
|
|
792
|
+
if (save_intermediate_data)
|
|
793
|
+
needs_prosodic_profile = 1
|
|
794
|
+
needs_pitchrange = 1
|
|
795
|
+
endif
|
|
796
|
+
if (polytonia_annotation)
|
|
797
|
+
needs_prosodic_profile = 1 ; because pitch range information is stored in profileID !!
|
|
798
|
+
needs_pitchrange = 1
|
|
799
|
+
hesitation_method$ = "all"
|
|
800
|
+
; mindur_ts = 0.05
|
|
801
|
+
show_hesitations = 1
|
|
802
|
+
show_pauses = 1
|
|
803
|
+
save_intermediate_data = 1
|
|
804
|
+
endif
|
|
805
|
+
if (boundary_annotation)
|
|
806
|
+
calc_prominence = 1
|
|
807
|
+
show_prominence = 0
|
|
808
|
+
show_length = 1
|
|
809
|
+
boundary_annotation_verbose = 1
|
|
810
|
+
save_intermediate_data = 1
|
|
811
|
+
endif
|
|
812
|
+
if (stress_annotation)
|
|
813
|
+
calc_prominence = 1
|
|
814
|
+
show_prominence = 1 ; plot prominence measures
|
|
815
|
+
pause_use_nuclei = 1 ; 1= use gaps between nuclei; 0= use pauses from annotation
|
|
816
|
+
endif
|
|
817
|
+
if (hesitation_annotation)
|
|
818
|
+
needs_segm_tg = 1
|
|
819
|
+
needs_pitchrange = 0
|
|
820
|
+
needs_picture_win = 0
|
|
821
|
+
endif
|
|
822
|
+
elsif (task == task_pitch_plot)
|
|
823
|
+
needs_stylization = 0
|
|
824
|
+
segmentation_name$ = ""
|
|
825
|
+
needs_picture_win = 1
|
|
826
|
+
show_settings = 0
|
|
827
|
+
show_vuv = 0
|
|
828
|
+
; show_intensity = 0
|
|
829
|
+
show_intbp = 0
|
|
830
|
+
rich = 1 ; because pitch is plotted in rich mode
|
|
831
|
+
show_pauses = 0
|
|
832
|
+
nrof_nuclei_analysed = 0
|
|
833
|
+
elsif (task == task_interactive)
|
|
834
|
+
needs_segm_type = 1
|
|
835
|
+
show_settings = 0
|
|
836
|
+
viewsize$ = "wide"
|
|
837
|
+
rich = 0
|
|
838
|
+
elsif (task == task_segmentation)
|
|
839
|
+
needs_segm_type = 1
|
|
840
|
+
needs_stylization = 0
|
|
841
|
+
draw_prosograms = 0
|
|
842
|
+
if (not variableExists ("batch_mode"))
|
|
843
|
+
segm_type = segm_asyll
|
|
844
|
+
endif
|
|
845
|
+
elsif (task == task_validate_tiers)
|
|
846
|
+
needs_segm_tg = 1
|
|
847
|
+
needs_parameters = 0
|
|
848
|
+
needs_stylization = 0
|
|
849
|
+
draw_prosograms = 0
|
|
850
|
+
endif
|
|
851
|
+
if (variableExists("rich"))
|
|
852
|
+
if (not rich)
|
|
853
|
+
show_tg_bound = 0
|
|
854
|
+
endif
|
|
855
|
+
endif
|
|
856
|
+
endproc
|
|
857
|
+
|
|
858
|
+
|
|
859
|
+
procedure process_form
|
|
860
|
+
; Interpret all fields from script form and set global flag variables accordingly
|
|
861
|
+
@debug_msg: "process_form: entry"
|
|
862
|
+
needs_segm_type = 0 ; true if task requires a segmentation type
|
|
863
|
+
|
|
864
|
+
; Task selection
|
|
865
|
+
if (index (task$, "Recalculate pitch"))
|
|
866
|
+
task = task_calc_pitch
|
|
867
|
+
elsif (index (task$, "Recalculate intensity"))
|
|
868
|
+
task = task_calc_bp_intensity
|
|
869
|
+
elsif (index (task$, "Recalculate loudness"))
|
|
870
|
+
task = task_calc_loudness
|
|
871
|
+
elsif (index (task$, "Prosogram") or index (task$, "intermediate data files") or index (task$, "Polytonia"))
|
|
872
|
+
task = task_prosogram
|
|
873
|
+
needs_segm_type = 1 ; may be modified later by procedure: map_segmentation_type
|
|
874
|
+
draw_prosograms = 0
|
|
875
|
+
save_intermediate_data = 0
|
|
876
|
+
if (index (task$, "Prosogram"))
|
|
877
|
+
draw_prosograms = 1
|
|
878
|
+
endif
|
|
879
|
+
if (index (task$, "intermediate data files"))
|
|
880
|
+
save_intermediate_data = 1
|
|
881
|
+
endif
|
|
882
|
+
if (index (task$, "Polytonia"))
|
|
883
|
+
polytonia_annotation = 1
|
|
884
|
+
draw_prosograms = 1
|
|
885
|
+
endif
|
|
886
|
+
elsif (index (task$, "Plot pitch"))
|
|
887
|
+
task = task_pitch_plot
|
|
888
|
+
elsif (index (task$, "automatic segmentation"))
|
|
889
|
+
task = task_segmentation
|
|
890
|
+
segm_type = segm_asyll
|
|
891
|
+
segmentation_name$ = "asyll"
|
|
892
|
+
needs_segm_type = 1
|
|
893
|
+
elsif (index (task$, "Interactive"))
|
|
894
|
+
task = task_interactive
|
|
895
|
+
needs_segm_type = 1
|
|
896
|
+
elsif (index (task$, "Draw annotation"))
|
|
897
|
+
task = task_annotation
|
|
898
|
+
elsif (index (task$, "boundary"))
|
|
899
|
+
task = task_autoanno
|
|
900
|
+
boundary_annotation = 1
|
|
901
|
+
needs_segm_type = 1
|
|
902
|
+
elsif (index (task$, "Validate"))
|
|
903
|
+
task = task_validate_tiers
|
|
904
|
+
else
|
|
905
|
+
@fatal_error: "Invalid task"
|
|
906
|
+
endif
|
|
907
|
+
|
|
908
|
+
; Map segmentation type descriptions in form to names that will also be displayed in prosogram graph.
|
|
909
|
+
if (needs_segm_type)
|
|
910
|
+
.method$ = segmentation_method$
|
|
911
|
+
if (index (.method$, "optimal"))
|
|
912
|
+
segmentation_name$ = "optimal"
|
|
913
|
+
elsif (index (.method$, "in vowels"))
|
|
914
|
+
segmentation_name$ = "vow-nucl"
|
|
915
|
+
elsif (index (.method$, "external"))
|
|
916
|
+
segmentation_name$ = "extern"
|
|
917
|
+
elsif (index (.method$, "loudness")) ; automatic, loudness peaks
|
|
918
|
+
segmentation_name$ = "loudness"
|
|
919
|
+
elsif (index (.method$, "BP-filtered")) ; automatic, peaks in bandpass filters speech
|
|
920
|
+
segmentation_name$ = "int-BP"
|
|
921
|
+
elsif (index (.method$, "Automatic: acoustic syllables"))
|
|
922
|
+
segmentation_name$ = "asyll"
|
|
923
|
+
elsif (index (.method$, "in rhyme"))
|
|
924
|
+
segmentation_name$ = "rhyme"
|
|
925
|
+
elsif (index_regex (.method$, "in syllables.*and vowels"))
|
|
926
|
+
segmentation_name$ = "syll+vow"
|
|
927
|
+
elsif (index_regex (.method$, "in syllables.*and local peak"))
|
|
928
|
+
segmentation_name$ = "syll"
|
|
929
|
+
elsif (index (.method$, "Automatic: voiced portions"))
|
|
930
|
+
segmentation_name$ = "voiced"
|
|
931
|
+
else
|
|
932
|
+
@fatal_error: "Unknown segmentation type: '.method$'"
|
|
933
|
+
endif
|
|
934
|
+
endif
|
|
935
|
+
|
|
936
|
+
; Time step (frame rate) used for calculation of intensity and pitch
|
|
937
|
+
time_step = 'frame_period$'
|
|
938
|
+
anal_t1 = left_Time_range
|
|
939
|
+
anal_t2 = right_Time_range
|
|
940
|
+
|
|
941
|
+
; Plotting options
|
|
942
|
+
viewsize$ = "wide"
|
|
943
|
+
if (index (view$, "Compact"))
|
|
944
|
+
viewsize$ = "compact"
|
|
945
|
+
elsif (index (view$, "Large"))
|
|
946
|
+
viewsize$ = "large"
|
|
947
|
+
endif
|
|
948
|
+
rich = 0
|
|
949
|
+
if (index (view$, "rich"))
|
|
950
|
+
rich = 1
|
|
951
|
+
show_f0 = 1
|
|
952
|
+
endif
|
|
953
|
+
show_pitchrange = 0 ; Show pitch range
|
|
954
|
+
if (index (view$, "pitch range"))
|
|
955
|
+
show_pitchrange = 1
|
|
956
|
+
endif
|
|
957
|
+
if (rich = 0 and index (view$, "f0"))
|
|
958
|
+
show_f0 = 1
|
|
959
|
+
endif
|
|
960
|
+
draw_pitch_target_values = 0 ; Draw pitch target values (in ST) in prosogram
|
|
961
|
+
if (index (view$, "pitch targets"))
|
|
962
|
+
draw_pitch_target_values = 1
|
|
963
|
+
if (index (view$, "pitch targets in Hertz"))
|
|
964
|
+
draw_pitch_target_values = 2
|
|
965
|
+
endif
|
|
966
|
+
endif
|
|
967
|
+
|
|
968
|
+
; Thresholds
|
|
969
|
+
j = index_regex (thresholds$, "/T")
|
|
970
|
+
s$ = mid$ (thresholds$, 3, j-3) ; s$ = glissando threshold(s)
|
|
971
|
+
if (index (thresholds$, "adaptive")) ; adaptive glissando threshold (lower before pause)
|
|
972
|
+
adaptive_glissando = 1
|
|
973
|
+
s2$ = left$ (s$, 4)
|
|
974
|
+
glissando_low = 's2$'
|
|
975
|
+
s2$ = mid$ (s$, 6, 4)
|
|
976
|
+
glissando = 's2$'
|
|
977
|
+
else ; fixed glissando threshold
|
|
978
|
+
adaptive_glissando = 0
|
|
979
|
+
glissando = 's$'
|
|
980
|
+
glissando_low = glissando
|
|
981
|
+
endif
|
|
982
|
+
diffgt = extractNumber (thresholds$, "DG=")
|
|
983
|
+
mindur_ts = 0.035 ; Minimum duration for a tonal segment (default)
|
|
984
|
+
j = index (thresholds$, "dmin")
|
|
985
|
+
if (j > 0)
|
|
986
|
+
s$ = mid$ (thresholds$, j+5, length (thresholds$) -(j+4))
|
|
987
|
+
mindur_ts = 's$'
|
|
988
|
+
endif
|
|
989
|
+
|
|
990
|
+
if (not variableExists ("output_mode"))
|
|
991
|
+
output_mode$ = "Fill page with strips"
|
|
992
|
+
endif
|
|
993
|
+
outputmode$ = left$ (output_mode$, index(output_mode$, " ") -1)
|
|
994
|
+
|
|
995
|
+
volatile = 0 ; by default, process full signal and store results
|
|
996
|
+
if (index (parameter_calculation$, "Partial") > 0)
|
|
997
|
+
volatile = 1
|
|
998
|
+
endif
|
|
999
|
+
|
|
1000
|
+
scale_signal_amplitude = 0
|
|
1001
|
+
if (variableExists ("scale_signal_intensity")) ; Form contains field "Scale signal intensity"
|
|
1002
|
+
if (index (scale_signal_intensity$, "Scale"))
|
|
1003
|
+
scale_signal_amplitude = 1
|
|
1004
|
+
endif
|
|
1005
|
+
endif
|
|
1006
|
+
|
|
1007
|
+
clip_level = 0
|
|
1008
|
+
if (variableExists ("clipping_threshold")) ; Form contains field "Clipping threshold"
|
|
1009
|
+
if (clipping_threshold > 100 or clipping_threshold < 0)
|
|
1010
|
+
@fatal_error: "Invalid value for clipping threshold"
|
|
1011
|
+
endif
|
|
1012
|
+
clip_level = clipping_threshold
|
|
1013
|
+
endif
|
|
1014
|
+
|
|
1015
|
+
; Frequency range for pitch detection
|
|
1016
|
+
minimum_pitch = left_F0_detection_range
|
|
1017
|
+
maximum_pitch = right_F0_detection_range
|
|
1018
|
+
if (minimum_pitch = 0)
|
|
1019
|
+
; autorange_f0detect = 1
|
|
1020
|
+
else
|
|
1021
|
+
if (minimum_pitch < 40 or maximum_pitch > 800)
|
|
1022
|
+
@fatal_error: "Invalid F0 range: expected to be within 40 - 800 Hz range"
|
|
1023
|
+
endif
|
|
1024
|
+
if (minimum_pitch >= maximum_pitch)
|
|
1025
|
+
@fatal_error: "Invalid F0 range: lower limit > higher limit"
|
|
1026
|
+
endif
|
|
1027
|
+
endif
|
|
1028
|
+
if (left_Pitch_range == 0)
|
|
1029
|
+
auto_pitchrange = 1 ; automatic pitch range adjustment in plot (not in F0 detection)
|
|
1030
|
+
else
|
|
1031
|
+
auto_pitchrange = 0 ; manual pitch range adjustment for plot
|
|
1032
|
+
if (right_Pitch_range <= left_Pitch_range or left_Pitch_range <= 0)
|
|
1033
|
+
@fatal_error: "Invalid values for pitch range"
|
|
1034
|
+
endif
|
|
1035
|
+
endif
|
|
1036
|
+
ySTmin = left_Pitch_range
|
|
1037
|
+
ySTmax = right_Pitch_range
|
|
1038
|
+
|
|
1039
|
+
timeincr = time_interval_per_strip ; time increment for plot
|
|
1040
|
+
|
|
1041
|
+
@task_flags
|
|
1042
|
+
|
|
1043
|
+
if (draw_prosograms and (index (output_format$,"EMF")) and not windows)
|
|
1044
|
+
@error_msg: "Windows Metafiles (EMF) are supported on Windows systems only."
|
|
1045
|
+
endif
|
|
1046
|
+
|
|
1047
|
+
if (draw_prosograms and (index (output_format$,"JPG")))
|
|
1048
|
+
if (not fileReadable (path_ghostscript$))
|
|
1049
|
+
@error_msg: "JPG output requires Ghostscript, which is not found. Verify configuration for Ghostcript."
|
|
1050
|
+
endif
|
|
1051
|
+
endif
|
|
1052
|
+
|
|
1053
|
+
s$ = input_files$
|
|
1054
|
+
s$ = replace$ (s$, " ", "", 0)
|
|
1055
|
+
if (length(s$) < 1)
|
|
1056
|
+
input_files$ = chooseReadFile$ ("Select file")
|
|
1057
|
+
endif
|
|
1058
|
+
@debug_msg: "process_form: exit"
|
|
1059
|
+
endproc
|
|
1060
|
+
|
|
1061
|
+
|
|
1062
|
+
procedure initialization_multiple_files
|
|
1063
|
+
@debug_msg: "initialization_multiple_files: entry"
|
|
1064
|
+
# Unit (ST or Hz) used for expressing shown pitch range
|
|
1065
|
+
units$ = "Semitones"
|
|
1066
|
+
ySTmin = 0
|
|
1067
|
+
ySTmax = 100
|
|
1068
|
+
# diffST = difference in ST between ST-scale relative to 100Hz and that rel to 1Hz
|
|
1069
|
+
.diffST = 12 * log2(100/1)
|
|
1070
|
+
if (units$ = "Semitones")
|
|
1071
|
+
yHzmin = semitonesToHertz(ySTmin-.diffST)
|
|
1072
|
+
yHzmax = semitonesToHertz(ySTmax-.diffST)
|
|
1073
|
+
else
|
|
1074
|
+
ySTmin = hertzToSemitones(yHzmin) - hertzToSemitones(1)
|
|
1075
|
+
ySTmax = hertzToSemitones(yHzmax) - hertzToSemitones(1)
|
|
1076
|
+
endif
|
|
1077
|
+
fc_low = 300 ; BP filter
|
|
1078
|
+
fc_high = 3500 ; BP filter
|
|
1079
|
+
mindiff = 3 ; intensity difference threshold for local dips in convex hull
|
|
1080
|
+
diff_left = 2 ; intensity difference between local peak and left boundary of nucleus
|
|
1081
|
+
@debug_msg: "initialization_multiple_files: exit"
|
|
1082
|
+
endproc
|
|
1083
|
+
|
|
1084
|
+
|
|
1085
|
+
procedure initialization_per_file
|
|
1086
|
+
@debug_msg: "initialization_per_file: entry"
|
|
1087
|
+
signal_available = 0 ; speech signal loaded
|
|
1088
|
+
intensity_available = 0
|
|
1089
|
+
pitch_available = 0
|
|
1090
|
+
harmonicity_available = 0
|
|
1091
|
+
intbp_available = 0
|
|
1092
|
+
inthp_available = 0
|
|
1093
|
+
nuclei_available = 0
|
|
1094
|
+
nucldat_available = 0
|
|
1095
|
+
loudness_available = 0
|
|
1096
|
+
segfile_available = 0 ; annotation file with alignements of phonemes, syllables, etc.
|
|
1097
|
+
segmentation_available = 0 ; 1 if segmentation has been read from saved <basename>_nucl.TextGrid file
|
|
1098
|
+
stylization_available = 0 ; 1 if stylization has been read from saved <basename>_styl.PitchTier file
|
|
1099
|
+
profile_available = 0 ; 1 if profile data has been read from saved <basename>_profile_data.txt
|
|
1100
|
+
reuse_nucl = 0 ; 1 if nucleus file could be read and used as segmentation
|
|
1101
|
+
reuse_styl = 0
|
|
1102
|
+
phones_available = 0 ; 1 if phoneme tier is found in annotation TextGrid
|
|
1103
|
+
syllables_available = 0 ; 1 if syllable tier is found in annotation TextGrid
|
|
1104
|
+
words_available = 0 ; 1 if word tier is found in annotation TextGrid
|
|
1105
|
+
creak_available = 0 ; creak textgrid/tier found and read ?
|
|
1106
|
+
speaker_available = 0 ; speaker textgrid/tier found and read ?
|
|
1107
|
+
hesitation_available = 0 ; hesitation textgrid/tier found and read ?
|
|
1108
|
+
newgrid_available = 0
|
|
1109
|
+
nrof_nuclei_analysed = 0
|
|
1110
|
+
success = 1 ; 0 if error encountered while processing input file
|
|
1111
|
+
endproc
|
|
1112
|
+
|
|
1113
|
+
|
|
1114
|
+
procedure construct_filenames: .infname$
|
|
1115
|
+
# Construct filenames to be used for TextGrid, parameters, graphics and intermediate data files,
|
|
1116
|
+
# starting from filename of sound input file
|
|
1117
|
+
# <.infname$> full path of input file, including filename extension
|
|
1118
|
+
@debug_msg: "construct_filenames: entry, infname=<'.infname$'>"
|
|
1119
|
+
if (rindex (.infname$,".") == 0)
|
|
1120
|
+
@fatal_error: "Invalid filename for input file. Should include filename extension."
|
|
1121
|
+
endif
|
|
1122
|
+
@fname_parts: .infname$
|
|
1123
|
+
.fext$ = replace_regex$ (result3$, "(.)", "\L\1", 0) ; get filename extension and convert it to lowercase
|
|
1124
|
+
if (task <> task_validate_tiers)
|
|
1125
|
+
if (index(filename_extensions_sound$, ":'.fext$':") == 0)
|
|
1126
|
+
@debug_msg: "construct_filenames: filename extension=<'.fext$'>"
|
|
1127
|
+
@fatal_error: "Input file ('.infname$') should be an audio file supported by Praat Open LongSound, with extension .wav, .aiff, .aifc, .nist, .sound, .flac, or .mp3"
|
|
1128
|
+
endif
|
|
1129
|
+
endif
|
|
1130
|
+
basename$ = result2$
|
|
1131
|
+
indir$ = result4$
|
|
1132
|
+
.path$ = indir$ + basename$
|
|
1133
|
+
signalfile$ = .path$ + "." + .fext$
|
|
1134
|
+
pitchfile$ = .path$ + ".Pitch"
|
|
1135
|
+
segfile$ = .path$ + ".TextGrid" ; TextGrid with already available segmentation (phonetic and/or syllabic alignment)
|
|
1136
|
+
globalfile$ = indir$ + "globalsheet.txt" ; speaker profile data, for all input speech files in run (headerless spreadsheet)
|
|
1137
|
+
collectfile$ = indir$ + "_collect_.txt" ; append output from multiple input files
|
|
1138
|
+
|
|
1139
|
+
if (batch_mode and length (corpus$)) ; in batch mode, file paths may be defined for a given corpus
|
|
1140
|
+
; "corpus_subdir_sound$" is handled by batch commands
|
|
1141
|
+
; "corpus_subdir_img$" is handled below
|
|
1142
|
+
if (variableExists ("corpus_subdir_pitch$"))
|
|
1143
|
+
pitchfile$ = corpus_home$ + corpus_subdir_pitch$ + basename$ + ".Pitch"
|
|
1144
|
+
endif
|
|
1145
|
+
if (variableExists ("corpus_subdir_tg$"))
|
|
1146
|
+
segfile$ = corpus_home$ + corpus_subdir_tg$ + basename$ + ".TextGrid"
|
|
1147
|
+
endif
|
|
1148
|
+
if (variableExists ("corpus_subdir_data$"))
|
|
1149
|
+
.path$ = corpus_home$ + corpus_subdir_data$ + basename$
|
|
1150
|
+
globalfile$ = corpus_home$ + corpus_subdir_data$ + "globalsheet.txt" ; no basename here
|
|
1151
|
+
endif
|
|
1152
|
+
endif
|
|
1153
|
+
|
|
1154
|
+
intensityfile$ = .path$ + ".Intensity"
|
|
1155
|
+
intbpfile$ = .path$ + "_BP.Intensity"
|
|
1156
|
+
inthpfile$ = .path$ + "_HP.Intensity"
|
|
1157
|
+
loudnessfile$ = .path$ + "_loud.Intensity"
|
|
1158
|
+
creakfile$ = .path$ + "_creak.TextGrid" ; TextGrid file with creak annotation
|
|
1159
|
+
speakerfile$ = .path$ + "_speaker.TextGrid" ; TextGrid file with speaker annotation
|
|
1160
|
+
nuclfile$ = .path$ + "_nucl.TextGrid" ; nuclfile contains automatic segmentation etc.
|
|
1161
|
+
stylfile$ = .path$ + "_styl.PitchTier" ; stylization as a sequence of targets, in Praat's PitchTier format
|
|
1162
|
+
stylpitchfile$ = .path$ + "_styl.Pitch" ; stylization as pitch samples
|
|
1163
|
+
targetsfile$ = .path$ + "_styl.txt" ; stylization as a sequence of targets and segmentations format
|
|
1164
|
+
statsfile$ = .path$ + "_profile.txt" ; prosodic profile report in text format
|
|
1165
|
+
sheetfile$ = .path$ + "_data.txt" ; syllable data (format: headerless spreadsheet file)
|
|
1166
|
+
tablefile$ = .path$ + "_table.txt" ; long format syllabic features, adding categorical variables for statistic analysis
|
|
1167
|
+
profile_file$ = .path$ + "_profile_data.txt" ; prosodic profile data, for current input speech file (headerless spreadsheet file)
|
|
1168
|
+
autosegfile$ = .path$ + "_auto.TextGrid" ; TextGrid file with output from automatic segmentation into syllables
|
|
1169
|
+
polytonia_file$ = .path$ + "_polytonia.TextGrid" ; TextGrid file with output from Polytonia
|
|
1170
|
+
boundaryfile$ = .path$ + "_boundary.TextGrid" ; TextGrid file with detected prosodic boundaries
|
|
1171
|
+
hesitfile$ = .path$ + "_hesit.TextGrid" ; TextGrid file with detected hesitations
|
|
1172
|
+
evalfile$ = .path$ + "_eval.txt"
|
|
1173
|
+
harmonicityfile$ = .path$ + ".Harmonicity"
|
|
1174
|
+
|
|
1175
|
+
|
|
1176
|
+
; Special handling of output filename in batch mode
|
|
1177
|
+
; The variable <output_filename$> is set in script form or in batch command "prosogram_variants"
|
|
1178
|
+
output_fname$ = output_filename$ ; local copy for modifications
|
|
1179
|
+
if (batch_mode and (task == task_prosogram or task == task_segmentation))
|
|
1180
|
+
if (length (output_filename$))
|
|
1181
|
+
; filename explicitly given, keep it as such; does not use corpus_home nor corpus_subdir_img
|
|
1182
|
+
file_numbering = 0
|
|
1183
|
+
elsif (length (corpus$))
|
|
1184
|
+
if (variableExists ("corpus_subdir_img$"))
|
|
1185
|
+
output_fname$ = corpus_home$ + corpus_subdir_img$ + basename$ + "_" + output_suffix$
|
|
1186
|
+
else
|
|
1187
|
+
output_fname$ = corpus_home$ + basename$ + "_" + output_suffix$
|
|
1188
|
+
endif
|
|
1189
|
+
else
|
|
1190
|
+
output_fname$ = indir$ + basename$ + "_" + output_suffix$
|
|
1191
|
+
endif
|
|
1192
|
+
if (task == task_prosogram and draw_prosograms)
|
|
1193
|
+
@msg: "Graphics file will be written to (path+basename): <'output_fname$'>"
|
|
1194
|
+
endif
|
|
1195
|
+
@debug_msg: "construct_filenames: batch_mode, output_fname=<'output_fname$'>"
|
|
1196
|
+
if (task == task_segmentation)
|
|
1197
|
+
autosegfile$ = corpus_home$ + basename$ + output_suffix$ + ".TextGrid"
|
|
1198
|
+
endif
|
|
1199
|
+
endif
|
|
1200
|
+
if (not batch_mode) ; script mode
|
|
1201
|
+
if (length (indir$))
|
|
1202
|
+
output_fname$ = replace$ (output_fname$, "<input_directory>", indir$, 1)
|
|
1203
|
+
else
|
|
1204
|
+
output_fname$ = replace$ (output_fname$, "<input_directory>/", "", 1)
|
|
1205
|
+
endif
|
|
1206
|
+
output_fname$ = replace$ (output_fname$, "<basename>", basename$, 1)
|
|
1207
|
+
output_fname$ = replace$ (output_fname$, "//", "/", 1)
|
|
1208
|
+
if (single_fname_graphics_output) ; all graphics filenames have same basename followed by number
|
|
1209
|
+
output_filename$ = output_fname$ ; keep name for all graphics output files
|
|
1210
|
+
else
|
|
1211
|
+
file_ctr = 1 ; counter for graphics filename
|
|
1212
|
+
nrof_pages = 0
|
|
1213
|
+
endif
|
|
1214
|
+
@debug_msg: "construct_filenames: script_mode, output_fname=<'output_fname$'>"
|
|
1215
|
+
endif
|
|
1216
|
+
|
|
1217
|
+
@debug_msg: "construct_filenames: exit, output_fname=<'output_fname$'>"
|
|
1218
|
+
endproc
|
|
1219
|
+
|
|
1220
|
+
|
|
1221
|
+
procedure read_annotation_file
|
|
1222
|
+
@debug_msg: "read_annotation_file: entry"
|
|
1223
|
+
segfile_available = 0
|
|
1224
|
+
if (fileReadable (segfile$))
|
|
1225
|
+
segfile_available = 1
|
|
1226
|
+
@gr_printline: "Loading annotation TextGrid from file <'segfile$'>"
|
|
1227
|
+
segmentationID = Read from file: segfile$
|
|
1228
|
+
Rename: "annotation_input"
|
|
1229
|
+
elsif (needs_segm_tg)
|
|
1230
|
+
@fatal_error: "Cannot open TextGrid with segmentation from file <'segfile$'>"
|
|
1231
|
+
endif
|
|
1232
|
+
@debug_msg: "read_annotation_file: exit"
|
|
1233
|
+
endproc
|
|
1234
|
+
|
|
1235
|
+
|
|
1236
|
+
procedure corpus_conversion: .step, .gridname$, .corpus$
|
|
1237
|
+
# In-place conversion of a TextGrid object
|
|
1238
|
+
# <.step> 1 = before preparation of plotted tiers, 2 = after preparation
|
|
1239
|
+
@debug_msg: "corpus_conversion: entry, gridname=<'.gridname$'>"
|
|
1240
|
+
.grid = '.gridname$'
|
|
1241
|
+
selectObject: .grid
|
|
1242
|
+
|
|
1243
|
+
if (.corpus$ = "Portes" and .step = 1)
|
|
1244
|
+
@tier_number_by_name: .grid, "speaker"
|
|
1245
|
+
if (not result)
|
|
1246
|
+
@tier_number_by_name: .grid, "turns"
|
|
1247
|
+
if (result)
|
|
1248
|
+
Duplicate tier: result, 1, "speaker"
|
|
1249
|
+
Replace interval texts: 1, 1, 0, "([a-zA-Z]*)[0-9]*", "\1"
|
|
1250
|
+
endif
|
|
1251
|
+
endif
|
|
1252
|
+
endif
|
|
1253
|
+
|
|
1254
|
+
if (.corpus$ = "rhapsodie" and .step = 1)
|
|
1255
|
+
@tier_number_by_name: .grid, "locuteur"
|
|
1256
|
+
if (result)
|
|
1257
|
+
Set tier name: result, "speaker"
|
|
1258
|
+
endif
|
|
1259
|
+
endif
|
|
1260
|
+
|
|
1261
|
+
if (.corpus$ = "cprom" and .step = 1)
|
|
1262
|
+
@process_cprom: .grid
|
|
1263
|
+
endif
|
|
1264
|
+
|
|
1265
|
+
if (.corpus$ = "gvlex" and .step = 1)
|
|
1266
|
+
@lpa_sampa: .grid, corpus_tier_phon$
|
|
1267
|
+
@lpa_sampa: .grid, corpus_tier_syll$
|
|
1268
|
+
endif
|
|
1269
|
+
|
|
1270
|
+
if (.corpus$ = "Kotsifas_2015" and .step = 1)
|
|
1271
|
+
@tier_number_by_name: .grid, "Intonation"
|
|
1272
|
+
if (result)
|
|
1273
|
+
Set tier name: result, "intonation"
|
|
1274
|
+
endif
|
|
1275
|
+
@tier_number_by_name: .grid, "phon"
|
|
1276
|
+
if (result)
|
|
1277
|
+
.phon_tier = result
|
|
1278
|
+
.ni = Get number of intervals: .phon_tier
|
|
1279
|
+
for .j to .ni
|
|
1280
|
+
.t1s = Get start point: .phon_tier, .j
|
|
1281
|
+
.label$ = Get label of interval: .phon_tier, .j
|
|
1282
|
+
.pos = index (.label$, """") ; position of SAMPA primary stress mark " (double quote)
|
|
1283
|
+
if (.pos > 0 and .pos <> 1) ; bad position
|
|
1284
|
+
.s$ = replace_regex$ (.label$, """", "", 0)
|
|
1285
|
+
Set interval text: .phon_tier, .j, .s$
|
|
1286
|
+
@msg "Error: interval in phon tier at '.t1s:4' contains stress in wrong position. (label= '.label$')"
|
|
1287
|
+
endif
|
|
1288
|
+
endfor
|
|
1289
|
+
endif
|
|
1290
|
+
endif
|
|
1291
|
+
|
|
1292
|
+
if (.corpus$ = "RUHRCAT" and .step = 2)
|
|
1293
|
+
@tier_number_by_name: .grid, "phon"
|
|
1294
|
+
if (result)
|
|
1295
|
+
Replace interval text... result 0 0 "^j([aOoueE@])" "i\1" Regular Expressions
|
|
1296
|
+
Replace interval text... result 0 0 "^w([o0iaeE])" "u\1" Regular Expressions
|
|
1297
|
+
Replace interval text... result 0 0 "^w\\ct" "uO" Regular Expressions
|
|
1298
|
+
Replace interval text... result 0 0 "^j\\ct" "iO" Regular Expressions
|
|
1299
|
+
Replace interval text... result 0 0 "^j\\ef" "iE" Regular Expressions
|
|
1300
|
+
Replace interval text... result 0 0 "(\\:f)*" "" Regular Expressions
|
|
1301
|
+
Replace interval text... result 0 0 "^mmm$" "m=" Regular Expressions
|
|
1302
|
+
endif
|
|
1303
|
+
@tier_number_by_name: .grid, "syll"
|
|
1304
|
+
.syll_tier = result
|
|
1305
|
+
if (result)
|
|
1306
|
+
Replace interval text... result 0 0 "^mmm$" "m=" Regular Expressions
|
|
1307
|
+
endif
|
|
1308
|
+
; Create prominence and hesitation tiers from "DM+Hes" tier
|
|
1309
|
+
@tier_number_by_name: .grid, "DM.Hes"
|
|
1310
|
+
if (result)
|
|
1311
|
+
.deliv_tier = result
|
|
1312
|
+
.nt = Get number of tiers
|
|
1313
|
+
.hes_tier = .nt + 1
|
|
1314
|
+
Duplicate tier: .syll_tier, .hes_tier, "hes"
|
|
1315
|
+
.ni = Get number of intervals: .syll_tier
|
|
1316
|
+
for .j to .ni
|
|
1317
|
+
.t1s = Get start time of interval: .syll_tier, .j
|
|
1318
|
+
.t2s = Get end time of interval: .syll_tier, .j
|
|
1319
|
+
@interval_from_time: .grid, .deliv_tier, .t1s+(.t2s-.t1s)/2, "interv"
|
|
1320
|
+
.label$ = Get label of interval: .deliv_tier, interv
|
|
1321
|
+
if (index (.label$, "Hes"))
|
|
1322
|
+
.s$ = "H"
|
|
1323
|
+
else
|
|
1324
|
+
.s$ = ""
|
|
1325
|
+
endif
|
|
1326
|
+
Set interval text: .hes_tier, .j, .s$
|
|
1327
|
+
endfor
|
|
1328
|
+
endif
|
|
1329
|
+
endif ; (.corpus$ = "RUHRCAT" and .step = 2)
|
|
1330
|
+
@debug_msg: "corpus_conversion: exit"
|
|
1331
|
+
endproc
|
|
1332
|
+
|
|
1333
|
+
|
|
1334
|
+
procedure prepare_plotted_textgrid: ltiers$
|
|
1335
|
+
# Prepare (create and fill) TextGrid with user-selected tiers plotted in prosograms
|
|
1336
|
+
# <ltiers> list of (comma-separated) tier numbers or names, to be copied from segmentation TextGrid
|
|
1337
|
+
# When tier name or number is preceded by "*", the tier content is converted from SAMPA to IPA symbols
|
|
1338
|
+
# Returns <nrofplottedtiers>, the number of tiers in resulting TextGrid
|
|
1339
|
+
@debug_msg: "prepare_plotted_textgrid: entry"
|
|
1340
|
+
if (index(ltiers$,""""))
|
|
1341
|
+
@fatal_error: "Invalid character (double quote) in field <Tiers to show>. Please check content of this field."
|
|
1342
|
+
endif
|
|
1343
|
+
nrofplottedtiers = 0
|
|
1344
|
+
if (segfile_available == 0 and (polytonia_annotation or boundary_annotation))
|
|
1345
|
+
newgridID = Create TextGrid: signal_start, signal_finish, "dummy", ""
|
|
1346
|
+
newgrid_available = 1
|
|
1347
|
+
endif
|
|
1348
|
+
if (segfile_available)
|
|
1349
|
+
selectObject: segmentationID
|
|
1350
|
+
tiers = 0
|
|
1351
|
+
nrofTiers = Get number of tiers
|
|
1352
|
+
; Get number of speaker tier in segmentation TextGrid
|
|
1353
|
+
@tier_number_by_name: segmentationID, "^[Ss]peaker$"
|
|
1354
|
+
speaker_tier_in = result
|
|
1355
|
+
; Select tiers (and convert) from segmentation TextGrid
|
|
1356
|
+
repeat
|
|
1357
|
+
@next_field: ltiers$
|
|
1358
|
+
ltiers$ = result3$ ; rest of fields
|
|
1359
|
+
if (result) ; next field found
|
|
1360
|
+
.field$ = result2$
|
|
1361
|
+
@debug_msg: "prepare_plotted_textgrid: next_field=<'.field$'>"
|
|
1362
|
+
.convert = 0 ; convert SAMPA to IPA
|
|
1363
|
+
if (left$ (.field$, 1) = "*")
|
|
1364
|
+
.field$ = right$ (.field$, length(.field$) - 1)
|
|
1365
|
+
.convert = 1
|
|
1366
|
+
endif
|
|
1367
|
+
@is_number: .field$
|
|
1368
|
+
if (result == 0) ; next field is not specified as a number
|
|
1369
|
+
@tier_number_by_name: segmentationID, .field$
|
|
1370
|
+
.tier_in = result
|
|
1371
|
+
if (result == 0) ; tier name not found
|
|
1372
|
+
@error_msg: "No tier named <'.field$'> found in input textgrid. Change content of field <Tiers to show>."
|
|
1373
|
+
@error_msg: "This tier will be skipped in prosogram."
|
|
1374
|
+
endif
|
|
1375
|
+
else ; next field is a number
|
|
1376
|
+
.tier_in = '.field$'
|
|
1377
|
+
if (.tier_in > nrofTiers)
|
|
1378
|
+
@error_msg: "Tier '.tier_in' not found. Input textgrid has only 'nrofTiers' tiers. Change content of field <Tiers to show>."
|
|
1379
|
+
@error_msg: "This tier will be skipped in prosogram."
|
|
1380
|
+
.tier_in = 0
|
|
1381
|
+
endif
|
|
1382
|
+
endif
|
|
1383
|
+
if (.tier_in > 0) ; valid tier to add
|
|
1384
|
+
selectObject: segmentationID
|
|
1385
|
+
if (tiers == 0) ; this is first tier to add to plotted grid
|
|
1386
|
+
.tmpID = Extract tier: .tier_in
|
|
1387
|
+
newgridID = Into TextGrid
|
|
1388
|
+
Rename: "plotted_tiers"
|
|
1389
|
+
newgrid_available = 1
|
|
1390
|
+
removeObject: .tmpID
|
|
1391
|
+
else
|
|
1392
|
+
@grid_append_tier: segmentationID, .tier_in, "newgridID"
|
|
1393
|
+
selectObject: newgridID
|
|
1394
|
+
endif
|
|
1395
|
+
tiers += 1
|
|
1396
|
+
if (.convert)
|
|
1397
|
+
@convert_sampa_ipa: newgridID, tiers
|
|
1398
|
+
endif
|
|
1399
|
+
endif
|
|
1400
|
+
endif
|
|
1401
|
+
until (result == 0) ; no more fields in list
|
|
1402
|
+
nrofplottedtiers = tiers
|
|
1403
|
+
if (nrofplottedtiers > 0)
|
|
1404
|
+
@textgrid_disable_textstyle: newgridID
|
|
1405
|
+
endif
|
|
1406
|
+
endif ; segfile_available
|
|
1407
|
+
if (show_pseudosyllables)
|
|
1408
|
+
nrofplottedtiers += 1 ; results will appear in extra tier
|
|
1409
|
+
endif
|
|
1410
|
+
if (polytonia_annotation)
|
|
1411
|
+
nrofplottedtiers += 1 ; results will appear in extra tier
|
|
1412
|
+
endif
|
|
1413
|
+
if (boundary_annotation)
|
|
1414
|
+
nrofplottedtiers += 1 ; results will appear in extra tier
|
|
1415
|
+
endif
|
|
1416
|
+
if (stress_annotation)
|
|
1417
|
+
nrofplottedtiers += 1 ; results will appear in extra tier
|
|
1418
|
+
endif
|
|
1419
|
+
@debug_msg: "prepare_plotted_textgrid: exit"
|
|
1420
|
+
endproc
|
|
1421
|
+
|
|
1422
|
+
|
|
1423
|
+
procedure peek_signal
|
|
1424
|
+
@debug_msg: "peek_signal: entry"
|
|
1425
|
+
@fname_parts: signalfile$
|
|
1426
|
+
.fext$ = replace_regex$ (result3$, "(.)", "\L\1", 0) ; to lowercase
|
|
1427
|
+
if (.fext$ = "sound")
|
|
1428
|
+
.tmpID = Read from file: signalfile$
|
|
1429
|
+
else
|
|
1430
|
+
.tmpID = Open long sound file: signalfile$
|
|
1431
|
+
endif
|
|
1432
|
+
signal_start = Get starting time
|
|
1433
|
+
signal_finish = Get finishing time
|
|
1434
|
+
removeObject: .tmpID
|
|
1435
|
+
@debug_msg: "peek_signal: exit"
|
|
1436
|
+
endproc
|
|
1437
|
+
|
|
1438
|
+
|
|
1439
|
+
procedure read_signal
|
|
1440
|
+
# Read signal. Apply clipping if clip > 0. Apply scaling of (scale_signal_amplitude > 0).
|
|
1441
|
+
@debug_msg: "read_signal: entry"
|
|
1442
|
+
soundID = Read from file: signalfile$
|
|
1443
|
+
fullsoundID = soundID
|
|
1444
|
+
signal_start = Get starting time
|
|
1445
|
+
signal_finish = Get finishing time
|
|
1446
|
+
if (clip_level > 0) ; Avoid error related to speech files clipped at max amplitude.
|
|
1447
|
+
@msg: "Clipping speech signal at 'clip_level' percent..."
|
|
1448
|
+
clip = clip_level/100
|
|
1449
|
+
Formula... if (abs(self)>clip) then if self>0 then clip else -clip fi else self fi
|
|
1450
|
+
endif
|
|
1451
|
+
if (scale_signal_amplitude)
|
|
1452
|
+
@msg: "Scaling speech signal at 70 dB..."
|
|
1453
|
+
Scale intensity: 70.0
|
|
1454
|
+
endif
|
|
1455
|
+
signal_available = 1
|
|
1456
|
+
@debug_msg: "read_signal: exit"
|
|
1457
|
+
endproc
|
|
1458
|
+
|
|
1459
|
+
|
|
1460
|
+
procedure read_parameter_files
|
|
1461
|
+
@debug_msg: "read_parameter_files: entry"
|
|
1462
|
+
if ((not volatile) and fileReadable (pitchfile$))
|
|
1463
|
+
pitchID = Read from file: pitchfile$
|
|
1464
|
+
.t = Get time step
|
|
1465
|
+
if (.t > time_step)
|
|
1466
|
+
@msg: "Pitch file on disk has larger time step."
|
|
1467
|
+
removeObject: pitchID
|
|
1468
|
+
else
|
|
1469
|
+
.t2 = Get end time
|
|
1470
|
+
if (.t2 + 5*time_step < signal_finish)
|
|
1471
|
+
@msg: "Pitch file end time ('.t2' s) < speech signal end time ('signal_finish' s)"
|
|
1472
|
+
endif
|
|
1473
|
+
pitch_available = 1
|
|
1474
|
+
@gr_printline: "Consulted pitch from file <'pitchfile$'> (time step='.t:3')"
|
|
1475
|
+
endif
|
|
1476
|
+
endif
|
|
1477
|
+
if (show_harmonicity)
|
|
1478
|
+
if (fileReadable (harmonicityfile$))
|
|
1479
|
+
harmonicityID = Read from file: harmonicityfile$
|
|
1480
|
+
@msg: "Consulted harmonicity from file <'harmonicityfile$'>"
|
|
1481
|
+
harmonicity_available = 1
|
|
1482
|
+
endif
|
|
1483
|
+
endif
|
|
1484
|
+
if (needs_loudness)
|
|
1485
|
+
if (fileReadable (loudnessfile$))
|
|
1486
|
+
loudnessID = Read from file: loudnessfile$
|
|
1487
|
+
@msg: "Consulted loudness from file <'loudnessfile$'>"
|
|
1488
|
+
loudness_available = 1
|
|
1489
|
+
endif
|
|
1490
|
+
endif
|
|
1491
|
+
if (needs_intbp and not volatile)
|
|
1492
|
+
if (fileReadable (intbpfile$))
|
|
1493
|
+
intbpID = Read from file: intbpfile$
|
|
1494
|
+
.t = Get time step
|
|
1495
|
+
intbp_available = 1
|
|
1496
|
+
@gr_printline: "Consulted BP intensity from file <'intbpfile$'> (time step='.t:3')"
|
|
1497
|
+
endif
|
|
1498
|
+
endif
|
|
1499
|
+
@debug_msg: "read_parameter_files: exit"
|
|
1500
|
+
endproc
|
|
1501
|
+
|
|
1502
|
+
|
|
1503
|
+
procedure calculate_pitch_int
|
|
1504
|
+
@debug_msg: "calculate_pitch_int: entry"
|
|
1505
|
+
if (not signal_available)
|
|
1506
|
+
@read_signal
|
|
1507
|
+
endif
|
|
1508
|
+
; Standard settings, see FAQ Pitch Analysis
|
|
1509
|
+
tstep = 0 ; automatic, i.e. 0.75/pitch_floor, e.g. for 60Hz, time step = 0.0125 s
|
|
1510
|
+
voicing_threshold = 0.45
|
|
1511
|
+
silence_threshold = 0.03
|
|
1512
|
+
octave_cost = 0.01
|
|
1513
|
+
octave_jump_cost = 0.35
|
|
1514
|
+
vuv_cost = 0.14
|
|
1515
|
+
; Modified settings
|
|
1516
|
+
tstep = time_step ; typically 0.01 or 0.005 s
|
|
1517
|
+
|
|
1518
|
+
; Actual pitch determination, either explicit F0 range or autorange (2 pass, when minimum_pitch = 0)
|
|
1519
|
+
.f0min = minimum_pitch ; local variable
|
|
1520
|
+
.f0max = maximum_pitch ; local variable
|
|
1521
|
+
if (.f0min = 0) ; autorange, pass 1
|
|
1522
|
+
@debug_msg: "calculate_pitch_int: pass 1"
|
|
1523
|
+
.tstep1 = 0.01 ; speed up pass 1
|
|
1524
|
+
.f0min = 65
|
|
1525
|
+
.f0max = 800
|
|
1526
|
+
@from_Hz_to_ST_rel_1: .f0min
|
|
1527
|
+
.minST = result
|
|
1528
|
+
@from_Hz_to_ST_rel_1: .f0max
|
|
1529
|
+
.maxST = result
|
|
1530
|
+
@msg: "Calculating pitch in autorange mode, Pass 1..."
|
|
1531
|
+
@msg: ".'tab$'Time step='.tstep1:2', F0 range='.f0min:0'-'.f0max:0' Hz ('.minST:1'-'.maxST:1' ST), Voicing threshold='voicing_threshold:2'"
|
|
1532
|
+
selectObject: soundID
|
|
1533
|
+
pitchID = To Pitch (ac): .tstep1, .f0min, 15, "no", silence_threshold, voicing_threshold, octave_cost, octave_jump_cost, vuv_cost, .f0max
|
|
1534
|
+
.p05 = Get quantile: 0, 0, 0.05, "Hertz"
|
|
1535
|
+
.p25 = Get quantile: 0, 0, 0.25, "Hertz"
|
|
1536
|
+
.p50 = Get quantile: 0, 0, 0.50, "Hertz"
|
|
1537
|
+
.p75 = Get quantile: 0, 0, 0.75, "Hertz"
|
|
1538
|
+
.p95 = Get quantile: 0, 0, 0.95, "Hertz"
|
|
1539
|
+
removeObject: pitchID
|
|
1540
|
+
@msg: ".'tab$'Pitch percentiles 5%='.p05:0' 25%='.p25:0' 50%='.p50:0' 75%='.p75:0' 95%='.p95:0' (Hz)"
|
|
1541
|
+
.f0min = floor ( max (semitonesToHertz(hertzToSemitones(.p50) - 12), 50) ) ; 12 ST below median pitch, with minimum of 50 Hz
|
|
1542
|
+
.f0max = floor ( min (semitonesToHertz(hertzToSemitones(.p50) + 18), 1000) ) ; 18 ST above median pitch, with maximum of 1000 Hz
|
|
1543
|
+
.minST = hertzToSemitones (.f0min) - hertzToSemitones(1)
|
|
1544
|
+
.maxST = hertzToSemitones (.f0max) - hertzToSemitones(1)
|
|
1545
|
+
@msg: ".'tab$'Selected F0 detection range for pass 2: '.f0min'-'.f0max' Hz, '.minST:1'-'.maxST:1' ST"
|
|
1546
|
+
endif
|
|
1547
|
+
@debug_msg: "calculate_pitch_int: pass 2"
|
|
1548
|
+
@from_Hz_to_ST_rel_1: .f0min
|
|
1549
|
+
.minST = result
|
|
1550
|
+
@from_Hz_to_ST_rel_1: .f0max
|
|
1551
|
+
.maxST = result
|
|
1552
|
+
@msg: "Calculating pitch for fixed frequency range..."
|
|
1553
|
+
@msg: ".'tab$'Time step='tstep:2', F0 range='.f0min'-'.f0max' Hz ('.minST:1'-'.maxST:1' ST), Voicing threshold='voicing_threshold:2'"
|
|
1554
|
+
selectObject: soundID
|
|
1555
|
+
pitchID = To Pitch (ac): tstep, .f0min, 15, "no", silence_threshold, voicing_threshold, octave_cost, octave_jump_cost, vuv_cost, .f0max
|
|
1556
|
+
.p50 = Get quantile: 0, 0, 0.50, "Hertz"
|
|
1557
|
+
.p95 = Get quantile: 0, 0, 0.95, "Hertz"
|
|
1558
|
+
.p05 = Get quantile: 0, 0, 0.05, "Hertz"
|
|
1559
|
+
@msg: ".'tab$'Pitch percentiles 50%='.p50:0' 5%='.p05:0' 95%='.p95:0' (Hz)"
|
|
1560
|
+
if (not volatile)
|
|
1561
|
+
@msg: "Writing Pitch object to 'pitchfile$'"
|
|
1562
|
+
Write to binary file: pitchfile$
|
|
1563
|
+
endif
|
|
1564
|
+
pitch_available = 1
|
|
1565
|
+
@debug_msg: "calculate_pitch_int: exit"
|
|
1566
|
+
endproc
|
|
1567
|
+
|
|
1568
|
+
|
|
1569
|
+
procedure calculate_intensity_bp_filtered
|
|
1570
|
+
@msg: "Calculating intensity of BP filtered signal (fc_low='fc_low:0', fc_high='fc_high:0') ..."
|
|
1571
|
+
if (avoid_insuff_memory)
|
|
1572
|
+
call intensity_bp_filter_longsound_compact "'signalfile$'" fc_low fc_high intbpID
|
|
1573
|
+
else
|
|
1574
|
+
if (not signal_available)
|
|
1575
|
+
@read_signal
|
|
1576
|
+
endif
|
|
1577
|
+
selectObject: soundID
|
|
1578
|
+
tmpfsID = Filter (pass Hann band): fc_low, fc_high, 100
|
|
1579
|
+
selectObject: tmpfsID
|
|
1580
|
+
intbpID = To Intensity: 100, time_step
|
|
1581
|
+
intbpID = selected ("Intensity", -1) ; keep this line !!
|
|
1582
|
+
removeObject: tmpfsID
|
|
1583
|
+
endif
|
|
1584
|
+
if (not volatile)
|
|
1585
|
+
selectObject: intbpID
|
|
1586
|
+
Write to text file: intbpfile$
|
|
1587
|
+
endif
|
|
1588
|
+
intbp_available = 1
|
|
1589
|
+
endproc
|
|
1590
|
+
|
|
1591
|
+
|
|
1592
|
+
procedure bp_filter: .soundin, .fc_low, .fc_high, .varname$
|
|
1593
|
+
# BP-filter sound object and put reference to resulting object in variable with name <.varname>
|
|
1594
|
+
selectObject: .soundin
|
|
1595
|
+
.out = Filter (pass Hann band): .fc_low, .fc_high, 100
|
|
1596
|
+
'.varname$' = .out
|
|
1597
|
+
endproc
|
|
1598
|
+
|
|
1599
|
+
|
|
1600
|
+
procedure intensity_bp_filter_longsound_compact .fname$ .fc_low .fc_high .varname$
|
|
1601
|
+
# Filter signal and compute intensity. Filtered signal is temporarily stored as file.
|
|
1602
|
+
# Use this procedure to avoid insufficient memory.
|
|
1603
|
+
# Use double quotes if filename contains blanks:
|
|
1604
|
+
# call intensity_bp_filter_longsound_compact "'filename$'" 300 2500 myvar
|
|
1605
|
+
# .fname$ filename of sound file to be filtered
|
|
1606
|
+
.tmpfile$ = prefix$ + "_filtered.wav"
|
|
1607
|
+
.longsound = Open long sound file: .fname$
|
|
1608
|
+
.t1 = Get start time
|
|
1609
|
+
.endtime = Get end time
|
|
1610
|
+
.maxdurpart = 300 ; duration of part of signal to be filtered separatedly due to available memory
|
|
1611
|
+
.partnr = 1
|
|
1612
|
+
repeat
|
|
1613
|
+
.t2 = min (.endtime, .t1 + .maxdurpart)
|
|
1614
|
+
selectObject: .longsound
|
|
1615
|
+
.part = Extract part... .t1 .t2 yes
|
|
1616
|
+
if (clip_level > 0)
|
|
1617
|
+
clip = clip_level/100
|
|
1618
|
+
Formula... if (abs(self)>clip) then if self>0 then clip else -clip fi else self fi
|
|
1619
|
+
endif
|
|
1620
|
+
.filt = Filter (pass Hann band)... .fc_low .fc_high 100
|
|
1621
|
+
Formula... if (abs(self)>0.98) then if self>0 then 0.98 else -0.98 fi else self fi
|
|
1622
|
+
if (.partnr == 1)
|
|
1623
|
+
Save as WAV file: .tmpfile$
|
|
1624
|
+
else
|
|
1625
|
+
Append to existing sound file... '.tmpfile$'
|
|
1626
|
+
endif
|
|
1627
|
+
removeObject: .filt, .part
|
|
1628
|
+
.t1 = .t2
|
|
1629
|
+
.partnr += 1
|
|
1630
|
+
until (.t2 >= .endtime)
|
|
1631
|
+
removeObject: .longsound ; free memory
|
|
1632
|
+
.tmp = Read from file... '.tmpfile$'
|
|
1633
|
+
intbpID = To Intensity: 100, time_step
|
|
1634
|
+
intbpID = selected ("Intensity", -1) ; keep this line !!
|
|
1635
|
+
'.varname$' = intbpID
|
|
1636
|
+
removeObject: .tmp
|
|
1637
|
+
filedelete '.tmpfile$'
|
|
1638
|
+
endproc
|
|
1639
|
+
|
|
1640
|
+
|
|
1641
|
+
procedure calculate_loudness
|
|
1642
|
+
if (not signal_available)
|
|
1643
|
+
@read_signal
|
|
1644
|
+
endif
|
|
1645
|
+
execute loudness.praat 'soundID' 'signal_start' 'signal_finish' 'time_step'
|
|
1646
|
+
loudnessID = selected ("Intensity", -1)
|
|
1647
|
+
if (not volatile)
|
|
1648
|
+
Save as short text file: loudnessfile$
|
|
1649
|
+
endif
|
|
1650
|
+
loudness_available = 1
|
|
1651
|
+
endproc
|
|
1652
|
+
|
|
1653
|
+
|
|
1654
|
+
procedure calculate_parameters
|
|
1655
|
+
@debug_msg: "calculate_parameters: entry"
|
|
1656
|
+
if (not signal_available)
|
|
1657
|
+
@read_signal
|
|
1658
|
+
endif
|
|
1659
|
+
if (volatile)
|
|
1660
|
+
# Calculate parameters on the fly only for interval to stylize
|
|
1661
|
+
selectObject: soundID
|
|
1662
|
+
tmpsoundID = Extract part: anal_t1, anal_t2, "rectangular", 1.0, "yes"
|
|
1663
|
+
removeObject: soundID ; remove full signal
|
|
1664
|
+
soundID = tmpsoundID ; redefine soundID
|
|
1665
|
+
selectObject: soundID
|
|
1666
|
+
if (scale_signal_amplitude)
|
|
1667
|
+
Scale intensity: scale_signal_amplitude
|
|
1668
|
+
endif
|
|
1669
|
+
intensityID = To Intensity: 100, time_step
|
|
1670
|
+
intensityID = selected ("Intensity", -1) ; keep this line !!
|
|
1671
|
+
intensity_available = 1
|
|
1672
|
+
@calculate_pitch_int
|
|
1673
|
+
if (needs_loudness and not loudness_available)
|
|
1674
|
+
execute loudness.praat 'soundID' 'anal_t1' 'anal_t2' 'time_step'
|
|
1675
|
+
loudnessID = selected ("Intensity", -1)
|
|
1676
|
+
loudness_available = 1
|
|
1677
|
+
endif
|
|
1678
|
+
if (needs_intbp and not intbp_available)
|
|
1679
|
+
@calculate_intensity_bp_filtered
|
|
1680
|
+
intbp_available = 1
|
|
1681
|
+
endif
|
|
1682
|
+
else ; (volatile == 0)
|
|
1683
|
+
if (not intensity_available)
|
|
1684
|
+
|
|
1685
|
+
if (corpus$ = "rhapsodie")
|
|
1686
|
+
prefilterHP100 = 0
|
|
1687
|
+
if (basename$ = "Rhap-D0008")
|
|
1688
|
+
prefilterHP100 = 1
|
|
1689
|
+
endif
|
|
1690
|
+
endif
|
|
1691
|
+
|
|
1692
|
+
if (prefilterHP100) ; Apply HP filtering to avoid problems with low quality recordings by filtering low frequency band
|
|
1693
|
+
@msg: "Preprocessing (high-pass filtering) speech signal ('basename$')..."
|
|
1694
|
+
selectObject: soundID
|
|
1695
|
+
fc_low = 100
|
|
1696
|
+
tmpID = Filter (pass Hann band): fc_low, 0, 100
|
|
1697
|
+
removeObject: soundID
|
|
1698
|
+
soundID = tmpID
|
|
1699
|
+
endif
|
|
1700
|
+
if (prefilterLP) ; Apply LP filtering when fricatives are very intense
|
|
1701
|
+
@msg: "Preprocessing (low-pass filtering) speech signal... 'basename$'"
|
|
1702
|
+
selectObject: soundID
|
|
1703
|
+
fc = 2500
|
|
1704
|
+
tmpID = Filter (pass Hann band): 0, fc, 100
|
|
1705
|
+
removeObject: soundID
|
|
1706
|
+
soundID = tmpID
|
|
1707
|
+
endif
|
|
1708
|
+
|
|
1709
|
+
selectObject: soundID
|
|
1710
|
+
tmpID = To Intensity: 100, time_step
|
|
1711
|
+
intensityID = selected ("Intensity", -1) ; keep this line !!
|
|
1712
|
+
intensity_available = 1
|
|
1713
|
+
endif
|
|
1714
|
+
if (not pitch_available)
|
|
1715
|
+
@calculate_pitch_int
|
|
1716
|
+
endif
|
|
1717
|
+
if (show_harmonicity and not harmonicity_available)
|
|
1718
|
+
selectObject: soundID
|
|
1719
|
+
silence_threshold = 0.1
|
|
1720
|
+
nrofperiods = 1.0
|
|
1721
|
+
harmonicityID = To Harmonicity (cc)... time_step minimum_pitch silence_threshold nrofperiods
|
|
1722
|
+
harmonicity_available = 1
|
|
1723
|
+
endif
|
|
1724
|
+
if (needs_loudness and not loudness_available)
|
|
1725
|
+
execute loudness.praat 'soundID' 'signal_start' 'signal_finish' 'time_step'
|
|
1726
|
+
loudnessID = selected ("Intensity", -1)
|
|
1727
|
+
Write to short text file: loudnessfile$
|
|
1728
|
+
loudness_available = 1
|
|
1729
|
+
endif
|
|
1730
|
+
if (needs_intbp and not intbp_available)
|
|
1731
|
+
@calculate_intensity_bp_filtered
|
|
1732
|
+
endif
|
|
1733
|
+
endif ; # volatile
|
|
1734
|
+
# Calculate V/UV decision and store in TextGrid
|
|
1735
|
+
selectObject: pitchID
|
|
1736
|
+
pointprocID = To PointProcess
|
|
1737
|
+
vuvgridID = To TextGrid (vuv): 0.02, 0.01
|
|
1738
|
+
Rename: "vuv"
|
|
1739
|
+
removeObject: pointprocID
|
|
1740
|
+
@debug_msg: "calculate_parameters: exit"
|
|
1741
|
+
endproc
|
|
1742
|
+
|
|
1743
|
+
|
|
1744
|
+
procedure peek_settings
|
|
1745
|
+
# Read settings from intermediate data file (_nucl.TextGrid).
|
|
1746
|
+
@debug_msg: "peek_settings: entry"
|
|
1747
|
+
if (not fileReadable (nuclfile$))
|
|
1748
|
+
@fatal_error: "Cannot find nucleus file: <'nuclfile$'>'newline$'Obtain it first using Prosogram task: Calculate intermediate data files"
|
|
1749
|
+
else
|
|
1750
|
+
.tmpID = Read from file: nuclfile$
|
|
1751
|
+
@tier_number_by_name: .tmpID, "^settings$"
|
|
1752
|
+
@debug_msg: "peek_settings: settings tier nr='result'"
|
|
1753
|
+
if (result)
|
|
1754
|
+
.s$ = Get label of interval: result, 1
|
|
1755
|
+
@debug_msg: "peek_settings: settings tier label=<'.s$'>"
|
|
1756
|
+
segmentation_name$ = extractWord$ (.s$, "SEG=")
|
|
1757
|
+
adaptive_glissando = extractNumber (.s$, "ADAPT=")
|
|
1758
|
+
glissando = extractNumber (.s$, "GT=")
|
|
1759
|
+
glissando_low = glissando
|
|
1760
|
+
if (adaptive_glissando = undefined)
|
|
1761
|
+
adaptive_glissando = 0
|
|
1762
|
+
endif
|
|
1763
|
+
if (adaptive_glissando)
|
|
1764
|
+
glissando_low = extractNumber (.s$, "GT_low=")
|
|
1765
|
+
endif
|
|
1766
|
+
diffgt = extractNumber (.s$, "DG=")
|
|
1767
|
+
mindur_ts = extractNumber (.s$, "MINTS=")
|
|
1768
|
+
else
|
|
1769
|
+
@fatal_error: "Cannot find settings tier in nucleus file"
|
|
1770
|
+
endif
|
|
1771
|
+
removeObject: .tmpID
|
|
1772
|
+
endif
|
|
1773
|
+
@debug_msg: "peek_settings: exit"
|
|
1774
|
+
endproc
|
|
1775
|
+
|
|
1776
|
+
|
|
1777
|
+
procedure load_intermediate_data_files
|
|
1778
|
+
# Reads intermediate data files into data objects. Used in interactive mode.
|
|
1779
|
+
.nerrs = 0
|
|
1780
|
+
if (not fileReadable (stylfile$))
|
|
1781
|
+
@gr_printline: "Cannot open stylization file <'stylfile$'>"
|
|
1782
|
+
.nerrs += 1
|
|
1783
|
+
endif
|
|
1784
|
+
if (not fileReadable (sheetfile$))
|
|
1785
|
+
@gr_printline: "Cannot open prosodic features from file <'sheetfile$'>"
|
|
1786
|
+
.nerrs += 1
|
|
1787
|
+
endif
|
|
1788
|
+
if (not fileReadable (profile_file$))
|
|
1789
|
+
@gr_printline: "Cannot open prosodic file <'profile_file$'>"
|
|
1790
|
+
.nerrs += 1
|
|
1791
|
+
endif
|
|
1792
|
+
if (.nerrs == 0)
|
|
1793
|
+
@gr_printline: "Loading stylization from file <'stylfile$'>..."
|
|
1794
|
+
stylID = Read from file: stylfile$
|
|
1795
|
+
stylization_available = 1
|
|
1796
|
+
reuse_styl = 1
|
|
1797
|
+
removeObject: nucldatID
|
|
1798
|
+
@gr_printline: "Loading prosodic features from file <'sheetfile$'>..."
|
|
1799
|
+
nucldatID = Read TableOfReal from headerless spreadsheet file: sheetfile$
|
|
1800
|
+
@gr_printline: "Loading prosodic profile from file <'profile_file$'>..."
|
|
1801
|
+
@profile_table_load
|
|
1802
|
+
else
|
|
1803
|
+
@cleanup_current_file
|
|
1804
|
+
@cleanup_global
|
|
1805
|
+
@fatal_error: "Some data files could not be found. Please run task <Calculate intermediate data files> again before using Interactive mode."
|
|
1806
|
+
endif
|
|
1807
|
+
selectObject: nucldatID
|
|
1808
|
+
.nrows = Get number of rows
|
|
1809
|
+
if (nrof_nuclei_analysed <> .nrows)
|
|
1810
|
+
@fatal_error: "Mismatch between data in intermediate data files. Please run task <Calculate intermediate data files> again before using Interactive mode."
|
|
1811
|
+
endif
|
|
1812
|
+
endproc
|
|
1813
|
+
|
|
1814
|
+
|
|
1815
|
+
procedure get_segmentation: method
|
|
1816
|
+
# Calculate segmentation or read it from file
|
|
1817
|
+
@debug_msg: "get_segmentation: entry"
|
|
1818
|
+
if (method == segm_vnucl) ; vowel nuclei
|
|
1819
|
+
segmentation_name$ = "vow-nucl"
|
|
1820
|
+
elsif (method == segm_extern)
|
|
1821
|
+
segmentation_name$ = "extern"
|
|
1822
|
+
elsif (method == segm_aloudness) ; automatic, loudness peaks
|
|
1823
|
+
segmentation_name$ = "loudness"
|
|
1824
|
+
elsif (method == segm_anucl) ; automatic, peaks in bandpass filters speech
|
|
1825
|
+
segmentation_name$ = "int-BP"
|
|
1826
|
+
elsif (method == segm_msyllvow) ; syllabic nuclei
|
|
1827
|
+
segmentation_name$ = "syll+vow"
|
|
1828
|
+
elsif (method == segm_msyllpeak) ; syllabic nuclei
|
|
1829
|
+
segmentation_name$ = "syll"
|
|
1830
|
+
elsif (method == segm_mrhyme) ; syllable rhyme
|
|
1831
|
+
segmentation_name$ = "rhyme"
|
|
1832
|
+
elsif (method == segm_asyll) ; pseudo-syllables
|
|
1833
|
+
segmentation_name$ = "asyll"
|
|
1834
|
+
elsif (method == segm_voiced) ; voiced portions
|
|
1835
|
+
segmentation_name$ = "voiced"
|
|
1836
|
+
elsif (method == segm_specsim) ;
|
|
1837
|
+
segmentation_name$ = "specsim"
|
|
1838
|
+
elsif (method == segm_pitchchange) ; coding not completed
|
|
1839
|
+
segmentation_name$ = "pitchchange"
|
|
1840
|
+
elsif (method == segm_pitchterrace) ; coding not completed
|
|
1841
|
+
segmentation_name$ = "pitchterrace"
|
|
1842
|
+
else
|
|
1843
|
+
segmentation_name$ = "unknown"
|
|
1844
|
+
endif
|
|
1845
|
+
@debug_msg: "get_segmentation: method='method' name='segmentation_name$'"
|
|
1846
|
+
|
|
1847
|
+
if (task == task_interactive)
|
|
1848
|
+
@read_nuclei_file_interactive_mode
|
|
1849
|
+
;@gr_printline: "segmentation_available='segmentation_available' reuse_nucl='reuse_nucl' nr='nrof_nuclei_analysed'"
|
|
1850
|
+
endif ; task_interactive
|
|
1851
|
+
|
|
1852
|
+
if (not segmentation_available)
|
|
1853
|
+
@debug_msg: "get_segmentation: segmentation not available"
|
|
1854
|
+
# Make segmentation into nuclei depending upon segmentation method
|
|
1855
|
+
# tiers in annotation TextGrid file
|
|
1856
|
+
phone_tier_in = 1 ; default location
|
|
1857
|
+
# create TextGrid nucleiID
|
|
1858
|
+
phone_tier = 1
|
|
1859
|
+
dip_tier = 2
|
|
1860
|
+
nucleus_tier = 3
|
|
1861
|
+
syllable_tier = 4
|
|
1862
|
+
vuv_tier = 5
|
|
1863
|
+
discontinuity_tier = 6
|
|
1864
|
+
safe_tier = 7
|
|
1865
|
+
pointer_tier = 8
|
|
1866
|
+
speaker_tier = 9
|
|
1867
|
+
creak_tier = 10
|
|
1868
|
+
settings_tier = 11
|
|
1869
|
+
hesitation_tier = 12
|
|
1870
|
+
polytonia_tier = 0 ; may be added later for polytonia_annotation
|
|
1871
|
+
polytonia2_tier = 0 ; may be added later for polytonia_annotation
|
|
1872
|
+
boundary_tier = 0 ; may be added later for automatic boundary detection
|
|
1873
|
+
stress_tier = 0 ; may be added later for stress_annotation
|
|
1874
|
+
.n = 12
|
|
1875
|
+
.tiers$ = "phone dip nucleus syll vuv discont safe pointer speaker creak settings hesitation"
|
|
1876
|
+
.point_tiers$ = "dip discont"
|
|
1877
|
+
if (boundary_annotation)
|
|
1878
|
+
.tiers$ += " boundary-auto"
|
|
1879
|
+
.point_tiers$ += " boundary-auto"
|
|
1880
|
+
.n += 1
|
|
1881
|
+
boundary_tier = .n
|
|
1882
|
+
endif
|
|
1883
|
+
if (polytonia_annotation)
|
|
1884
|
+
.tiers$ += " polytonia polytonia-iw"
|
|
1885
|
+
polytonia_tier = .n+1
|
|
1886
|
+
polytonia2_tier = .n+2
|
|
1887
|
+
.n += 2
|
|
1888
|
+
endif
|
|
1889
|
+
if (stress_annotation)
|
|
1890
|
+
.tiers$ += " stress"
|
|
1891
|
+
.n += 1
|
|
1892
|
+
stress_tier = .n
|
|
1893
|
+
endif
|
|
1894
|
+
nucleiID = Create TextGrid: signal_start, signal_finish, .tiers$, .point_tiers$
|
|
1895
|
+
Rename: "nucl"
|
|
1896
|
+
nuclei_available = 1
|
|
1897
|
+
|
|
1898
|
+
# copy VUV decision from vuvgrid to nuclei grid
|
|
1899
|
+
@copy_tier: vuvgridID, 1, nucleiID, vuv_tier
|
|
1900
|
+
removeObject: vuvgridID
|
|
1901
|
+
|
|
1902
|
+
|
|
1903
|
+
if (segfile_available) ; and not internal data format available
|
|
1904
|
+
@debug_msg: "get_segmentation: segfile available"
|
|
1905
|
+
@gr_printline: "Copying annotation tiers..."
|
|
1906
|
+
@copy_tiers_from_annotation_tg
|
|
1907
|
+
endif
|
|
1908
|
+
|
|
1909
|
+
if (not speaker_available)
|
|
1910
|
+
if (fileReadable (speakerfile$))
|
|
1911
|
+
@msg: "Reading speaker information from 'speakerfile$'"
|
|
1912
|
+
tmpID = Read from file: speakerfile$
|
|
1913
|
+
@tier_number_by_name: tmpID, "^[Ss]peaker$"
|
|
1914
|
+
if (result > 0)
|
|
1915
|
+
@copy_tier: tmpID, result, nucleiID, speaker_tier
|
|
1916
|
+
speaker_available = 1
|
|
1917
|
+
else
|
|
1918
|
+
@msg: "No speaker tier found in 'speakerfile$'"
|
|
1919
|
+
endif
|
|
1920
|
+
removeObject: tmpID
|
|
1921
|
+
endif
|
|
1922
|
+
endif
|
|
1923
|
+
|
|
1924
|
+
if (not creak_available)
|
|
1925
|
+
if (fileReadable (creakfile$))
|
|
1926
|
+
@msg: "Reading 'creakfile$'"
|
|
1927
|
+
tmpID = Read from file: creakfile$
|
|
1928
|
+
@tier_number_by_name: tmpID, "^[Cc]reaky?$"
|
|
1929
|
+
if (result > 0)
|
|
1930
|
+
@msg: "Reading creak information from tier 'result' of 'creakfile$'"
|
|
1931
|
+
@copy_tier: tmpID, result, nucleiID, creak_tier
|
|
1932
|
+
creak_available = 1
|
|
1933
|
+
endif
|
|
1934
|
+
removeObject: tmpID
|
|
1935
|
+
endif
|
|
1936
|
+
endif
|
|
1937
|
+
|
|
1938
|
+
if (corpus$ = "cprom" or corpus$ = "rhapsodie")
|
|
1939
|
+
@tier_number_by_name: segmentationID, "^contour$"
|
|
1940
|
+
if (result)
|
|
1941
|
+
@grid_append_tier: segmentationID, result, "nucleiID"
|
|
1942
|
+
endif
|
|
1943
|
+
@tier_number_by_name: segmentationID, "^prom$"
|
|
1944
|
+
if (result)
|
|
1945
|
+
@grid_append_tier: segmentationID, result, "nucleiID"
|
|
1946
|
+
endif
|
|
1947
|
+
endif
|
|
1948
|
+
|
|
1949
|
+
if (needs_phon_tier and not phones_available)
|
|
1950
|
+
@fatal_error: "Cannot find phoneme tier (named phon) in annotation TextGrid"
|
|
1951
|
+
endif
|
|
1952
|
+
|
|
1953
|
+
if (needs_syll_tier and not syllables_available)
|
|
1954
|
+
@fatal_error: "Cannot find syllable tier (named syll) in annotation TextGrid"
|
|
1955
|
+
endif
|
|
1956
|
+
|
|
1957
|
+
@msg: "Calculating actual segmentation. Method='segmentation_name$'..."
|
|
1958
|
+
if (segm_type != segm_extern)
|
|
1959
|
+
@make_segmentation: segm_type, t1s, t2s, nucleiID, mindiff
|
|
1960
|
+
else ; copy consulted external segmentation in tier "segm" of file to nucleus tier of object nucleiID
|
|
1961
|
+
@tier_get: segmentationID, "^segm$", "tier_in", "Cannot find segmentation tier (named segm) in segmentation TextGrid", 1
|
|
1962
|
+
@copy_tier: segmentationID, tier_in, nucleiID, nucleus_tier
|
|
1963
|
+
selectObject: nucleiID
|
|
1964
|
+
.n = Get number of intervals: nucleus_tier
|
|
1965
|
+
for .j from 1 to .n
|
|
1966
|
+
label$ = Get label of interval: nucleus_tier, .j
|
|
1967
|
+
@is_vowel: label$
|
|
1968
|
+
if (is_vowel)
|
|
1969
|
+
Set interval text: nucleus_tier, .j, "a"
|
|
1970
|
+
endif
|
|
1971
|
+
endfor
|
|
1972
|
+
endif
|
|
1973
|
+
|
|
1974
|
+
@safe_nuclei: t1s, t2s
|
|
1975
|
+
nrof_nuclei_analysed = result
|
|
1976
|
+
|
|
1977
|
+
# Store analysis, segmentation and stylization settings in nucleiID TextGrid
|
|
1978
|
+
selectObject: nucleiID
|
|
1979
|
+
Set interval text: settings_tier, 1, "File='basename$' SEG='segmentation_name$' t1='t1s' t2='t2s' " +
|
|
1980
|
+
... "GT='glissando' GT_low='glissando_low' ADAPT='adaptive_glissando' DG='diffgt' MINTS='mindur_ts' MINPAUSE='mindur_pause_gap'"
|
|
1981
|
+
|
|
1982
|
+
|
|
1983
|
+
if (hesitation_available == 0)
|
|
1984
|
+
.tier = nucleus_tier
|
|
1985
|
+
if (syllables_available)
|
|
1986
|
+
.tier = syllable_tier
|
|
1987
|
+
endif
|
|
1988
|
+
@copy_tier: nucleiID, .tier, nucleiID, hesitation_tier
|
|
1989
|
+
@tier_clear_text: nucleiID, hesitation_tier
|
|
1990
|
+
hesitation_available = 1
|
|
1991
|
+
endif
|
|
1992
|
+
if (stress_annotation)
|
|
1993
|
+
if (segm_type == segm_msyllvow or segm_type == segm_msyllpeak or segm_type == segm_mrhyme)
|
|
1994
|
+
@tier_replace2: nucleiID, syllable_tier, nucleiID, stress_tier
|
|
1995
|
+
elsif (segm_type == segm_vnucl and segfile_available and syllables_available)
|
|
1996
|
+
@tier_replace2: nucleiID, syllable_tier, nucleiID, stress_tier
|
|
1997
|
+
else
|
|
1998
|
+
@tier_replace2: nucleiID, nucleus_tier, nucleiID, stress_tier
|
|
1999
|
+
endif
|
|
2000
|
+
@tier_clear_text: nucleiID, stress_tier
|
|
2001
|
+
endif
|
|
2002
|
+
endif ; not segmentation available
|
|
2003
|
+
@debug_msg: "get_segmentation: exit"
|
|
2004
|
+
endproc
|
|
2005
|
+
|
|
2006
|
+
|
|
2007
|
+
procedure read_nuclei_file_interactive_mode
|
|
2008
|
+
; Read settings and tiers from saved nuclei file into nucleiID TextGrid:
|
|
2009
|
+
if (task == task_interactive)
|
|
2010
|
+
if (not fileReadable (nuclfile$))
|
|
2011
|
+
@gr_printline: "Cannot find internal data TextGrid file <'nuclfile$'>"
|
|
2012
|
+
else
|
|
2013
|
+
@gr_printline: "Loading internal data TextGrid from file <'nuclfile$'>"
|
|
2014
|
+
tmpID = Read from file: nuclfile$
|
|
2015
|
+
@tier_number_by_name: tmpID, "^settings$"
|
|
2016
|
+
if (result)
|
|
2017
|
+
@debug_msg: "get_segmentation: interactive mode, found settings"
|
|
2018
|
+
settings_tier = result
|
|
2019
|
+
s$ = Get label of interval: settings_tier, 1
|
|
2020
|
+
method_$ = extractWord$ (s$, "SEG=")
|
|
2021
|
+
t1_ = extractNumber (s$, "t1=")
|
|
2022
|
+
t2_ = extractNumber (s$, "t2=")
|
|
2023
|
+
gt_ = extractNumber (s$, "GT=")
|
|
2024
|
+
gt_low_ = extractNumber (s$, "GT_low=")
|
|
2025
|
+
if (gt_low_ == undefined)
|
|
2026
|
+
gt_low_ = gt_
|
|
2027
|
+
endif
|
|
2028
|
+
.adapt = extractNumber (s$, "ADAPT=")
|
|
2029
|
+
if (.adapt == undefined)
|
|
2030
|
+
.adapt = 0
|
|
2031
|
+
endif
|
|
2032
|
+
dg_ = extractNumber (s$, "DG=")
|
|
2033
|
+
mints_ = extractNumber (s$, "MINTS=")
|
|
2034
|
+
if (anal_t1 >= t1_ and abs(anal_t2 - t2_) < time_step) ; Correct for rounding errors in 32 vs. 64 bit
|
|
2035
|
+
segmentation_name$ = method_$
|
|
2036
|
+
@tier_number_by_name: tmpID, "^phone$"
|
|
2037
|
+
phone_tier = result
|
|
2038
|
+
if (result)
|
|
2039
|
+
@tier_number_by_name: tmpID, "^syll$"
|
|
2040
|
+
syllable_tier = result
|
|
2041
|
+
endif
|
|
2042
|
+
if (result)
|
|
2043
|
+
@tier_number_by_name: tmpID, "^nucleus$"
|
|
2044
|
+
nucleus_tier = result
|
|
2045
|
+
endif
|
|
2046
|
+
if (result)
|
|
2047
|
+
@tier_number_by_name: tmpID, "^vuv$"
|
|
2048
|
+
vuv_tier = result
|
|
2049
|
+
removeObject: vuvgridID
|
|
2050
|
+
endif
|
|
2051
|
+
if (result)
|
|
2052
|
+
@tier_number_by_name: tmpID, "^discont$"
|
|
2053
|
+
discontinuity_tier = result
|
|
2054
|
+
endif
|
|
2055
|
+
if (result)
|
|
2056
|
+
@tier_number_by_name: tmpID, "^safe$"
|
|
2057
|
+
safe_tier = result
|
|
2058
|
+
endif
|
|
2059
|
+
if (result)
|
|
2060
|
+
@tier_number_by_name: tmpID, "^pointer$"
|
|
2061
|
+
pointer_tier = result
|
|
2062
|
+
endif
|
|
2063
|
+
if (result)
|
|
2064
|
+
@tier_number_by_name: tmpID, "^speaker$"
|
|
2065
|
+
speaker_tier = result
|
|
2066
|
+
if (speaker_tier)
|
|
2067
|
+
speaker_available = 1 ; speaker tier available
|
|
2068
|
+
endif
|
|
2069
|
+
endif
|
|
2070
|
+
if (result) ; success
|
|
2071
|
+
reuse_nucl = 1
|
|
2072
|
+
segmentation_available = 1
|
|
2073
|
+
nucleiID = tmpID
|
|
2074
|
+
nuclei_available = 1
|
|
2075
|
+
; nrof_nuclei_analysed = Count labels: safe_tier, "a"
|
|
2076
|
+
nrof_nuclei_analysed = Count labels: nucleus_tier, "a"
|
|
2077
|
+
@gr_printline: "Loaded internal data TextGrid from file <'nuclfile$'>"
|
|
2078
|
+
endif
|
|
2079
|
+
endif ; if anal_t1
|
|
2080
|
+
endif ; tier with settings found
|
|
2081
|
+
if (not segmentation_available) ; textgrid read but not used
|
|
2082
|
+
removeObject: tmpID
|
|
2083
|
+
endif
|
|
2084
|
+
endif ; nuclfile found
|
|
2085
|
+
endif
|
|
2086
|
+
endproc
|
|
2087
|
+
|
|
2088
|
+
|
|
2089
|
+
procedure copy_tiers_from_annotation_tg
|
|
2090
|
+
; Copy tiers (phon, syll, speaker, words, creak, hes) from segmentation TextGrid to nucleiID
|
|
2091
|
+
@tier_number_by_name: segmentationID, "^phon"
|
|
2092
|
+
# The following if-block may override default name of phoneme tier
|
|
2093
|
+
if (variableExists ("corpus$") and variableExists ("corpus_tier_phon$"))
|
|
2094
|
+
if (length (corpus_tier_phon$))
|
|
2095
|
+
@tier_number_by_name: segmentationID, "^'corpus_tier_phon$'$"
|
|
2096
|
+
if (result = 0)
|
|
2097
|
+
@msg: "Cannot find phoneme tier named <'corpus_tier_phon$'>"
|
|
2098
|
+
endif
|
|
2099
|
+
endif
|
|
2100
|
+
endif
|
|
2101
|
+
if (result) ; found phoneme tier
|
|
2102
|
+
phone_tier_in = result
|
|
2103
|
+
@msg: "Using phonetic alignment from tier named <'result2$'>"
|
|
2104
|
+
phones_available = 1
|
|
2105
|
+
else
|
|
2106
|
+
phone_tier_in = 0
|
|
2107
|
+
phones_available = 0
|
|
2108
|
+
@msg: "Cannot find tier with phonetic alignment"
|
|
2109
|
+
endif
|
|
2110
|
+
if (phones_available)
|
|
2111
|
+
@tier_replace: segmentationID, phone_tier_in, nucleiID, phone_tier, "nucleiID"
|
|
2112
|
+
selectObject: nucleiID
|
|
2113
|
+
Set tier name: phone_tier, "phone"
|
|
2114
|
+
endif
|
|
2115
|
+
|
|
2116
|
+
@tier_number_by_name: segmentationID, "^syll"
|
|
2117
|
+
# The following if-block may override default name of syllable tier
|
|
2118
|
+
if (variableExists ("corpus$") and variableExists ("corpus_tier_syll$"))
|
|
2119
|
+
if (length (corpus_tier_syll$))
|
|
2120
|
+
@tier_number_by_name: segmentationID, "^'corpus_tier_syll$'$"
|
|
2121
|
+
endif
|
|
2122
|
+
endif
|
|
2123
|
+
if (result and segm_type <> segm_asyll)
|
|
2124
|
+
; In segm_asyll, the syllable tier is used for pseudo-syllables and should not be used for true syllables.
|
|
2125
|
+
syll_tier_in = result
|
|
2126
|
+
@msg: "Using syllable alignment from tier named <'result2$'>"
|
|
2127
|
+
@copy_tier: segmentationID, syll_tier_in, nucleiID, syllable_tier
|
|
2128
|
+
; call tier_replace segmentationID syll_tier_in nucleiID syllable_tier nucleiID
|
|
2129
|
+
; selectObject: nucleiID
|
|
2130
|
+
; Set tier name... syllable_tier syll
|
|
2131
|
+
syllables_available = 1
|
|
2132
|
+
endif
|
|
2133
|
+
|
|
2134
|
+
@tier_number_by_name: segmentationID, "^[Ss]peaker$"
|
|
2135
|
+
# The following if-block may override default name of speaker tier
|
|
2136
|
+
if (variableExists ("corpus$") and variableExists ("corpus_tier_speaker$"))
|
|
2137
|
+
if (length (corpus_tier_speaker$))
|
|
2138
|
+
@tier_number_by_name: segmentationID, "^'corpus_tier_speaker$'$"
|
|
2139
|
+
endif
|
|
2140
|
+
endif
|
|
2141
|
+
if (result)
|
|
2142
|
+
speaker_tier_in = result
|
|
2143
|
+
@msg: "Using speaker information from tier named <'result2$'>"
|
|
2144
|
+
if (corpus$ = "rhapsodie")
|
|
2145
|
+
@copy_tier_tweeked: segmentationID, speaker_tier_in, nucleiID, speaker_tier
|
|
2146
|
+
else
|
|
2147
|
+
@copy_tier: segmentationID, speaker_tier_in, nucleiID, speaker_tier
|
|
2148
|
+
endif
|
|
2149
|
+
speaker_available = 1 ; speaker tier available
|
|
2150
|
+
else ; input segmentation textgrid does not contain speaker tier
|
|
2151
|
+
selectObject: nucleiID
|
|
2152
|
+
Set interval text: speaker_tier, 1, "ANON"
|
|
2153
|
+
endif
|
|
2154
|
+
|
|
2155
|
+
@tier_number_by_name: segmentationID, "^[Cc]reaky?$"
|
|
2156
|
+
if (result)
|
|
2157
|
+
@msg: "Using creak information from tier named <'result2$'> of <'segfile$'>"
|
|
2158
|
+
@copy_tier: segmentationID, result, nucleiID, creak_tier
|
|
2159
|
+
creak_available = 1
|
|
2160
|
+
endif
|
|
2161
|
+
|
|
2162
|
+
word_tier_in = 0
|
|
2163
|
+
@tier_number_by_name: segmentationID, "^words?$"
|
|
2164
|
+
# The following if-block may override default name of word tier
|
|
2165
|
+
if (variableExists ("corpus$") and variableExists ("corpus_tier_word$"))
|
|
2166
|
+
if (length (corpus_tier_word$))
|
|
2167
|
+
@tier_number_by_name: segmentationID, "^'corpus_tier_word$'$"
|
|
2168
|
+
endif
|
|
2169
|
+
endif
|
|
2170
|
+
if (result)
|
|
2171
|
+
word_tier_in = result
|
|
2172
|
+
endif
|
|
2173
|
+
|
|
2174
|
+
if (segm_type <> segm_asyll) ; syllable_tier will be used for storing asyll intervals
|
|
2175
|
+
@tier_number_by_name: segmentationID, "^hes$"
|
|
2176
|
+
if (result)
|
|
2177
|
+
hesitation_tier_in = result
|
|
2178
|
+
@msg: "Using hesitation information from tier named <'result2$'> of <'segfile$'>"
|
|
2179
|
+
@copy_tier: segmentationID, hesitation_tier_in, nucleiID, hesitation_tier
|
|
2180
|
+
hesitation_available = 1
|
|
2181
|
+
endif
|
|
2182
|
+
endif
|
|
2183
|
+
endproc
|
|
2184
|
+
|
|
2185
|
+
|
|
2186
|
+
|
|
2187
|
+
procedure clipPitchTier: objectID, .ymin, .ymax, .xmin, .xmax
|
|
2188
|
+
# Clip a PitchTier by replacing data points outside clip range by values 1 ST outside range
|
|
2189
|
+
@debug_msg: "clipPitchTier: entry"
|
|
2190
|
+
selectObject: objectID
|
|
2191
|
+
.firsti = Get low index from time: .xmin
|
|
2192
|
+
if (.firsti = 0)
|
|
2193
|
+
.firsti = 1
|
|
2194
|
+
endif
|
|
2195
|
+
.lasti = Get nearest index from time: .xmax
|
|
2196
|
+
for .i from .firsti to .lasti
|
|
2197
|
+
.x = Get time from index: .i
|
|
2198
|
+
.y = Get value at index: .i
|
|
2199
|
+
.y = min(.y,.ymax+1)
|
|
2200
|
+
.y = max(.y,.ymin-1)
|
|
2201
|
+
endif
|
|
2202
|
+
Remove point: .i
|
|
2203
|
+
Add point: .x, .y
|
|
2204
|
+
endfor
|
|
2205
|
+
@debug_msg: "clipPitchTier: exit"
|
|
2206
|
+
endproc
|
|
2207
|
+
|
|
2208
|
+
|
|
2209
|
+
procedure cleanup_current_file
|
|
2210
|
+
# Delete temporary objects created for current input file
|
|
2211
|
+
@debug_msg: "cleanup_current_file: entry"
|
|
2212
|
+
if (do_cleanup)
|
|
2213
|
+
if (signal_available)
|
|
2214
|
+
removeObject: soundID
|
|
2215
|
+
if (volatile)
|
|
2216
|
+
removeObject: fullsoundID
|
|
2217
|
+
endif
|
|
2218
|
+
signal_available = 0
|
|
2219
|
+
endif
|
|
2220
|
+
if (intensity_available)
|
|
2221
|
+
removeObject: intensityID
|
|
2222
|
+
endif
|
|
2223
|
+
if (pitch_available)
|
|
2224
|
+
removeObject: pitchID
|
|
2225
|
+
endif
|
|
2226
|
+
if (loudness_available)
|
|
2227
|
+
removeObject: loudnessID
|
|
2228
|
+
endif
|
|
2229
|
+
if (intbp_available)
|
|
2230
|
+
removeObject: intbpID
|
|
2231
|
+
endif
|
|
2232
|
+
if (inthp_available)
|
|
2233
|
+
removeObject: inthpID
|
|
2234
|
+
endif
|
|
2235
|
+
if (segfile_available)
|
|
2236
|
+
if not boundary_annotation
|
|
2237
|
+
removeObject: segmentationID
|
|
2238
|
+
endif
|
|
2239
|
+
endif
|
|
2240
|
+
if (draw_prosograms)
|
|
2241
|
+
if (variableExists("newgridID"))
|
|
2242
|
+
; if (segfile_available and nrofplottedtiers > 0) or (polytonia_annotation)
|
|
2243
|
+
removeObject: newgridID
|
|
2244
|
+
endif
|
|
2245
|
+
endif
|
|
2246
|
+
if (nuclei_available)
|
|
2247
|
+
removeObject: nucleiID
|
|
2248
|
+
endif
|
|
2249
|
+
if (needs_stylization)
|
|
2250
|
+
if (nucldat_available)
|
|
2251
|
+
removeObject: nucldatID
|
|
2252
|
+
endif
|
|
2253
|
+
if (stylization_available)
|
|
2254
|
+
removeObject: stylID, stylSTID
|
|
2255
|
+
endif
|
|
2256
|
+
endif
|
|
2257
|
+
if (profile_available)
|
|
2258
|
+
removeObject: profileID
|
|
2259
|
+
endif
|
|
2260
|
+
endif ; if (do_cleanup)
|
|
2261
|
+
@debug_msg: "cleanup_current_file: exit"
|
|
2262
|
+
endproc
|
|
2263
|
+
|
|
2264
|
+
|
|
2265
|
+
procedure cleanup_global
|
|
2266
|
+
@debug_msg: "cleanup_global: entry"
|
|
2267
|
+
if (do_cleanup)
|
|
2268
|
+
if (globalsheet_available)
|
|
2269
|
+
removeObject: globalprofileID
|
|
2270
|
+
endif
|
|
2271
|
+
if ((not batch_mode) and variableExists ("filelistID"))
|
|
2272
|
+
removeObject: filelistID
|
|
2273
|
+
endif
|
|
2274
|
+
endif ; if (do_cleanup)
|
|
2275
|
+
@debug_msg: "cleanup_global: exit"
|
|
2276
|
+
endproc
|
|
2277
|
+
|
|
2278
|
+
|
|
2279
|
+
procedure convert_sampa_ipa: .objectID, .tier
|
|
2280
|
+
selectObject: .objectID
|
|
2281
|
+
.n = Get number of intervals: .tier
|
|
2282
|
+
for .j from 1 to .n
|
|
2283
|
+
label$ = Get label of interval: .tier, .j
|
|
2284
|
+
@sampa_ipa: label$
|
|
2285
|
+
Set interval text: .tier, .j, label$
|
|
2286
|
+
endfor
|
|
2287
|
+
endproc
|
|
2288
|
+
|
|
2289
|
+
|
|
2290
|
+
procedure sampa_ipa: label$
|
|
2291
|
+
# replace SAMPA label$ by IPA representation in Praat's "special symbols" format
|
|
2292
|
+
.len = length (label$)
|
|
2293
|
+
s$ = "" ; resulting string
|
|
2294
|
+
.i = 1
|
|
2295
|
+
while (.i <= .len)
|
|
2296
|
+
b$ = "" ; translation (i.e. output) of current SAMPA symbol
|
|
2297
|
+
c1$ = mid$ (label$,.i,1) ; next char
|
|
2298
|
+
c2$ = mid$ (label$,.i,2) ; next 2 chars
|
|
2299
|
+
c3$ = mid$ (label$,.i,3) ; next 3 chars
|
|
2300
|
+
c4$ = mid$ (label$,.i,4) ; next 4 chars
|
|
2301
|
+
cn$ = mid$ (label$,.i+1,1) ; following char
|
|
2302
|
+
.restlen = .len - .i
|
|
2303
|
+
if (.restlen >= 1 and (cn$ = "~" or cn$ = "`" or cn$ = "="))
|
|
2304
|
+
if (cn$ = "~") ; nasal vowels (length=2)
|
|
2305
|
+
if (c2$ = "a~" or c2$ = "A~")
|
|
2306
|
+
b$ = "\as\~^"
|
|
2307
|
+
elsif (c2$ = "o~" or c2$ = "O~")
|
|
2308
|
+
b$ = "\ct\~^"
|
|
2309
|
+
elsif (c2$ = "e~" or c2$ = "E~")
|
|
2310
|
+
b$ = "\ef\~^"
|
|
2311
|
+
elsif (c2$ = "9~")
|
|
2312
|
+
b$ = "\oe\~^"
|
|
2313
|
+
endif
|
|
2314
|
+
elsif (cn$ = "`") ; rhoticity
|
|
2315
|
+
if (index ("uoi", c1$))
|
|
2316
|
+
b$ = c1$ + "\hr"
|
|
2317
|
+
elsif (c2$ = "@`")
|
|
2318
|
+
b$ = "\sr"
|
|
2319
|
+
elsif (c2$ = "s`")
|
|
2320
|
+
b$ = "\s."
|
|
2321
|
+
elsif (c2$ = "z`")
|
|
2322
|
+
b$ = "\z."
|
|
2323
|
+
endif
|
|
2324
|
+
elsif (cn$ = "=") ; syllabicity in x-sampa
|
|
2325
|
+
if (index_regex (c1$, "[mnNJlrR]"))
|
|
2326
|
+
b$ = c1$ + "\|v"
|
|
2327
|
+
endif
|
|
2328
|
+
endif
|
|
2329
|
+
if (length (b$) > 0)
|
|
2330
|
+
s$ = s$ + b$
|
|
2331
|
+
.i += 2
|
|
2332
|
+
else
|
|
2333
|
+
s$ = s$ + c1$
|
|
2334
|
+
.i += 1
|
|
2335
|
+
endif
|
|
2336
|
+
# SAMPA symbols of length 1
|
|
2337
|
+
elsif (index ("BCDSTZRMNHJGAEIOQVY@2679?:{}&", c1$) or c1$ = """") ; additional regex to speed up things
|
|
2338
|
+
if (c1$ = """") ; primary stress
|
|
2339
|
+
b$ = "\'1"
|
|
2340
|
+
elsif (c1$ = "A")
|
|
2341
|
+
b$ = "\as"
|
|
2342
|
+
elsif (c1$ = "E")
|
|
2343
|
+
; b$ = "\ep"
|
|
2344
|
+
b$ = "\ef"
|
|
2345
|
+
elsif (c1$ = "I")
|
|
2346
|
+
b$ = "\ic"
|
|
2347
|
+
elsif (c1$ = "O")
|
|
2348
|
+
b$ = "\ct"
|
|
2349
|
+
elsif (c1$ = "Y")
|
|
2350
|
+
b$ = "\yc"
|
|
2351
|
+
elsif (c1$ = "Q")
|
|
2352
|
+
b$ = "\ab"
|
|
2353
|
+
elsif (c1$ = "V")
|
|
2354
|
+
b$ = "\vt"
|
|
2355
|
+
elsif (c1$ = "2")
|
|
2356
|
+
b$ = "\o/"
|
|
2357
|
+
elsif (c1$ = "6")
|
|
2358
|
+
b$ = "\sr"
|
|
2359
|
+
elsif (c1$ = "7")
|
|
2360
|
+
b$ = "\rh"
|
|
2361
|
+
elsif (c1$ = "9")
|
|
2362
|
+
b$ = "\oe"
|
|
2363
|
+
elsif (c1$ = "@") ; schwa
|
|
2364
|
+
b$ = "\sw"
|
|
2365
|
+
elsif (c1$ = "B") ; voiced bilabial fricative
|
|
2366
|
+
b$ = "\bf"
|
|
2367
|
+
elsif (c1$ = "C") ; voiceless palatal fricative
|
|
2368
|
+
b$ = "\c,"
|
|
2369
|
+
elsif (c1$ = "D") ; voiced dental fricative
|
|
2370
|
+
b$ = "\dh"
|
|
2371
|
+
elsif (c1$ = "T") ; voiceless dental fricative
|
|
2372
|
+
b$ = "\tf"
|
|
2373
|
+
elsif (c1$ = "S")
|
|
2374
|
+
b$ = "\sh"
|
|
2375
|
+
elsif (c1$ = "Z")
|
|
2376
|
+
b$ = "\zh"
|
|
2377
|
+
elsif (c1$ = "R")
|
|
2378
|
+
b$ = "\rc"
|
|
2379
|
+
elsif (c1$ = "M")
|
|
2380
|
+
b$ = "\mj"
|
|
2381
|
+
elsif (c1$ = "N")
|
|
2382
|
+
b$ = "\ng"
|
|
2383
|
+
elsif (c1$ = "H")
|
|
2384
|
+
b$ = "\ht"
|
|
2385
|
+
elsif (c1$ = "J")
|
|
2386
|
+
b$ = "\nj"
|
|
2387
|
+
elsif (c1$ = "G")
|
|
2388
|
+
b$ = "\gf"
|
|
2389
|
+
elsif (c1$ = "?")
|
|
2390
|
+
b$ = "\?g"
|
|
2391
|
+
elsif (c1$ = ":")
|
|
2392
|
+
b$ = "\:f"
|
|
2393
|
+
elsif (c1$ = "{")
|
|
2394
|
+
b$ = "\ae"
|
|
2395
|
+
elsif (c1$ = "}")
|
|
2396
|
+
b$ = "\u-"
|
|
2397
|
+
elsif (c1$ = "&")
|
|
2398
|
+
b$ = "\Oe"
|
|
2399
|
+
endif
|
|
2400
|
+
if (length (b$) > 0)
|
|
2401
|
+
s$ = s$ + b$ ; append b$ to output string
|
|
2402
|
+
.i += 1
|
|
2403
|
+
else
|
|
2404
|
+
s$ = s$ + c1$
|
|
2405
|
+
.i += 1
|
|
2406
|
+
endif
|
|
2407
|
+
elsif (c1$ = "\")
|
|
2408
|
+
if (index ("_\:f__\a~__\o~__\o/__\ef_", "_'c3$'_"))
|
|
2409
|
+
s$ = s$ + c3$
|
|
2410
|
+
.i += 3
|
|
2411
|
+
; elsif (.restlen >= 3 and index (":\a~:\o~:\o/:", ":'c3$':") > 0)
|
|
2412
|
+
; s$ = s$ + c3$
|
|
2413
|
+
; .i += 3
|
|
2414
|
+
elsif (.restlen >= 4 and index (":\ep~:", c4$) > 0)
|
|
2415
|
+
s$ = s$ + c4$
|
|
2416
|
+
.i += 4
|
|
2417
|
+
else ; others: just copy them
|
|
2418
|
+
s$ = s$ + c1$
|
|
2419
|
+
.i += 1
|
|
2420
|
+
endif
|
|
2421
|
+
elsif (c3$ = "i_d")
|
|
2422
|
+
s$ = s$ + "i\Nv"
|
|
2423
|
+
.i += 3
|
|
2424
|
+
else ; others: just copy them
|
|
2425
|
+
s$ = s$ + c1$
|
|
2426
|
+
.i += 1
|
|
2427
|
+
endif
|
|
2428
|
+
endwhile
|
|
2429
|
+
label$ = s$
|
|
2430
|
+
endproc
|
|
2431
|
+
|
|
2432
|
+
|
|
2433
|
+
procedure speaker_info_get
|
|
2434
|
+
# 1. Make list of speakers appearing in tier "speaker" of annotated segmentation textgrid.
|
|
2435
|
+
# 2. Add speaker identification to nucldatID, as a number (in column <speaker_id>) and as a string (in row label).
|
|
2436
|
+
@debug_msg: "speaker_info_get: entry, speaker_available='speaker_available'"
|
|
2437
|
+
speakers = 0 ; nrof different speakers found
|
|
2438
|
+
speakers$ = "" ; used to store speaker names and their number
|
|
2439
|
+
if (speaker_available)
|
|
2440
|
+
speaker = 0 ; nrof current speaker
|
|
2441
|
+
selectObject: nucldatID
|
|
2442
|
+
.rows = Get number of rows
|
|
2443
|
+
for .row to .rows
|
|
2444
|
+
.t1 = Get value: .row, j_nucl_t1
|
|
2445
|
+
.t2 = Get value: .row, j_nucl_t2
|
|
2446
|
+
selectObject: nucleiID
|
|
2447
|
+
.i = Get interval at time: speaker_tier, .t1+(.t2-.t1)/2
|
|
2448
|
+
speaker$ = Get label of interval: speaker_tier, .i
|
|
2449
|
+
speaker$ = replace_regex$ (speaker$, "^ +(.*) +$", "\1", 1) ; trim spaces at left and right
|
|
2450
|
+
if (index (speakers$, "<'speaker$'>") > 0) ; speaker already encountered in file
|
|
2451
|
+
speaker = extractNumber (speakers$, "<'speaker$'>:") ; get his number
|
|
2452
|
+
else ; encountered a new speaker
|
|
2453
|
+
speakers += 1
|
|
2454
|
+
speaker = speakers ; number of speaker
|
|
2455
|
+
speakers$ = speakers$ + "<'speaker$'>:'speaker' " ; store name and number in list
|
|
2456
|
+
speaker_label'speaker'$ = speaker$
|
|
2457
|
+
endif
|
|
2458
|
+
selectObject: nucldatID
|
|
2459
|
+
Set value: .row, j_speaker_id, speaker
|
|
2460
|
+
Set row label (index): .row, speaker$
|
|
2461
|
+
endfor
|
|
2462
|
+
else
|
|
2463
|
+
@msg: "No speaker tier in input textgrid. Assuming 1 speaker."
|
|
2464
|
+
speakers = 1
|
|
2465
|
+
speaker$ = "ANON"
|
|
2466
|
+
speaker_label1$ = speaker$
|
|
2467
|
+
speakers$ = "<'speaker$'>:'speakers' "
|
|
2468
|
+
selectObject: nucleiID
|
|
2469
|
+
Set interval text: speaker_tier, 1, speaker$
|
|
2470
|
+
selectObject: nucldatID
|
|
2471
|
+
.rows = Get number of rows
|
|
2472
|
+
for .row from 1 to .rows
|
|
2473
|
+
Set value: .row, j_speaker_id, 1
|
|
2474
|
+
endfor
|
|
2475
|
+
endif
|
|
2476
|
+
|
|
2477
|
+
; Correct values at speaker turns, for pause duration, internucleus duration
|
|
2478
|
+
selectObject: nucldatID
|
|
2479
|
+
.rows = Get number of rows
|
|
2480
|
+
for .row to .rows
|
|
2481
|
+
if (.row > 1)
|
|
2482
|
+
selectObject: nucldatID
|
|
2483
|
+
.sp = Get value: .row, j_speaker_id
|
|
2484
|
+
.prevsp = Get value: .row-1, j_speaker_id
|
|
2485
|
+
if (.sp <> .prevsp)
|
|
2486
|
+
Set value: .row, j_intersyl, 0
|
|
2487
|
+
Set value: .row, j_internucldur, 0
|
|
2488
|
+
.t1 = Get value: .row-1, j_nucl_t1
|
|
2489
|
+
.t2 = Get value: .row-1, j_nucl_t2
|
|
2490
|
+
; pause ends and speaker turn, even when next turn starts with silence
|
|
2491
|
+
selectObject: nucleiID
|
|
2492
|
+
.i = Get interval at time: speaker_tier, .t1
|
|
2493
|
+
.t2s = Get end time of interval: speaker_tier, .i
|
|
2494
|
+
selectObject: nucldatID
|
|
2495
|
+
if (.t2s-.t2 >= mindur_pause_gap)
|
|
2496
|
+
Set value: .row-1, j_pause_dur, .t2s-.t2
|
|
2497
|
+
else
|
|
2498
|
+
Set value: .row-1, j_pause_dur, 0
|
|
2499
|
+
endif
|
|
2500
|
+
endif
|
|
2501
|
+
endif
|
|
2502
|
+
endfor
|
|
2503
|
+
@debug_msg: "speaker_info_get: exit"
|
|
2504
|
+
endproc
|
|
2505
|
+
|
|
2506
|
+
|
|
2507
|
+
procedure speaker_autorange: t1, t2
|
|
2508
|
+
; propose a pitch range suitable for plotting the range of all speakers in time interval <t1>..<t2>
|
|
2509
|
+
; return range in ST in <ymin>..<ymax>
|
|
2510
|
+
@debug_msg: "speaker_autorange: entry"
|
|
2511
|
+
if (task == task_interactive and profile_available)
|
|
2512
|
+
speakers$ = ""
|
|
2513
|
+
selectObject: profileID
|
|
2514
|
+
.nrows = Get number of rows
|
|
2515
|
+
for .j to .nrows
|
|
2516
|
+
.s$ = Get row label: .j
|
|
2517
|
+
.s$ = replace_regex$(.s$, "^('basename$'_)", "", 1)
|
|
2518
|
+
speakers$ += "<'.s$'>:'.j' "
|
|
2519
|
+
speaker_label'.j'$ = .s$
|
|
2520
|
+
.topST = Get value: .j, j_pitch_top_ST
|
|
2521
|
+
.bottomST = Get value: .j, j_pitch_bottom_ST
|
|
2522
|
+
.medianST = Get value: .j, j_pitch_median_ST
|
|
2523
|
+
speaker_range_'.j'$ = "TOP_ST='.topST:1' BOTTOM_ST='.bottomST:1' MEDIAN_ST='.medianST:1' "
|
|
2524
|
+
endfor
|
|
2525
|
+
endif
|
|
2526
|
+
if (task == task_pitch_plot or not needs_pitchrange) ; no nuclei available
|
|
2527
|
+
selectObject: pitchID
|
|
2528
|
+
.y = Get mean: t1, t2, "semitones re 1 Hz"
|
|
2529
|
+
if (.y == undefined)
|
|
2530
|
+
@msg: "speaker_autorange: mean pitch undefined for time 't1:3'-'t2:3', using default 85 ST"
|
|
2531
|
+
.y = 85
|
|
2532
|
+
endif
|
|
2533
|
+
ymin = .y - 12
|
|
2534
|
+
ymax = .y + 12
|
|
2535
|
+
else ; compute total range for all speakers in time interval <t1>..<t2>
|
|
2536
|
+
ymin = 1000
|
|
2537
|
+
ymax = 0
|
|
2538
|
+
.t = t1
|
|
2539
|
+
selectObject: nucleiID
|
|
2540
|
+
.n = Get number of intervals: speaker_tier
|
|
2541
|
+
.t2n = Get end point: speaker_tier, .n
|
|
2542
|
+
.t2 = min (t2, .t2n)
|
|
2543
|
+
repeat ; all speakers in range t1..t2
|
|
2544
|
+
.i = Get interval at time: speaker_tier, .t
|
|
2545
|
+
speaker$ = Get label of interval: speaker_tier, .i
|
|
2546
|
+
speaker$ = replace_regex$ (speaker$, "^ +(.*) +$", "\1", 1) ; trim left and right
|
|
2547
|
+
speaker_j = extractNumber (speakers$, "<'speaker$'>:") ; get his number
|
|
2548
|
+
if (speaker_j == 0 or speaker_j == undefined)
|
|
2549
|
+
; @msg: "speaker_autorange: speaker undefined at time '.t:3'"
|
|
2550
|
+
; This happens when there are no nuclei for this speaker
|
|
2551
|
+
selectObject: pitchID
|
|
2552
|
+
ymin = Get mean: 0, 0, "semitones re 1 Hz"
|
|
2553
|
+
ymax = ymin
|
|
2554
|
+
else
|
|
2555
|
+
.y = extractNumber (speaker_range_'speaker_j'$, "BOTTOM_ST=")
|
|
2556
|
+
if (.y == undefined)
|
|
2557
|
+
@msg: "speaker_autorange: BOTTOM_ST undefined t='.t:3'"
|
|
2558
|
+
.y = 85
|
|
2559
|
+
endif
|
|
2560
|
+
ymin = min (ymin, .y)
|
|
2561
|
+
.y = extractNumber (speaker_range_'speaker_j'$, "TOP_ST=")
|
|
2562
|
+
if (.y == undefined)
|
|
2563
|
+
@msg: "speaker_autorange: TOP_ST undefined t='.t:3'"
|
|
2564
|
+
.y = 85
|
|
2565
|
+
endif
|
|
2566
|
+
ymax = max (ymax, .y)
|
|
2567
|
+
endif
|
|
2568
|
+
selectObject: nucleiID
|
|
2569
|
+
.t = Get end point: speaker_tier, .i
|
|
2570
|
+
until (.t >= .t2)
|
|
2571
|
+
if (ymax - ymin < 24) ; default range 24 ST or 2 octaves
|
|
2572
|
+
ymin = max(0, (ymin + (ymax - ymin)/2) - 12)
|
|
2573
|
+
ymax = ymin + 24
|
|
2574
|
+
endif
|
|
2575
|
+
endif ; task
|
|
2576
|
+
@debug_msg: "speaker_autorange: exit"
|
|
2577
|
+
endproc
|
|
2578
|
+
|
|
2579
|
+
|
|
2580
|
+
procedure nucleus_tier_postproc: .grid, .nucl_tier, .syll_tier
|
|
2581
|
+
selectObject: .grid
|
|
2582
|
+
interval_tier = Is interval tier: .nucl_tier
|
|
2583
|
+
if (interval_tier)
|
|
2584
|
+
Set tier name: .nucl_tier, "segm"
|
|
2585
|
+
.n = Get number of intervals: .nucl_tier
|
|
2586
|
+
; replace parts other than nucleus by zero-length strings
|
|
2587
|
+
for .i to .n
|
|
2588
|
+
.s$ = Get label of interval: .nucl_tier, .i
|
|
2589
|
+
if (index (":<:>:<>:xL:xR:U:skip:reject:short:", ":'.s$':"))
|
|
2590
|
+
Set interval text: .nucl_tier, .i, ""
|
|
2591
|
+
endif
|
|
2592
|
+
endfor
|
|
2593
|
+
; reduce contiguous empty intervals
|
|
2594
|
+
.i = 2
|
|
2595
|
+
while (.i <= .n)
|
|
2596
|
+
.prev$ = Get label of interval: .nucl_tier, .i-1
|
|
2597
|
+
.s$ = Get label of interval: .nucl_tier, .i
|
|
2598
|
+
if (.prev$ = "" and .s$ = "")
|
|
2599
|
+
Remove left boundary: .nucl_tier, .i
|
|
2600
|
+
.n -= 1
|
|
2601
|
+
else
|
|
2602
|
+
.i += 1
|
|
2603
|
+
endif
|
|
2604
|
+
endwhile
|
|
2605
|
+
; add syllable boundaries to nucleus tier
|
|
2606
|
+
.n = Get number of intervals: .syll_tier
|
|
2607
|
+
for .i to .n - 1
|
|
2608
|
+
.t = Get end point: .syll_tier, .i
|
|
2609
|
+
.j = Get interval at time: .nucl_tier, .t
|
|
2610
|
+
.x1 = Get start point: .nucl_tier, .j
|
|
2611
|
+
.x2 = Get end point: .nucl_tier, .j
|
|
2612
|
+
if (.t - .x1 > time_step and .t < .x2)
|
|
2613
|
+
Insert boundary: .nucl_tier, .t
|
|
2614
|
+
endif
|
|
2615
|
+
Set interval text: .syll_tier, .i, ""
|
|
2616
|
+
endfor
|
|
2617
|
+
endif
|
|
2618
|
+
endproc
|
|
2619
|
+
|
|
2620
|
+
|
|
2621
|
+
procedure copy_tier_tweeked: srcgrid, srctier, destgrid, desttier
|
|
2622
|
+
; Special treatment for Rhapsodie tier "locuteur". Force interval contiguity.
|
|
2623
|
+
selectObject: destgrid
|
|
2624
|
+
.n = Get number of intervals: desttier
|
|
2625
|
+
.endtime = Get end time of interval: desttier, .n
|
|
2626
|
+
selectObject: srcgrid
|
|
2627
|
+
.n = Get number of intervals: srctier
|
|
2628
|
+
for .i to .n
|
|
2629
|
+
selectObject: srcgrid
|
|
2630
|
+
.t1 = Get start time of interval: srctier, .i
|
|
2631
|
+
.t2 = Get end time of interval: srctier, .i
|
|
2632
|
+
.label$ = Get label of interval: srctier, .i
|
|
2633
|
+
if (.i < .n)
|
|
2634
|
+
.t3 = Get start time of interval: srctier, .i+1
|
|
2635
|
+
if (.t2 < .t3)
|
|
2636
|
+
.t2 = .t3 ; intervals should be contiguous
|
|
2637
|
+
endif
|
|
2638
|
+
endif
|
|
2639
|
+
selectObject: destgrid
|
|
2640
|
+
if (.t2 < .endtime and .i < .n)
|
|
2641
|
+
Insert boundary: desttier, .t2
|
|
2642
|
+
endif
|
|
2643
|
+
if (.t1 <= .endtime)
|
|
2644
|
+
.t2 = min (.t2, .endtime) ; when destgrid is shorter than srcgrid
|
|
2645
|
+
j = Get interval at time: desttier, .t1 + (.t2-.t1)/2
|
|
2646
|
+
Set interval text: desttier, j, .label$
|
|
2647
|
+
endif
|
|
2648
|
+
endfor
|
|
2649
|
+
endproc
|
|
2650
|
+
|
|
2651
|
+
|
|
2652
|
+
procedure validate_phoneme_tier: .infname$
|
|
2653
|
+
@fname_parts: .infname$
|
|
2654
|
+
.basename$ = result2$
|
|
2655
|
+
.dataID = Read from file: .infname$
|
|
2656
|
+
@tier_get: .dataID, "^phon", "phone_tier_", "Cannot find phoneme tier in '.basename$'", 1
|
|
2657
|
+
.n = Get number of intervals: phone_tier_
|
|
2658
|
+
for .phon from 1 to .n
|
|
2659
|
+
.t1_phon = Get start time of interval: phone_tier_, .phon
|
|
2660
|
+
.label_phon$ = Get label of interval: phone_tier_, .phon
|
|
2661
|
+
@verify_tier: .label_phon$, .basename$, "phoneme", .t1_phon
|
|
2662
|
+
endfor
|
|
2663
|
+
removeObject: .dataID
|
|
2664
|
+
endproc
|
|
2665
|
+
|
|
2666
|
+
|
|
2667
|
+
procedure validate_syllable_tier: .infname$
|
|
2668
|
+
; (1) Validate symbols in phoneme tier: should be valid SAMPA or Praat phonetic symbols
|
|
2669
|
+
; (2) Validate symbols in syllable tier
|
|
2670
|
+
; (3) Validante alignment of syllable and phoneme tiers
|
|
2671
|
+
; (4) Check number of vowels (or syllabic consonants) within syllable.
|
|
2672
|
+
; Uses tiers named "phon...", "syll...", and optionally "word...".
|
|
2673
|
+
; Test is skipped when syllable label equals "_" or when word label equals "%".
|
|
2674
|
+
@fname_parts: .infname$
|
|
2675
|
+
.basename$ = result2$
|
|
2676
|
+
dataID = Read from file: .infname$
|
|
2677
|
+
@corpus_conversion: 1, "dataID", corpus$
|
|
2678
|
+
@tier_get: dataID, "^phon", "phone_tier", "Cannot find phoneme tier in '.basename$'", 1
|
|
2679
|
+
@tier_get: dataID, "^syll", "syllable_tier", "Cannot find syllable tier in '.basename$'", 1
|
|
2680
|
+
@tier_get: dataID, "^word", "word_tier", "Cannot find word tier in '.basename$'", 0
|
|
2681
|
+
|
|
2682
|
+
@debug_msg: "validate_syllable_tier: infname='.infname$' phone_tier='phone_tier'"
|
|
2683
|
+
.ns = Get number of intervals: syllable_tier
|
|
2684
|
+
for .j from 1 to .ns
|
|
2685
|
+
.t1_syll = Get start time of interval: syllable_tier, .j
|
|
2686
|
+
.t2_syll = Get end time of interval: syllable_tier, .j
|
|
2687
|
+
.label_syll$ = Get label of interval: syllable_tier, .j
|
|
2688
|
+
.dur = .t2_syll - .t1_syll
|
|
2689
|
+
if (.dur < 0.005)
|
|
2690
|
+
@msg: "Syllable with label '.label_syll$', starting at '.t1_syll:4', is only '.dur:4' s long"
|
|
2691
|
+
endif
|
|
2692
|
+
if (word_tier)
|
|
2693
|
+
.k = Get interval at time: word_tier, (.t1_syll+(.t2_syll-.t1_syll)/2)
|
|
2694
|
+
.word$ = Get label of interval: word_tier, .k
|
|
2695
|
+
.word$ = replace_regex$ (.word$, " ", "", 0) ; rm whitespace
|
|
2696
|
+
else
|
|
2697
|
+
.word$ = ""
|
|
2698
|
+
endif
|
|
2699
|
+
if (index("_#", .label_syll$)== 0 and .word$ <> "%") ; not pause, not noise, not incomprehensible
|
|
2700
|
+
errors = 0
|
|
2701
|
+
; Check label in tier
|
|
2702
|
+
@verify_tier: .label_syll$, .basename$, "syllable", .t1_syll
|
|
2703
|
+
errors += result
|
|
2704
|
+
@intervals_from_time_range_excl: dataID, phone_tier, .t1_syll, .t2_syll, "ph1", "ph2"
|
|
2705
|
+
nrof_vowels = 0
|
|
2706
|
+
nrof_syllabics = 0
|
|
2707
|
+
if (phone_tier)
|
|
2708
|
+
for .phon from ph1 to ph2
|
|
2709
|
+
.t1_phon = Get start time of interval: phone_tier, .phon
|
|
2710
|
+
.t2_phon = Get end time of interval: phone_tier, .phon
|
|
2711
|
+
.label_phon$ = Get label of interval: phone_tier, .phon
|
|
2712
|
+
.dur = .t2_phon - .t1_phon
|
|
2713
|
+
if (.dur < 0.005)
|
|
2714
|
+
@msg: "Phoneme with label '.label_phon$', starting at '.t1_phon:4', is only '.dur:4' s long"
|
|
2715
|
+
endif
|
|
2716
|
+
@verify_tier: .label_phon$, .basename$, "phoneme", .t1_phon
|
|
2717
|
+
errors += result
|
|
2718
|
+
; Check alignment between syllable tier and phoneme tier
|
|
2719
|
+
if (.phon == ph1 and .t1_phon <> .t1_syll)
|
|
2720
|
+
@msg: "Syllable ('.label_syll$') starting at '.t1_syll:4' not aligned with phoneme ('.label_phon$') start ('.t1_phon:4')"
|
|
2721
|
+
errors += 1
|
|
2722
|
+
endif
|
|
2723
|
+
if (.phon == ph2 and .t2_phon <> .t2_syll)
|
|
2724
|
+
@msg: "Syllable ('.label_syll$') ending at '.t2_syll:4' not aligned with phoneme ('.label_phon$') end ('.t2_phon:4')"
|
|
2725
|
+
errors += 1
|
|
2726
|
+
endif
|
|
2727
|
+
; Check number of vowels/syllabics
|
|
2728
|
+
@is_syllabic: .label_phon$
|
|
2729
|
+
if (result)
|
|
2730
|
+
nrof_syllabics += 1
|
|
2731
|
+
if (is_vowel)
|
|
2732
|
+
nrof_vowels += 1
|
|
2733
|
+
endif
|
|
2734
|
+
if (nrof_syllabics > 1)
|
|
2735
|
+
@msg: "Multiple vowels (or syllabics) in syllable ('.label_syll$') at time '.t1_syll:4'"
|
|
2736
|
+
endif
|
|
2737
|
+
endif ; syllabic
|
|
2738
|
+
if (.phon == ph2 and nrof_vowels == 0)
|
|
2739
|
+
@msg: "No vowel in syllable ('.label_syll$') at time '.t1_syll:4'"
|
|
2740
|
+
endif
|
|
2741
|
+
endfor ; phonemes in syllable
|
|
2742
|
+
endif ; phone_tier
|
|
2743
|
+
endif
|
|
2744
|
+
endfor
|
|
2745
|
+
removeObject: dataID
|
|
2746
|
+
endproc
|
|
2747
|
+
|
|
2748
|
+
|
|
2749
|
+
procedure verify_tier: .str$, .fname$, .type$, .time
|
|
2750
|
+
# Verify the label in a tier of type <.type> (either "phoneme", "syllable", "hesitation", "prominence", "polytonia")
|
|
2751
|
+
# <result> 0 when valid or 1 when error found.
|
|
2752
|
+
result = 0
|
|
2753
|
+
if (index(.str$, " "))
|
|
2754
|
+
@error_msg: "Label in tier at time '.time:2' of file <'.fname$'> contains whitespace"
|
|
2755
|
+
endif
|
|
2756
|
+
if (.type$ = "phoneme")
|
|
2757
|
+
.str$ = replace_regex$(.str$, "^""", "", 0) ; remove stress symbol
|
|
2758
|
+
; .str$ = replace_regex$(.str$, "^_$", "", 0) ; remove pause symbol
|
|
2759
|
+
.str$ = replace$(.str$, "\rc", "R", 0)
|
|
2760
|
+
.str$ = replace$(.str$, "\ep", "E", 0)
|
|
2761
|
+
if (.str$ = "_" or .str$ = "")
|
|
2762
|
+
elsif (not index_regex(.str$, "^[ptcCkbdgfsTDSvzZxGhmnNJlrRiyue2oE9OajwH@\?][=~]?$"))
|
|
2763
|
+
@msg: "Label in phone tier at time '.time:2' contains invalid character: <'.str$'>"
|
|
2764
|
+
result = 1
|
|
2765
|
+
endif
|
|
2766
|
+
elsif (.type$ = "syllable")
|
|
2767
|
+
.str$ = replace$(.str$, "\rc", "R", 0)
|
|
2768
|
+
.str$ = replace$(.str$, "\ep", "E", 0)
|
|
2769
|
+
if (not index_regex(.str$, "^[ptcCkbdgfsTDSvzZxGhmnNJlrRiyue2oE9OajwH@\?=~]+$"))
|
|
2770
|
+
@msg: "Label in syllable tier at time '.time:2' contains invalid character: <'.str$'>"
|
|
2771
|
+
result = 1
|
|
2772
|
+
endif
|
|
2773
|
+
elsif (.type$ = "hesitation")
|
|
2774
|
+
if (not index_regex(.str$, "^H?$"))
|
|
2775
|
+
@msg: "Label in hesitation tier at time '.time:2' contains invalid character: <'.str$'>"
|
|
2776
|
+
result = 1
|
|
2777
|
+
endif
|
|
2778
|
+
elsif (.type$ = "prominence")
|
|
2779
|
+
if (not index_regex(.str$, "^[0SW]$"))
|
|
2780
|
+
@msg: "Label in prominence tier at time '.time:2' contains invalid character: <'.str$'>"
|
|
2781
|
+
result = 1
|
|
2782
|
+
endif
|
|
2783
|
+
elsif (.type$ = "polytonia")
|
|
2784
|
+
if (not index_regex(.str$, "^[BLMHT_RrFfSC,]*$"))
|
|
2785
|
+
@msg: "Label in polytonia tier at time '.time:2' contains invalid sequence: <'.str$'>"
|
|
2786
|
+
result = 1
|
|
2787
|
+
endif
|
|
2788
|
+
if (index_regex(.str$, "CC") or index_regex(.str$, "__"))
|
|
2789
|
+
@msg: "Label in polytonia tier at time '.time:2' contains invalid sequence: <'.str$'>"
|
|
2790
|
+
result = 1
|
|
2791
|
+
endif
|
|
2792
|
+
elsif (.type$ = "stressref")
|
|
2793
|
+
if (not index_regex(.str$, "^[SW0]$"))
|
|
2794
|
+
@msg: "Label in stress-ref tier at time '.time:2' contains invalid character: <'.str$'>"
|
|
2795
|
+
result = 1
|
|
2796
|
+
endif
|
|
2797
|
+
else
|
|
2798
|
+
@msg: "verify_tier: ERROR invalid tier type: <'.type$'>"
|
|
2799
|
+
endif
|
|
2800
|
+
endproc
|
|
2801
|
+
|
|
2802
|
+
|
|
2803
|
+
procedure store_stylization: paramID, nucleiID, .filename$
|
|
2804
|
+
# Store stylization targets, indicating whether they are initial/final/intermediate relative to nucleus boundaries
|
|
2805
|
+
@debug_msg: "store_stylization: entry"
|
|
2806
|
+
filedelete '.filename$'
|
|
2807
|
+
selectObject: nucleiID ; segmentation into nuclei, etc
|
|
2808
|
+
.nnucl = Get number of intervals: nucleus_tier
|
|
2809
|
+
select paramID ; stylization (PitchTier)
|
|
2810
|
+
.ni = Get number of points
|
|
2811
|
+
for .interv from 1 to .nnucl ; indices of nuclei for which to plot stylization
|
|
2812
|
+
selectObject: nucleiID ; segmentation into nuclei, etc
|
|
2813
|
+
@is_nucleus: .interv
|
|
2814
|
+
if (result)
|
|
2815
|
+
.nx1 = Get start point... nucleus_tier .interv ; time of start of nucleus
|
|
2816
|
+
.nx2 = Get end point... nucleus_tier .interv ; time of end of nucleus
|
|
2817
|
+
select paramID ; stylization (PitchTier)
|
|
2818
|
+
# Check that the nearest indices are within nucleus
|
|
2819
|
+
.i1 = Get nearest index from time... .nx1
|
|
2820
|
+
repeat
|
|
2821
|
+
.t1 = Get time from index... .i1
|
|
2822
|
+
if (.t1 < .nx1)
|
|
2823
|
+
.i1 += 1
|
|
2824
|
+
endif
|
|
2825
|
+
until (.t1 >= .nx1 or .t1 >= .nx2 or .i1 >= .ni)
|
|
2826
|
+
.i2 = Get nearest index from time... .nx2
|
|
2827
|
+
repeat
|
|
2828
|
+
.t2 = Get time from index... .i2
|
|
2829
|
+
if (.t2 > .nx2)
|
|
2830
|
+
.i2 -= 1
|
|
2831
|
+
endif
|
|
2832
|
+
until (.t2 <= .nx2 or .t2 <= .nx1 or .i2 <= 1)
|
|
2833
|
+
for .i from .i1 to .i2 ; each stylization (tonal) segment i
|
|
2834
|
+
selectObject: paramID ; stylization (PitchTier)
|
|
2835
|
+
.x = Get time from index... .i
|
|
2836
|
+
.y = Get value at index... .i
|
|
2837
|
+
.yST = hertzToSemitones(.y) - hertzToSemitones(1)
|
|
2838
|
+
if (.i == .i1)
|
|
2839
|
+
if (syllables_available) ; write syllable start, end and label
|
|
2840
|
+
selectObject: nucleiID
|
|
2841
|
+
.isyll = Get interval at time... syllable_tier .x
|
|
2842
|
+
.t1s = Get start point... syllable_tier .isyll
|
|
2843
|
+
.t2s = Get end point... syllable_tier .isyll
|
|
2844
|
+
.s$ = Get label of interval... syllable_tier .isyll
|
|
2845
|
+
fileappend '.filename$' syll'tab$''.t1s:4''tab$''.t2s:4''tab$''.s$''newline$'
|
|
2846
|
+
endif
|
|
2847
|
+
fileappend '.filename$' nucl'tab$''.nx1:4''tab$''.nx2:4''newline$'
|
|
2848
|
+
.type$ = "tstart" ; target at start of nucleus
|
|
2849
|
+
elsif (.i == .i2)
|
|
2850
|
+
.type$ = "tend" ; target at end of nucleus
|
|
2851
|
+
; fileappend '.filename$' Ne'tab$''.nx2:4''newline$'
|
|
2852
|
+
else
|
|
2853
|
+
.type$ = "tin" ; target inside nucleus
|
|
2854
|
+
endif
|
|
2855
|
+
fileappend '.filename$' '.type$''tab$''.x:4''tab$''.y:1''tab$''.yST:1''newline$'
|
|
2856
|
+
endfor
|
|
2857
|
+
endif
|
|
2858
|
+
endfor
|
|
2859
|
+
@debug_msg: "store_stylization: exit"
|
|
2860
|
+
endproc
|
|
2861
|
+
|
|
2862
|
+
|
|
2863
|
+
procedure store_stylization_resampled paramID nucleiID .dx .filename$
|
|
2864
|
+
# Store resampled stylization, as if it were a pitch curve, sampled at .dx (second)
|
|
2865
|
+
# <nucleiID> segmentation into nuclei, etc
|
|
2866
|
+
filedelete '.filename$'
|
|
2867
|
+
selectObject: paramID ; stylization (PitchTier)
|
|
2868
|
+
.np = Get number of points
|
|
2869
|
+
; nrof targets in Pitch Tier
|
|
2870
|
+
.starttime = Get start time
|
|
2871
|
+
.endtime = Get end time
|
|
2872
|
+
.nx = floor((.endtime-.starttime)/.dx) ; nrof pitch samples in Pitch object
|
|
2873
|
+
fileappend '.filename$' File type = "ooTextFile"'newline$'
|
|
2874
|
+
fileappend '.filename$' Object class = "Pitch 1"'newline$''newline$'
|
|
2875
|
+
fileappend '.filename$' xmin = '.starttime''newline$'
|
|
2876
|
+
fileappend '.filename$' xmax = '.endtime''newline$'
|
|
2877
|
+
fileappend '.filename$' nx = '.nx''newline$'
|
|
2878
|
+
fileappend '.filename$' dx = '.dx''newline$'
|
|
2879
|
+
fileappend '.filename$' x1 = '.starttime''newline$'
|
|
2880
|
+
fileappend '.filename$' ceiling = 600'newline$'
|
|
2881
|
+
fileappend '.filename$' maxnCandidates = 1'newline$'
|
|
2882
|
+
fileappend '.filename$' frame []:'newline$'
|
|
2883
|
+
.i = 0 ; frame index
|
|
2884
|
+
.t = .starttime
|
|
2885
|
+
while (.t < .endtime)
|
|
2886
|
+
.i += 1 ; frame index for current time
|
|
2887
|
+
.value = 0 ; default F0 (outside nuclei)
|
|
2888
|
+
selectObject: nucleiID ; segmentation into nuclei, etc
|
|
2889
|
+
.interv = Get interval at time... nucleus_tier .t
|
|
2890
|
+
.label$ = Get label of interval... nucleus_tier .interv
|
|
2891
|
+
.nx1 = Get starting point... nucleus_tier .interv ; time of start of nucleus
|
|
2892
|
+
.nx2 = Get end point... nucleus_tier .interv ; time of end of nucleus
|
|
2893
|
+
selectObject: paramID ; stylization (PitchTier)
|
|
2894
|
+
.i1 = Get low index from time... .t ; index of point in pitch tier >= .t
|
|
2895
|
+
if (not .i1 == undefined) and (.i1 > 0)
|
|
2896
|
+
.t1 = Get time from index... .i1
|
|
2897
|
+
if (.i1 < .np)
|
|
2898
|
+
.i2 = .i1 + 1 ; index of next point in pitch tier
|
|
2899
|
+
.t2 = Get time from index... .i2 ; time of next point
|
|
2900
|
+
else
|
|
2901
|
+
.i2 = .i1
|
|
2902
|
+
.t2 = .nx2
|
|
2903
|
+
endif
|
|
2904
|
+
if (.label$ = "a" and .t >= .nx1 and .t <= .nx2 and .t2 <= .nx2)
|
|
2905
|
+
.y1 = Get value at index... .i1
|
|
2906
|
+
.y2 = Get value at index... .i2
|
|
2907
|
+
.value = .y1 + ((.y2-.y1)*(.t-.t1)/(.t2-.t1))
|
|
2908
|
+
endif
|
|
2909
|
+
endif
|
|
2910
|
+
fileappend '.filename$' 'tab$'frame ['.i']:'newline$'
|
|
2911
|
+
fileappend '.filename$' 'tab$''tab$'intensity = 0'newline$'
|
|
2912
|
+
fileappend '.filename$' 'tab$''tab$'nCandidates = 1'newline$'
|
|
2913
|
+
fileappend '.filename$' 'tab$''tab$'candidate = []'newline$'
|
|
2914
|
+
fileappend '.filename$' 'tab$''tab$''tab$'candidate = [1]:'newline$'
|
|
2915
|
+
fileappend '.filename$' 'tab$''tab$''tab$''tab$'frequency = '.value:0''newline$'
|
|
2916
|
+
fileappend '.filename$' 'tab$''tab$''tab$''tab$'strength = 1'newline$'
|
|
2917
|
+
.t += .dx
|
|
2918
|
+
endwhile
|
|
2919
|
+
endproc
|
|
2920
|
+
|
|
2921
|
+
|
|
2922
|
+
procedure option_get: .type$, .optname$, .varname$, .default$, .data$
|
|
2923
|
+
# Extract an option value from a command string.
|
|
2924
|
+
# <.type> option type: {word|real|boolean}
|
|
2925
|
+
# word: a string without whitespace
|
|
2926
|
+
# line: a string with whitespace
|
|
2927
|
+
# real: a number
|
|
2928
|
+
# boolean: a boolean, where TRUE is encoded by one of: y, Y, yes, 1, T, TRUE
|
|
2929
|
+
# <.optname> option name, e.g. "file=", "threshold="
|
|
2930
|
+
# <.varname> name of variable, where value of option will be stored
|
|
2931
|
+
# <.default> default option value, when option is not mentioned in command
|
|
2932
|
+
# <.data> command argument string, where options will be searched
|
|
2933
|
+
if (.type$ = "word")
|
|
2934
|
+
.s$ = extractWord$ (.data$, .optname$)
|
|
2935
|
+
'.varname$'$ = .default$
|
|
2936
|
+
if (length (.s$))
|
|
2937
|
+
'.varname$'$ = .s$
|
|
2938
|
+
endif
|
|
2939
|
+
elsif (.type$ = "line")
|
|
2940
|
+
.s$ = extractLine$ (.data$, .optname$)
|
|
2941
|
+
'.varname$'$ = .default$
|
|
2942
|
+
if (length (.s$))
|
|
2943
|
+
'.varname$'$ = .s$
|
|
2944
|
+
endif
|
|
2945
|
+
elsif (.type$ = "real")
|
|
2946
|
+
.v = extractNumber (.data$, .optname$)
|
|
2947
|
+
'.varname$' = '.default$'
|
|
2948
|
+
if not (.v == undefined)
|
|
2949
|
+
'.varname$' = .v
|
|
2950
|
+
endif
|
|
2951
|
+
elsif (.type$ = "boolean")
|
|
2952
|
+
.s$ = extractWord$ (.data$, .optname$)
|
|
2953
|
+
'.varname$' = 0 ; default value
|
|
2954
|
+
if (length (.s$) < 1)
|
|
2955
|
+
.s$ = .default$
|
|
2956
|
+
endif
|
|
2957
|
+
if (index (":y:1:Y:yes:T:TRUE:", ":'.s$':") > 0)
|
|
2958
|
+
'.varname$' = 1
|
|
2959
|
+
endif
|
|
2960
|
+
endif
|
|
2961
|
+
endproc
|
|
2962
|
+
|
|
2963
|
+
|
|
2964
|
+
procedure option_check: .type$, .optname$, .valids$, .value$
|
|
2965
|
+
# Check whether the value of an option belongs to a set of accepted values (separated by a colon)
|
|
2966
|
+
if (.type$ = "word")
|
|
2967
|
+
if (index (.valids$, ":'.value$':") <= 0)
|
|
2968
|
+
@msg: "invalid value <'.value$'> for option <'.optname$'>"
|
|
2969
|
+
endif
|
|
2970
|
+
endif
|
|
2971
|
+
endproc
|
|
2972
|
+
|
|
2973
|
+
|
|
2974
|
+
procedure batch_cmd_start: .cmd$, .args$
|
|
2975
|
+
@msg: " 'newline$'Batch command> '.cmd$': '.args$''newline$'"
|
|
2976
|
+
endproc
|
|
2977
|
+
|
|
2978
|
+
|
|
2979
|
+
procedure batch_cmd_ready: .cmd$
|
|
2980
|
+
@msg: " 'newline$'Batch command> '.cmd$': ... Ready"
|
|
2981
|
+
endproc
|
|
2982
|
+
|
|
2983
|
+
|
|
2984
|
+
procedure check_regex: .re$
|
|
2985
|
+
if (index_regex (.re$, "^\*"))
|
|
2986
|
+
@fatal_error: "Invalid regular expression: <'.re$'> (\* should be preceded by some string)"
|
|
2987
|
+
endif
|
|
2988
|
+
endproc
|
|
2989
|
+
|
|
2990
|
+
|
|
2991
|
+
# Batch command "calculate_pitch"
|
|
2992
|
+
# arguments: option_name, default_value
|
|
2993
|
+
# file= "" ; input file of files specified by a regular expression
|
|
2994
|
+
# time_step= 0.005 ; frame period
|
|
2995
|
+
# f0min= 0 ; lower value for F0 detection - f0min=0 selects auto pitch range detection
|
|
2996
|
+
# f0max= 450 ; upper value for F0 detection
|
|
2997
|
+
procedure calculate_pitch: .cmda$
|
|
2998
|
+
@debug_msg: "calculate_pitch: entry"
|
|
2999
|
+
@batch_cmd_start: "calculate_pitch", .cmda$
|
|
3000
|
+
@initialization_main
|
|
3001
|
+
@option_get: "word", "file=", "inputfname", "", .cmda$
|
|
3002
|
+
@option_get: "real", "time_step=", "time_step", "0.005", .cmda$
|
|
3003
|
+
@option_get: "real", "f0min=", "minimum_pitch", "0", .cmda$
|
|
3004
|
+
@option_get: "real", "f0max=", "maximum_pitch", "450", .cmda$
|
|
3005
|
+
|
|
3006
|
+
if (variableExists ("corpus_home$"))
|
|
3007
|
+
if (variableExists ("corpus_subdir_sound$"))
|
|
3008
|
+
inputfname$ = corpus_home$ + corpus_subdir_sound$ + inputfname$
|
|
3009
|
+
else
|
|
3010
|
+
inputfname$ = corpus_home$ + inputfname$
|
|
3011
|
+
endif
|
|
3012
|
+
endif
|
|
3013
|
+
@check_regex: inputfname$
|
|
3014
|
+
@files_get_regex: "filelistID", inputfname$
|
|
3015
|
+
selectObject: filelistID
|
|
3016
|
+
.nfiles = Get number of strings
|
|
3017
|
+
if (.nfiles == 0)
|
|
3018
|
+
@msg: "No input files found for <'inputfname$'>"
|
|
3019
|
+
endif
|
|
3020
|
+
task = task_calc_pitch
|
|
3021
|
+
for .ifile to .nfiles
|
|
3022
|
+
selectObject: filelistID
|
|
3023
|
+
.fname$ = Get string: .ifile
|
|
3024
|
+
@initialization_per_file
|
|
3025
|
+
@construct_filenames: .fname$
|
|
3026
|
+
@fname_parts: pitchfile$
|
|
3027
|
+
.dir$ = result4$
|
|
3028
|
+
if (not fileReadable (.dir$))
|
|
3029
|
+
@fatal_error: "calculate_pitch: Cannot find directory <'.dir$'>. Create it first."
|
|
3030
|
+
endif
|
|
3031
|
+
@calculate_pitch_int
|
|
3032
|
+
removeObject: soundID, pitchID
|
|
3033
|
+
endfor
|
|
3034
|
+
removeObject: filelistID
|
|
3035
|
+
@batch_cmd_ready: "calculate_pitch"
|
|
3036
|
+
@debug_msg: "calculate_pitch: exit"
|
|
3037
|
+
endproc
|
|
3038
|
+
|
|
3039
|
+
|
|
3040
|
+
# Batch command "prosogram_variants"
|
|
3041
|
+
# arguments: option_name, default_value
|
|
3042
|
+
# file= "" ; input file of files specified by a regular expression
|
|
3043
|
+
# time_step= 0.005
|
|
3044
|
+
# f0min= 0 ; lower value for F0 detection - f0min=0 selects auto pitch range detection
|
|
3045
|
+
# f0max= 450 ; upper value for F0 detection
|
|
3046
|
+
# t1= 0 ; start time of analysis
|
|
3047
|
+
# t2= 0 ; end time of analysis
|
|
3048
|
+
# volatile= no ; analysis of speech signal restricted to analysis time interval
|
|
3049
|
+
# segmentation= optimal ; segmentation type, selected from: {optimal, vow-nucl, extern, int-BP, asyll, rhyme, syll+vow, syll, voiced}
|
|
3050
|
+
# g= ; glissando threshold; default selects adaptive threshold 0.16 + 0.32
|
|
3051
|
+
# dg= 30 ; differential glissando threshold
|
|
3052
|
+
# dmin= 0.035 ; minimum duration of tonal segment
|
|
3053
|
+
# draw= yes ; draw Prosogram in Graphics window and write graphics files
|
|
3054
|
+
# time_incr= 3.0 ; duration of prosogram pane
|
|
3055
|
+
# wide= yes ; selects wide size for Prosogram, otherwise compact size
|
|
3056
|
+
# rich= yes ; selects rich format, otherwise light format
|
|
3057
|
+
# pitch= yes ; plot pitch (overrides value of "rich")
|
|
3058
|
+
# pitchrange= yes ; plot pitch range
|
|
3059
|
+
# targets= 0 ; draw pitch targets in ST (=1) or in Hz (=2)
|
|
3060
|
+
# tiers= *1,2,3 ; tiers shown in the graphics output; tiers are specified by name or number (no whitespace allowed); * indicates SAMPA to IPA conversion.
|
|
3061
|
+
# outputfile= "" ; filename of graphics file (without filename extension). Zero-length string selects corpus-defined path and numbering in filename.
|
|
3062
|
+
# outputformat=PNG300 ; format of graphics file, selected from: {EPS, EMF, PNG300, PNG600, JPG300, JPG600}
|
|
3063
|
+
# number= yes ; number graphics files (path+basename+number+extension)
|
|
3064
|
+
# collect= no ; collect output of duration calculation in a single output file
|
|
3065
|
+
# save= no ; save intermediate data in files
|
|
3066
|
+
# cleanup= yes ; cleanup data after processing
|
|
3067
|
+
# settings= yes ; show settings (segmentation type, thresholds) in Prosogram
|
|
3068
|
+
# x_scale= yes ; show X scale (numbers on axis)
|
|
3069
|
+
# y_scale= yes ; show Y scale (numbers on axis)
|
|
3070
|
+
# y_scale_r= yes ; show Y scale (values in Hz on right axis)
|
|
3071
|
+
# pauses= yes ; show pauses (by "P")
|
|
3072
|
+
# tg_bound= yes/no ; show vertical boundaries of tier intervals in textgrid (default is yes in rich mode and no in light mode
|
|
3073
|
+
# portee= yes ; show portee (horizontal ST calibration lines)
|
|
3074
|
+
# traject= no ; show intrasyllab up/down and intersyllab pitch intervals
|
|
3075
|
+
# hesit= no ; show hesitations
|
|
3076
|
+
# hesit_method=all ; set method for hesitation detection
|
|
3077
|
+
# rhythm= no ; show rhythm
|
|
3078
|
+
|
|
3079
|
+
|
|
3080
|
+
procedure prosogram_variants: .cmda$
|
|
3081
|
+
; @logging: "reset debug timed", "_log.txt"
|
|
3082
|
+
@debug_msg: "prosogram_variants: entry, command=<'.cmda$'>"
|
|
3083
|
+
@initialization_main
|
|
3084
|
+
task = task_prosogram
|
|
3085
|
+
auto_pitchrange = 1 ; automatic pitch range adjustment in plot
|
|
3086
|
+
outputmode$ = "Fill"
|
|
3087
|
+
@option_get: "word", "file=", "inputfname", "", .cmda$
|
|
3088
|
+
@option_get: "real", "time_step=", "time_step", "0.005", .cmda$
|
|
3089
|
+
@option_get: "real", "f0min=", "minimum_pitch", "0", .cmda$
|
|
3090
|
+
@option_get: "real", "f0max=", "maximum_pitch", "450", .cmda$
|
|
3091
|
+
@option_get: "real", "t1=", "pmf_t1", "0", .cmda$
|
|
3092
|
+
@option_get: "real", "t2=", "pmf_t2", "0", .cmda$
|
|
3093
|
+
@option_get: "real", "time_incr=", "timeincr", "3.0", .cmda$
|
|
3094
|
+
@option_get: "word", "segmentation=", "segmentation_name", "optimal", .cmda$
|
|
3095
|
+
@option_check: "word", "segmentation", ":optimal:vow-nucl:extern:loudness:int-BP:asyll:rhyme:syll+vow:syll:voiced:", segmentation_name$
|
|
3096
|
+
@option_get: "real", "g=", "glissando", "-0.32", .cmda$
|
|
3097
|
+
@option_get: "real", "dg=", "diffgt", "30", .cmda$
|
|
3098
|
+
@option_get: "real", "dmin=", "mindur_ts", "0.035", .cmda$
|
|
3099
|
+
@option_get: "boolean", "wide=", "wide", "yes", .cmda$
|
|
3100
|
+
@option_get: "boolean", "rich=", "rich", "yes", .cmda$
|
|
3101
|
+
if (rich)
|
|
3102
|
+
tg_bound_default$ = "yes"
|
|
3103
|
+
show_intensity = 1
|
|
3104
|
+
show_f0 = 1
|
|
3105
|
+
else
|
|
3106
|
+
tg_bound_default$ = "no"
|
|
3107
|
+
show_intensity = 0
|
|
3108
|
+
show_f0 = 0
|
|
3109
|
+
endif
|
|
3110
|
+
@option_get: "word", "tiers=", "tiers_to_show", "*1,2,3", .cmda$
|
|
3111
|
+
@option_get: "boolean", "collect=", "collect_output", "no", .cmda$
|
|
3112
|
+
@option_get: "word", "outputfile=", "output_filename", "", .cmda$
|
|
3113
|
+
@option_get: "word", "outputsuffix=", "output_suffix", "", .cmda$
|
|
3114
|
+
@option_get: "word", "outputformat=", "output_format", "PNG300", .cmda$
|
|
3115
|
+
@option_check: "word", "outputformat", ":EPS:EMF:PNG300:PNG600:JPG300:JPG600:", output_format$
|
|
3116
|
+
output_format$ = replace_regex$ (output_format$, "PNG(\d+)", "PNG \1 dpi", 1)
|
|
3117
|
+
@option_get: "boolean", "number=", "file_numbering", "yes", .cmda$
|
|
3118
|
+
@option_get: "real", "jpegq=", "jpegq", "75", .cmda$
|
|
3119
|
+
@option_get: "boolean", "save=", "save_intermediate_data", "no", .cmda$
|
|
3120
|
+
@option_get: "boolean", "polytonia=", "polytonia_annotation", "no", .cmda$
|
|
3121
|
+
@option_get: "boolean", "boundaries=", "boundary_annotation", "no", .cmda$
|
|
3122
|
+
@option_get: "boolean", "cleanup=", "do_cleanup", "yes", .cmda$
|
|
3123
|
+
glissando_low = 0.16
|
|
3124
|
+
if (glissando < 0)
|
|
3125
|
+
adaptive_glissando = 1
|
|
3126
|
+
glissando = abs(glissando)
|
|
3127
|
+
else
|
|
3128
|
+
adaptive_glissando = 0
|
|
3129
|
+
endif
|
|
3130
|
+
if (wide)
|
|
3131
|
+
viewsize$ = "wide"
|
|
3132
|
+
else
|
|
3133
|
+
viewsize$ = "compact"
|
|
3134
|
+
endif
|
|
3135
|
+
if (variableExists ("corpus_home$"))
|
|
3136
|
+
if (variableExists ("corpus_subdir_sound$"))
|
|
3137
|
+
inputfname$ = corpus_home$ + corpus_subdir_sound$ + inputfname$
|
|
3138
|
+
else
|
|
3139
|
+
inputfname$ = corpus_home$ + inputfname$
|
|
3140
|
+
endif
|
|
3141
|
+
endif
|
|
3142
|
+
|
|
3143
|
+
@debug_msg: "prosogram_variants: inputfile='inputfname$' t1='pmf_t1' t2='pmf_t2' time_step='time_step:3' segmentation_name='segmentation_name$' wide='wide' rich='rich' adapative='adaptive_glissando'"
|
|
3144
|
+
@debug_msg: "prosogram_variants: outputfile='output_filename$'"
|
|
3145
|
+
|
|
3146
|
+
@check_regex: inputfname$
|
|
3147
|
+
@files_get_regex: "filelistID", inputfname$
|
|
3148
|
+
selectObject: filelistID
|
|
3149
|
+
.nfiles = Get number of strings
|
|
3150
|
+
if (.nfiles == 0)
|
|
3151
|
+
@msg: "No input files found for <'inputfname$'>"
|
|
3152
|
+
endif
|
|
3153
|
+
; @task_flags and @initialization_multiple_files initialize various default settings, which may be overridden by optional settings in batch command
|
|
3154
|
+
@task_flags
|
|
3155
|
+
if (save_intermediate_data)
|
|
3156
|
+
needs_prosodic_profile = 1
|
|
3157
|
+
endif
|
|
3158
|
+
|
|
3159
|
+
@initialization_multiple_files
|
|
3160
|
+
; The following lines override some variables initialized by @task_flags and @initialization_multiple_files
|
|
3161
|
+
@option_get: "boolean", "volatile=", "volatile", "no", .cmda$
|
|
3162
|
+
if (task == task_pitch_plot)
|
|
3163
|
+
@option_get: "boolean", "intensity=", "show_intensity", "no", .cmda$
|
|
3164
|
+
endif
|
|
3165
|
+
@option_get: "boolean", "pitchrange=", "show_pitchrange", "no", .cmda$
|
|
3166
|
+
if (variableExists ("show_pitchrange"))
|
|
3167
|
+
if (show_pitchrange)
|
|
3168
|
+
needs_pitchrange = 1
|
|
3169
|
+
endif
|
|
3170
|
+
endif
|
|
3171
|
+
@option_get: "boolean", "draw=", "draw_prosograms", "yes", .cmda$
|
|
3172
|
+
@option_get: "boolean", "pitch=", "show_f0", "yes", .cmda$
|
|
3173
|
+
@option_get: "boolean", "portee=", "show_portee", "yes", .cmda$
|
|
3174
|
+
@option_get: "boolean", "tg_bound=", "show_tg_boundaries", tg_bound_default$, .cmda$
|
|
3175
|
+
@option_get: "boolean", "settings=", "show_settings", "yes", .cmda$
|
|
3176
|
+
@option_get: "boolean", "x_scale=", "show_x_scale", "yes", .cmda$
|
|
3177
|
+
@option_get: "boolean", "y_scale=", "show_y_scale", "yes", .cmda$
|
|
3178
|
+
@option_get: "boolean", "y_scale_r=", "show_y_scale_r", "yes", .cmda$
|
|
3179
|
+
@option_get: "boolean", "tiernames=", "show_tiernames", "yes", .cmda$
|
|
3180
|
+
@option_get: "boolean", "pauses=", "show_pauses", "no", .cmda$
|
|
3181
|
+
@option_get: "real", "targets=", "draw_pitch_target_values", "0", .cmda$
|
|
3182
|
+
@option_get: "boolean", "traject=", "show_trajectories", "no", .cmda$
|
|
3183
|
+
@option_get: "boolean", "grey=", "greyscale", "no", .cmda$
|
|
3184
|
+
@option_get: "boolean", "hesit=", "show_hesitations", "no", .cmda$
|
|
3185
|
+
@option_get: "boolean", "hesit_anno=", "hesitation_annotation", "no", .cmda$
|
|
3186
|
+
@option_get: "word", "hesit_method=", "hesitation_method", "all", .cmda$
|
|
3187
|
+
@option_check: "word", "hesitation_method", ":all:none:phon+pros:words:annotation:", hesitation_method$
|
|
3188
|
+
@option_get: "boolean", "rhythm=", "show_rhythm", "no", .cmda$
|
|
3189
|
+
@debug_msg: "prosogram_variants: wide='wide' rich='rich' pitchrange='show_pitchrange' adapative='adaptive_glissando'"
|
|
3190
|
+
@debug_msg: "prosogram_variants: show_tiernames='show_tiernames'"
|
|
3191
|
+
for .ifile to .nfiles
|
|
3192
|
+
selectObject: filelistID
|
|
3193
|
+
.fname$ = Get string: .ifile
|
|
3194
|
+
@debug_msg: "prosogram_variants: pmf_t1='pmf_t1:4', pmf_t2='pmf_t2:4', fname='.fname$'"
|
|
3195
|
+
@process_one_input_file: pmf_t1, pmf_t2, .ifile, .nfiles, .fname$
|
|
3196
|
+
endfor
|
|
3197
|
+
@cleanup_global
|
|
3198
|
+
removeObject: filelistID
|
|
3199
|
+
@debug_msg: "prosogram_variants: exit"
|
|
3200
|
+
endproc
|
|
3201
|
+
|
|
3202
|
+
|
|
3203
|
+
procedure prosogram: .cmda$
|
|
3204
|
+
@batch_cmd_start: "prosogram", .cmda$
|
|
3205
|
+
@prosogram_variants: .cmda$
|
|
3206
|
+
@batch_cmd_ready: "prosogram"
|
|
3207
|
+
endproc
|
|
3208
|
+
|
|
3209
|
+
|
|
3210
|
+
procedure polytonia: .cmda$
|
|
3211
|
+
@batch_cmd_start: "polytonia", .cmda$
|
|
3212
|
+
@prosogram_variants: .cmda$ + " polytonia=yes"
|
|
3213
|
+
@batch_cmd_ready: "polytonia"
|
|
3214
|
+
endproc
|
|
3215
|
+
|
|
3216
|
+
|
|
3217
|
+
procedure detect_boundaries: .cmda$
|
|
3218
|
+
@batch_cmd_start: "detect_boundaries", .cmda$
|
|
3219
|
+
@prosogram_variants: .cmda$ + " boundaries=yes"
|
|
3220
|
+
@batch_cmd_ready: "detect_boundaries"
|
|
3221
|
+
endproc
|
|
3222
|
+
|
|
3223
|
+
|
|
3224
|
+
procedure hesitations: .cmda$
|
|
3225
|
+
@batch_cmd_start: "hesitations", .cmda$
|
|
3226
|
+
@prosogram_variants: .cmda$ + " hesit_anno=yes"
|
|
3227
|
+
@batch_cmd_ready: "hesitations"
|
|
3228
|
+
endproc
|
|
3229
|
+
|
|
3230
|
+
|
|
3231
|
+
procedure plot_pitch: .cmda$
|
|
3232
|
+
@batch_cmd_start: "plot_pitch", .cmda$
|
|
3233
|
+
task = task_pitch_plot
|
|
3234
|
+
@prosogram_variants: .cmda$
|
|
3235
|
+
@batch_cmd_ready: "plot_pitch"
|
|
3236
|
+
endproc
|
|
3237
|
+
|
|
3238
|
+
|
|
3239
|
+
procedure prominence: .cmda$
|
|
3240
|
+
@batch_cmd_start: "prominence", .cmda$
|
|
3241
|
+
task = task_prosogram
|
|
3242
|
+
needs_loudness = 1
|
|
3243
|
+
calc_prominence = 1
|
|
3244
|
+
show_prominence = 1
|
|
3245
|
+
@prosogram_variants: .cmda$
|
|
3246
|
+
calc_prominence = 0
|
|
3247
|
+
show_prominence = 0
|
|
3248
|
+
@batch_cmd_ready: "prominence"
|
|
3249
|
+
endproc
|
|
3250
|
+
|
|
3251
|
+
|
|
3252
|
+
procedure segmentation: .cmda$
|
|
3253
|
+
@debug_msg: "segmentation: entry"
|
|
3254
|
+
@batch_cmd_start: "segmentation", .cmda$
|
|
3255
|
+
@initialization_main
|
|
3256
|
+
@option_get: "word", "file=", "inputfname", "", .cmda$
|
|
3257
|
+
@option_get: "real", "t1=", "pmf_t1", "0", .cmda$
|
|
3258
|
+
@option_get: "real", "t2=", "pmf_t2", "0", .cmda$
|
|
3259
|
+
@option_get: "real", "time_step=", "time_step", "0.005", .cmda$
|
|
3260
|
+
@option_get: "word", "method=", "segmentation_name", "asyll", .cmda$
|
|
3261
|
+
@option_check: "word", "method", ":asyll:specsim:pitchchange:pitchterrace:", segmentation_name$
|
|
3262
|
+
@option_get: "real", "left=", "trimleft", "0", .cmda$
|
|
3263
|
+
@option_get: "real", "right=", "trimright", "0", .cmda$
|
|
3264
|
+
@option_get: "word", "output_suffix=", "output_suffix", "", .cmda$
|
|
3265
|
+
@option_get: "boolean", "save_BP=", "save_BP", "no", .cmda$
|
|
3266
|
+
if (variableExists ("corpus_home$"))
|
|
3267
|
+
if (variableExists ("corpus_subdir_sound$"))
|
|
3268
|
+
inputfname$ = corpus_home$ + corpus_subdir_sound$ + inputfname$
|
|
3269
|
+
else
|
|
3270
|
+
inputfname$ = corpus_home$ + inputfname$
|
|
3271
|
+
endif
|
|
3272
|
+
endif
|
|
3273
|
+
task = task_segmentation
|
|
3274
|
+
@map_segmentation_type: segmentation_name$
|
|
3275
|
+
@debug_msg: "segmentation: segmentation_name='segmentation_name$'"
|
|
3276
|
+
|
|
3277
|
+
@check_regex: inputfname$
|
|
3278
|
+
@files_get_regex: "filelistID", inputfname$
|
|
3279
|
+
selectObject: filelistID
|
|
3280
|
+
.nfiles = Get number of strings
|
|
3281
|
+
if (.nfiles == 0)
|
|
3282
|
+
@msg: "No input files found for <'inputfname$'>"
|
|
3283
|
+
endif
|
|
3284
|
+
@task_flags
|
|
3285
|
+
@initialization_multiple_files
|
|
3286
|
+
; The following lines override some variables initialized by @task_flags and @initialization_multiple_files
|
|
3287
|
+
if (segmentation_name$ == "asyll")
|
|
3288
|
+
@option_get: "real", "fc_low=", "fc_low", "300", .cmda$
|
|
3289
|
+
@option_get: "real", "fc_high=", "fc_high", "3500", .cmda$
|
|
3290
|
+
endif
|
|
3291
|
+
@option_get: "real", "mindiff=", "mindiff", "3", .cmda$
|
|
3292
|
+
@option_get: "real", "leftdiff=", "diff_left", "2", .cmda$
|
|
3293
|
+
if (segmentation_name$ == "pitchchange")
|
|
3294
|
+
@option_get: "real", "threshold=", "pc_threshold", "2", .cmda$
|
|
3295
|
+
endif
|
|
3296
|
+
if (segmentation_name$ == "pitchterrace")
|
|
3297
|
+
@option_get: "real", "threshold=", "pc_threshold", "2", .cmda$
|
|
3298
|
+
@option_get: "real", "mindur=", "mindur_terrace", "0", .cmda$
|
|
3299
|
+
endif
|
|
3300
|
+
volatile = 1 ; avoid reading and overwriting pitch and intensity of BP filtered speech
|
|
3301
|
+
@msg: " 'newline$'Batch command segmentation: method='segmentation_name$' fc_low='fc_low:0' fc_high='fc_high:0' mindiff='mindiff' leftdiff='diff_left' time_step='time_step' output_suffix=<'output_suffix$'>"
|
|
3302
|
+
for .ifile to .nfiles
|
|
3303
|
+
selectObject: filelistID
|
|
3304
|
+
.fname$ = Get string: .ifile
|
|
3305
|
+
@msg: "Processing file <'.fname$'>"
|
|
3306
|
+
@process_one_input_file: pmf_t1, pmf_t2, .ifile, .nfiles, .fname$
|
|
3307
|
+
endfor
|
|
3308
|
+
@cleanup_global
|
|
3309
|
+
removeObject: filelistID
|
|
3310
|
+
@batch_cmd_ready: "segmentation"
|
|
3311
|
+
endproc
|
|
3312
|
+
|
|
3313
|
+
|
|
3314
|
+
# Batch command "calculate_duration_data"
|
|
3315
|
+
# arguments: option_name, default_value
|
|
3316
|
+
# file= "" ; input file or files specified by a regular expression
|
|
3317
|
+
# unit= "sound" ; speech segment for which to obtain duration data: sound, syllable...
|
|
3318
|
+
# collect= "yes" ; collect results of multiple input files into same output file
|
|
3319
|
+
# out= default ; output filename, default = "_dur_<unit>.txt"
|
|
3320
|
+
procedure calculate_duration_data: .cmda$
|
|
3321
|
+
@batch_cmd_start: "calculate_duration_data", .cmda$
|
|
3322
|
+
@initialization_main
|
|
3323
|
+
; Specify the input filename as TextGrid file
|
|
3324
|
+
@option_get: "word", "file=", "filespec", "", .cmda$
|
|
3325
|
+
@option_get: "word", "unit=", "unit", "sound", .cmda$
|
|
3326
|
+
; Collect output = append results of multiple input files in one output file
|
|
3327
|
+
@option_get: "boolean", "collect=", "collect_output", "yes", .cmda$
|
|
3328
|
+
; @option_get: "word", "out=", collectfile, "_dur_'unit$'.txt", .cmda$
|
|
3329
|
+
|
|
3330
|
+
@check_regex: filespec$
|
|
3331
|
+
@files_get_regex: "filelistID", filespec$
|
|
3332
|
+
selectObject: filelistID
|
|
3333
|
+
.nfiles = Get number of strings
|
|
3334
|
+
if (.nfiles == 0)
|
|
3335
|
+
@msg: "No input files found for <'filespec$'>"
|
|
3336
|
+
endif
|
|
3337
|
+
; prepare default filename for outputfile
|
|
3338
|
+
@fname_parts: filespec$
|
|
3339
|
+
indir$ = result4$
|
|
3340
|
+
@option_get: "word", "out=", "outfname", "'indir$'_dur_'unit$'.txt", .cmda$
|
|
3341
|
+
@msg: "calculate_duration_data: filespec='filespec$', '.nfiles' files, collect='collect_output', outfname='outfname$'"
|
|
3342
|
+
for .j to .nfiles
|
|
3343
|
+
selectObject: filelistID
|
|
3344
|
+
.fname$ = Get string: .j
|
|
3345
|
+
@fname_parts: .fname$
|
|
3346
|
+
.basename$ = result2$
|
|
3347
|
+
;.fname$ = indir$ + .basename$ + ".TextGrid"
|
|
3348
|
+
@msg: "file='.j' <'.fname$'> out='outfname$'"
|
|
3349
|
+
dataID = Read from file: .fname$
|
|
3350
|
+
@corpus_conversion: 1, "dataID", corpus$
|
|
3351
|
+
if (unit$ = "sound")
|
|
3352
|
+
.f = .j
|
|
3353
|
+
if (not collect_output)
|
|
3354
|
+
.f = 1
|
|
3355
|
+
endif
|
|
3356
|
+
@duration_data_from_textgrid: dataID, .f, unit$, outfname$, .basename$
|
|
3357
|
+
endif
|
|
3358
|
+
removeObject: dataID
|
|
3359
|
+
endfor
|
|
3360
|
+
removeObject: filelistID
|
|
3361
|
+
@batch_cmd_ready: "calculate_duration_data"
|
|
3362
|
+
endproc
|
|
3363
|
+
|
|
3364
|
+
|
|
3365
|
+
procedure convert_eps .cmda$
|
|
3366
|
+
# Assumes Ghostscript, nconvert and pdftk are installed.
|
|
3367
|
+
# nconvert is required when converting to GIF format. http://www.xnview.com/
|
|
3368
|
+
# pdftk is required when creating multipage PDF files. http://www.pdflabs.com/tools/pdttk-the-pdf-toolkit/
|
|
3369
|
+
; Specify the filename as eps files
|
|
3370
|
+
call option_get word file= filespec "*.eps" '.cmda$'
|
|
3371
|
+
call fname_parts 'filespec$'
|
|
3372
|
+
basename$ = result2$
|
|
3373
|
+
indir$ = result4$
|
|
3374
|
+
call option_get word outfname= outfname "'basename$'.pdf" '.cmda$'
|
|
3375
|
+
call option_get word format= graphics_format "JPG" '.cmda$'
|
|
3376
|
+
call option_check word graphics_format :PNG:GIF:JPG:PDF:multipage_PDF: 'graphics_format$'
|
|
3377
|
+
call option_get word resolution= resolution "300" '.cmda$'
|
|
3378
|
+
call option_check word resolution :72:96:120:300:600: 'resolution$'
|
|
3379
|
+
call option_get word outdir= outdir "<same_as_input>" '.cmda$'
|
|
3380
|
+
if (outdir$ = "<same_as_input>")
|
|
3381
|
+
outdir$ = indir$
|
|
3382
|
+
else
|
|
3383
|
+
call fname_parts 'outdir$'
|
|
3384
|
+
outdir$ = result4$
|
|
3385
|
+
endif
|
|
3386
|
+
Create Strings as file list... filelist 'filespec$'
|
|
3387
|
+
.nfiles = Get number of strings
|
|
3388
|
+
pdf_files$ = ""
|
|
3389
|
+
call msg convert_eps: filespec='filespec$', '.nfiles' files, outdir='outdir$'
|
|
3390
|
+
call msg convert_eps: format='graphics_format$'
|
|
3391
|
+
eps_pdf$ = path_ghostscript$ + " -dNOPAUSE -dBATCH -dQUIET -sDEVICE=pdfwrite"
|
|
3392
|
+
eps_jpg$ = path_ghostscript$ + " -dNOPAUSE -dBATCH -dQUIET -sDEVICE=jpeg -sEPSCrop"
|
|
3393
|
+
eps_png$ = path_ghostscript$ + " -dNOPAUSE -dBATCH -dQUIET -sDEVICE=png16 -sEPSCrop"
|
|
3394
|
+
png_gif$ = path_nconvert$
|
|
3395
|
+
for .j to .nfiles
|
|
3396
|
+
select Strings filelist
|
|
3397
|
+
.fname$ = Get string... .j
|
|
3398
|
+
call fname_parts '.fname$'
|
|
3399
|
+
basename$ = result2$
|
|
3400
|
+
pdf_files$ = pdf_files$ + "'basename$'.pdf "
|
|
3401
|
+
src$ = indir$ + basename$
|
|
3402
|
+
dst$ = outdir$ + basename$
|
|
3403
|
+
res = 'resolution$'
|
|
3404
|
+
if (graphics_format$ = "GIF")
|
|
3405
|
+
system 'eps_png$' -r'res' -sOutputFile="'dst$'.png" "'src$'.eps"
|
|
3406
|
+
system 'png_gif$' -quiet -out gif 'dst$'.png
|
|
3407
|
+
filedelete 'dst$'.png
|
|
3408
|
+
elsif (graphics_format$ = "PNG")
|
|
3409
|
+
system 'eps_png$' -r'res' -sOutputFile="'dst$'.png" "'src$'.eps"
|
|
3410
|
+
elsif (index (graphics_format$, "PDF")) ; PDF or multipage PDF
|
|
3411
|
+
if (res == 96 or res == 120)
|
|
3412
|
+
@msg: "Invalid resolution for PDF. Using 300 dpi."
|
|
3413
|
+
res = 300
|
|
3414
|
+
endif
|
|
3415
|
+
;printline command= system 'eps_pdf$' -r'res' -sOutputFile="'dst$'.pdf" "'src$'.eps"
|
|
3416
|
+
system 'eps_pdf$' -r'res' -sOutputFile="'dst$'.pdf" "'src$'.eps"
|
|
3417
|
+
elsif (graphics_format$ = "JPG")
|
|
3418
|
+
if (res == 96 or res == 120)
|
|
3419
|
+
@msg: "Invalid resolution for JPEG. Using 300 dpi."
|
|
3420
|
+
endif
|
|
3421
|
+
system 'eps_jpg$' -r'res' -sOutputFile="'dst$'.jpg" "'src$'.eps"
|
|
3422
|
+
endif
|
|
3423
|
+
endfor
|
|
3424
|
+
if (index (graphics_format$, "multipage"))
|
|
3425
|
+
; use wildcard expansion by pdftk in order to avoid command line buffer overflow
|
|
3426
|
+
;pdf_out$ = replace_regex$ (filespec$, "[0-9\*]*\.eps", "\.pdf", 1)
|
|
3427
|
+
pdf_out$ = outdir$ + "/" + outfname$
|
|
3428
|
+
pdf_out$ = replace$ (pdf_out$, "//", "/", 0)
|
|
3429
|
+
pdf_in$ = replace_regex$ (filespec$, "\.eps", "\.pdf", 1)
|
|
3430
|
+
printline convert_eps: system pdftk 'pdf_in$' cat output 'pdf_out$'
|
|
3431
|
+
system pdftk 'pdf_in$' cat output 'pdf_out$'
|
|
3432
|
+
endif
|
|
3433
|
+
select Strings filelist
|
|
3434
|
+
Remove
|
|
3435
|
+
@msg: "Batch command convert_eps Ready"
|
|
3436
|
+
endproc
|
|
3437
|
+
|
|
3438
|
+
|
|
3439
|
+
procedure corpus_open: .cmda$
|
|
3440
|
+
@initialization_main
|
|
3441
|
+
; Load signal files and corresponding TextGrid files. Specify filenames as wav files regular expression.
|
|
3442
|
+
@option_get: "word", "file=", "inputfname", ".*.wav$", .cmda$
|
|
3443
|
+
|
|
3444
|
+
if (variableExists ("corpus_home$"))
|
|
3445
|
+
if (variableExists ("corpus_subdir_sound$"))
|
|
3446
|
+
inputfname$ = corpus_home$ + corpus_subdir_sound$ + inputfname$
|
|
3447
|
+
else
|
|
3448
|
+
inputfname$ = corpus_home$ + inputfname$
|
|
3449
|
+
endif
|
|
3450
|
+
endif
|
|
3451
|
+
@files_get_regex: "filelistID", inputfname$
|
|
3452
|
+
selectObject: filelistID
|
|
3453
|
+
.nrofFiles = Get number of strings
|
|
3454
|
+
if (.nrofFiles == 0)
|
|
3455
|
+
@msg: "No input files found for <'inputfname$'>"
|
|
3456
|
+
endif
|
|
3457
|
+
for .j to .nrofFiles
|
|
3458
|
+
selectObject: filelistID
|
|
3459
|
+
.fname$ = Get string: .j
|
|
3460
|
+
@fname_parts: .fname$
|
|
3461
|
+
.basename$ = result2$
|
|
3462
|
+
Open long sound file: .fname$
|
|
3463
|
+
.dur = Get total duration
|
|
3464
|
+
@msg: "'.basename$''tab$''.dur:3'"
|
|
3465
|
+
.tg$ = .basename$ + ".TextGrid"
|
|
3466
|
+
if (variableExists ("corpus_home$"))
|
|
3467
|
+
.tg$ = corpus_home$ + .basename$ + ".TextGrid"
|
|
3468
|
+
if (variableExists ("corpus_subdir_tg$"))
|
|
3469
|
+
.tg$ = corpus_home$ + corpus_subdir_tg$ + .basename$ + ".TextGrid"
|
|
3470
|
+
endif
|
|
3471
|
+
endif
|
|
3472
|
+
if (fileReadable (.tg$))
|
|
3473
|
+
Read from file: .tg$
|
|
3474
|
+
endif
|
|
3475
|
+
endfor
|
|
3476
|
+
removeObject: filelistID
|
|
3477
|
+
@msg: "Batch command corpus_open Ready"
|
|
3478
|
+
endproc
|
|
3479
|
+
|
|
3480
|
+
|
|
3481
|
+
procedure validate_tier: .cmda$
|
|
3482
|
+
@initialization_main
|
|
3483
|
+
@option_get: "word", "file=", "inputfname", ".*.TextGrid$", .cmda$
|
|
3484
|
+
; call option_get word tiertype= tiertype "phoneme" '.cmda$'
|
|
3485
|
+
|
|
3486
|
+
if (variableExists ("corpus$") and length (corpus$) and variableExists ("corpus_home$"))
|
|
3487
|
+
if (variableExists ("corpus_subdir_tg$"))
|
|
3488
|
+
inputfname$ = corpus_home$ + corpus_subdir_tg$ + inputfname$
|
|
3489
|
+
else
|
|
3490
|
+
inputfname$ = corpus_home$ + inputfname$
|
|
3491
|
+
endif
|
|
3492
|
+
endif
|
|
3493
|
+
@files_get_regex: "filelistID", inputfname$
|
|
3494
|
+
selectObject: filelistID
|
|
3495
|
+
.nrofFiles = Get number of strings
|
|
3496
|
+
if (.nrofFiles == 0)
|
|
3497
|
+
@msg: "No input files found for <'inputfname$'>"
|
|
3498
|
+
endif
|
|
3499
|
+
for .j to .nrofFiles
|
|
3500
|
+
selectObject: filelistID
|
|
3501
|
+
.tg$ = Get string: .j
|
|
3502
|
+
if (fileReadable (.tg$))
|
|
3503
|
+
@validate_phoneme_tier: .tg$
|
|
3504
|
+
endif
|
|
3505
|
+
endfor
|
|
3506
|
+
removeObject: filelistID
|
|
3507
|
+
@msg: "Batch command validate_tier Ready"
|
|
3508
|
+
endproc
|
|
3509
|
+
|
|
3510
|
+
|
|
3511
|
+
procedure find_nucleus: .type$, .t1, .t2, .forward
|
|
3512
|
+
; Find nucleus of given <.type$> in time range <.t1>..<.t2>
|
|
3513
|
+
; <.type$> "a" = valid_nucleus, "-" = rejected nucleus
|
|
3514
|
+
; <.forward> search from .t1 to .t2, otherwise from .t2 to .t1
|
|
3515
|
+
; Return index of nucleus (in nucleus_tier) in variable <result> or 0 if not found
|
|
3516
|
+
result = 0
|
|
3517
|
+
selectObject: nucleiID
|
|
3518
|
+
.endt = Get end time
|
|
3519
|
+
.startt = Get start time
|
|
3520
|
+
.endt = min (.endt, .t2)
|
|
3521
|
+
.startt = max (.startt, .t1)
|
|
3522
|
+
if (.forward)
|
|
3523
|
+
.t = .t1
|
|
3524
|
+
else
|
|
3525
|
+
.t = .t2 - 0.002
|
|
3526
|
+
endif
|
|
3527
|
+
while ((.forward and .t < .endt) or (.forward == 0 and .t > .startt) and result == 0)
|
|
3528
|
+
.i = Get interval at time: nucleus_tier, .t
|
|
3529
|
+
.s$ = Get label of interval: nucleus_tier, .i
|
|
3530
|
+
if (.s$ == .type$)
|
|
3531
|
+
result = .i
|
|
3532
|
+
endif
|
|
3533
|
+
if (.forward)
|
|
3534
|
+
.t = Get end time of interval: nucleus_tier, .i
|
|
3535
|
+
else
|
|
3536
|
+
.t = .startt
|
|
3537
|
+
if (.i > 1)
|
|
3538
|
+
.t = Get start time of interval: nucleus_tier, .i-1
|
|
3539
|
+
endif
|
|
3540
|
+
endif
|
|
3541
|
+
endwhile
|
|
3542
|
+
endproc
|
|
3543
|
+
|
|
3544
|
+
|
|
3545
|
+
procedure detect_hesitations: .t1a, .t2a, .method$
|
|
3546
|
+
# Detect hesitations, either from hesitation tier in TextGrid of from acoustic cues and phoneme identity.
|
|
3547
|
+
# <.method>
|
|
3548
|
+
# "phon+pros" use phoneme label and prosodic features (phon_dur >= 0.3 && intrasyll <= 0 && intrasyll > -3 ST)
|
|
3549
|
+
# phoneme is "\oe", "\o/", "9", "@" or "2"
|
|
3550
|
+
# "words" use word tier from annotation TextGrid, where hesitations are labeled "euh" or "eh"
|
|
3551
|
+
# "all" combine above methods
|
|
3552
|
+
# "none" don't attempt detection
|
|
3553
|
+
# "annotation" use hesitation tier from annotation TextGrid, where a hesitation is marked as "H"
|
|
3554
|
+
@debug_msg: "detect_hesitations: entry, method='.method$'"
|
|
3555
|
+
|
|
3556
|
+
if segfile_available and (.method$ = "phon+pros" or .method$ = "words" or .method$ = "all")
|
|
3557
|
+
@intervals_from_time_range: nucleiID, nucleus_tier, .t1a, .t2a, "first_interval", "last_interval"
|
|
3558
|
+
for .j from first_interval to last_interval
|
|
3559
|
+
selectObject: nucleiID
|
|
3560
|
+
@is_nucleus: .j
|
|
3561
|
+
if (result)
|
|
3562
|
+
.hesit = 0 ; default = no hesitation detected
|
|
3563
|
+
.t1 = Get start time of interval: nucleus_tier, .j
|
|
3564
|
+
.t2 = Get end time of interval: nucleus_tier, .j
|
|
3565
|
+
.tmid = .t1 + (.t2-.t1)/2
|
|
3566
|
+
if ((.method$ = "phon+pros" or .method$ = "all") and phones_available)
|
|
3567
|
+
@interval_from_time: nucleiID, phone_tier, .tmid, "i"
|
|
3568
|
+
.label$ = Get label of interval: phone_tier, i
|
|
3569
|
+
.label$ = replace_regex$(.label$, """", "", 0) ; remove double quotes from match string
|
|
3570
|
+
.px1 = Get start time of interval: phone_tier, i
|
|
3571
|
+
.px2 = Get end time of interval: phone_tier, i
|
|
3572
|
+
.phon_dur = .px2-.px1
|
|
3573
|
+
.s$ = Get label of interval: pointer_tier, .j
|
|
3574
|
+
.pj = '.s$'
|
|
3575
|
+
selectObject: nucldatID
|
|
3576
|
+
.intST = Get value: .pj, j_intrasyl
|
|
3577
|
+
if (index ("=\o/=\oe=9=2=@=", "='.label$'=")
|
|
3578
|
+
... and .phon_dur >= 0.3 and .intST <= 0 and abs(.intST) < 3)
|
|
3579
|
+
.hesit = 1
|
|
3580
|
+
endif
|
|
3581
|
+
endif
|
|
3582
|
+
if ((.method$ = "words" or .method$ = "all") and word_tier_in > 0)
|
|
3583
|
+
@interval_from_time: segmentationID, word_tier_in, .tmid, "jw"
|
|
3584
|
+
.label$ = Get label of interval: word_tier_in, jw
|
|
3585
|
+
if (index ("=euh=eh=", "='.label$'="))
|
|
3586
|
+
.hesit = 1
|
|
3587
|
+
endif
|
|
3588
|
+
endif
|
|
3589
|
+
if ((.method$ = "annotation" or .method$ = "all") and hesitation_available)
|
|
3590
|
+
@interval_from_time: nucleiID, hesitation_tier, .tmid, "jw"
|
|
3591
|
+
.label$ = Get label of interval: hesitation_tier, jw
|
|
3592
|
+
if (.label$ == "H")
|
|
3593
|
+
.hesit = 1
|
|
3594
|
+
endif
|
|
3595
|
+
endif
|
|
3596
|
+
if (.hesit)
|
|
3597
|
+
; @msg: "detect_hesitations: hesitation detected at '.t1:3', method='.method$'"
|
|
3598
|
+
@interval_from_time: nucleiID, hesitation_tier, .tmid, "i"
|
|
3599
|
+
Set interval text: hesitation_tier, i, "H"
|
|
3600
|
+
endif
|
|
3601
|
+
endif
|
|
3602
|
+
endfor
|
|
3603
|
+
hesitation_available = 1
|
|
3604
|
+
endif
|
|
3605
|
+
|
|
3606
|
+
; Copy info in hesitation tier to column of nucldatID
|
|
3607
|
+
selectObject: nucldatID
|
|
3608
|
+
.nrows = Get number of rows
|
|
3609
|
+
for .j to .nrows
|
|
3610
|
+
.t1n = Get value: .j, j_nucl_t1
|
|
3611
|
+
.t2n = Get value: .j, j_nucl_t2
|
|
3612
|
+
selectObject: nucleiID
|
|
3613
|
+
.i = Get interval at time: hesitation_tier, .t1n+(.t2n-.t1n)/2
|
|
3614
|
+
.label$ = Get label of interval: hesitation_tier, .i
|
|
3615
|
+
.hesit = 0
|
|
3616
|
+
if (.label$ = "H")
|
|
3617
|
+
.hesit = 1
|
|
3618
|
+
endif
|
|
3619
|
+
selectObject: nucldatID
|
|
3620
|
+
Set value: .j, j_hesitation, .hesit
|
|
3621
|
+
endfor
|
|
3622
|
+
@debug_msg: "detect_hesitations: exit"
|
|
3623
|
+
endproc
|
|
3624
|
+
|
|
3625
|
+
|
|
3626
|
+
procedure process_cprom: .grid
|
|
3627
|
+
@tier_number_by_name: .grid, "^ss$"
|
|
3628
|
+
if (result)
|
|
3629
|
+
Remove tier: result
|
|
3630
|
+
endif
|
|
3631
|
+
@tier_number_by_name: .grid, "phones"
|
|
3632
|
+
if (result) ; Remove labels for pauses ("_"), respirations ("*"), etc. "+"
|
|
3633
|
+
Replace interval text... result 0 0 "[\+\*_#%]" "" Regular Expressions
|
|
3634
|
+
endif
|
|
3635
|
+
@tier_number_by_name: .grid, "syll"
|
|
3636
|
+
.syll_tier = result
|
|
3637
|
+
if (.syll_tier) ; Simplify complex labels for pauses, respirations, etc. to "_"
|
|
3638
|
+
Replace interval text... .syll_tier 0 0 "[\+\*_#]+" "_" Regular Expressions
|
|
3639
|
+
endif
|
|
3640
|
+
; Create prominence and hesitation tiers from "delivery" tier in TextGrid <.grid>
|
|
3641
|
+
if (0)
|
|
3642
|
+
@tier_number_by_name: .grid, "delivery"
|
|
3643
|
+
if (result)
|
|
3644
|
+
.deliv_tier = result
|
|
3645
|
+
.nt = Get number of tiers
|
|
3646
|
+
.prom_tier = .nt + 1
|
|
3647
|
+
.hes_tier = .nt + 2
|
|
3648
|
+
.ni = Get number of intervals: .deliv_tier
|
|
3649
|
+
Duplicate tier: .syll_tier, .prom_tier, "prom"
|
|
3650
|
+
Duplicate tier: .syll_tier, .hes_tier, "hes"
|
|
3651
|
+
for .j to .ni
|
|
3652
|
+
.label$ = Get label of interval: .deliv_tier, .j
|
|
3653
|
+
if (index (.label$, "P"))
|
|
3654
|
+
.s$ = "S"
|
|
3655
|
+
elsif (index (.label$, "p"))
|
|
3656
|
+
.s$ = "W"
|
|
3657
|
+
else
|
|
3658
|
+
.s$ = "0"
|
|
3659
|
+
endif
|
|
3660
|
+
Set interval text: .prom_tier, .j, .s$
|
|
3661
|
+
if (index (.label$, "z"))
|
|
3662
|
+
.s$ = "H"
|
|
3663
|
+
else
|
|
3664
|
+
.s$ = ""
|
|
3665
|
+
endif
|
|
3666
|
+
Set interval text: .hes_tier, .j, .s$
|
|
3667
|
+
endfor
|
|
3668
|
+
endif
|
|
3669
|
+
endif
|
|
3670
|
+
@tier_number_by_name: .grid, "prom"
|
|
3671
|
+
if (result) ; Remove or rename labels
|
|
3672
|
+
Replace interval text... result 0 0 "0" "" Regular Expressions
|
|
3673
|
+
Replace interval text... result 0 0 "W" "p" Regular Expressions
|
|
3674
|
+
Replace interval text... result 0 0 "S" "P" Regular Expressions
|
|
3675
|
+
endif
|
|
3676
|
+
@tier_number_by_name: .grid, "rg"
|
|
3677
|
+
if (result) ; Remove labels
|
|
3678
|
+
Replace interval text... result 0 0 ".*" "" Regular Expressions
|
|
3679
|
+
endif
|
|
3680
|
+
endproc
|
|
3681
|
+
|
|
3682
|
+
|
|
3683
|
+
procedure lpa_sampa: .grid, .tiername$
|
|
3684
|
+
# Convert LPA (Limsi Phonetic Alphabet) to SAMPA
|
|
3685
|
+
@tier_number_by_name: .grid, .tiername$
|
|
3686
|
+
.tier = result
|
|
3687
|
+
if (result)
|
|
3688
|
+
; first 3 lines deal with pause, incomprehensible sound and respiration
|
|
3689
|
+
Replace interval texts: .tier, 0, 0, "\.", "_", "Regular Expressions"
|
|
3690
|
+
Replace interval texts: .tier, 0, 0, "\*", "", "Regular Expressions"
|
|
3691
|
+
Replace interval texts: .tier, 0, 0, "H", "", "Regular Expressions"
|
|
3692
|
+
; next line deals with intervals containing several non speech elements
|
|
3693
|
+
Replace interval texts: .tier, 0, 0, "_+", "_", "Regular Expressions"
|
|
3694
|
+
Replace interval texts: .tier, 0, 0, "@", "2", "Regular Expressions"
|
|
3695
|
+
Replace interval texts: .tier, 0, 0, "x", "@", "Regular Expressions"
|
|
3696
|
+
Replace interval texts: .tier, 0, 0, "h", "H", "Regular Expressions"
|
|
3697
|
+
Replace interval texts: .tier, 0, 0, "I", "e~", "Regular Expressions"
|
|
3698
|
+
Replace interval texts: .tier, 0, 0, "O", "o~", "Regular Expressions"
|
|
3699
|
+
Replace interval texts: .tier, 0, 0, "A", "a~", "Regular Expressions"
|
|
3700
|
+
Replace interval texts: .tier, 0, 0, "c", "O", "Regular Expressions"
|
|
3701
|
+
Replace interval texts: .tier, 0, 0, "X", "9", "Regular Expressions"
|
|
3702
|
+
Replace interval texts: .tier, 0, 0, "r", "R", "Regular Expressions"
|
|
3703
|
+
Replace interval texts: .tier, 0, 0, "N", "J", "Regular Expressions"
|
|
3704
|
+
endif
|
|
3705
|
+
endproc
|
|
3706
|
+
|
|
3707
|
+
|
|
3708
|
+
procedure tier_trim_nucleus: .grid, .tier, .dtleft, .dtright
|
|
3709
|
+
; Change boundaries of nucleus <.tier>, trimming left and right by amount <.dtleft> and <.dtright>, when duration remaining nucleus > 0.1 s
|
|
3710
|
+
@debug_msg: "tier_trim_nucleus: entry, .tier='.tier', .dtleft='.dtleft:3'"
|
|
3711
|
+
selectObject: .grid
|
|
3712
|
+
.n = Get number of tiers
|
|
3713
|
+
if (.tier <= .n)
|
|
3714
|
+
.ni = Get number of intervals: .tier
|
|
3715
|
+
.i = 1
|
|
3716
|
+
while (.i <= .ni)
|
|
3717
|
+
.s$ = Get label of interval: .tier, .i
|
|
3718
|
+
if (.s$ = "a")
|
|
3719
|
+
.t1 = Get starting point: .tier, .i
|
|
3720
|
+
.t2 = Get end point: .tier, .i
|
|
3721
|
+
if (.dtleft > 0 and .t2-.t1 > .dtleft) ; left side
|
|
3722
|
+
.t = min (.t1 + .dtleft, .t2 - 0.1)
|
|
3723
|
+
if (.t > .t1)
|
|
3724
|
+
Set interval text: .tier, .i, ""
|
|
3725
|
+
Insert boundary: .tier, .t
|
|
3726
|
+
Set interval text: .tier, .i+1, .s$
|
|
3727
|
+
.ni += 1
|
|
3728
|
+
.t1 = .t
|
|
3729
|
+
.i += 1
|
|
3730
|
+
endif
|
|
3731
|
+
endif
|
|
3732
|
+
if (.dtright > 0 and .t2-.t1 > .dtright) ; right side
|
|
3733
|
+
.t = max (.t1 + 0.1, .t2 - .dtright)
|
|
3734
|
+
if (.t < .t2)
|
|
3735
|
+
Set interval text: .tier, .i, ""
|
|
3736
|
+
Insert boundary: .tier, .t
|
|
3737
|
+
Set interval text: .tier, .i, .s$
|
|
3738
|
+
.ni += 1
|
|
3739
|
+
.i += 1
|
|
3740
|
+
endif
|
|
3741
|
+
endif
|
|
3742
|
+
endif
|
|
3743
|
+
.i += 1
|
|
3744
|
+
endwhile
|
|
3745
|
+
endif
|
|
3746
|
+
endproc
|
|
3747
|
+
|
|
3748
|
+
|
|
3749
|
+
procedure textgrid_disable_textstyle: .gridID
|
|
3750
|
+
selectObject: .gridID
|
|
3751
|
+
.n = Get number of tiers
|
|
3752
|
+
for .tier to .n
|
|
3753
|
+
@tier_disable_textstyle: .gridID, .tier
|
|
3754
|
+
endfor
|
|
3755
|
+
endproc
|
|
3756
|
+
|
|
3757
|
+
|
|
3758
|
+
procedure tier_disable_textstyle: .gridID, .tier
|
|
3759
|
+
selectObject: .gridID
|
|
3760
|
+
.n = Get number of intervals: .tier
|
|
3761
|
+
for .interval to .n
|
|
3762
|
+
.label$ = Get label of interval: .tier, .interval
|
|
3763
|
+
.label$ = replace$ (.label$, "_", "\_ ", 0) ; avoid interpretation of underscore as subscript
|
|
3764
|
+
.label$ = replace$ (.label$, "%", "\% ", 0) ; avoid interpretation of percent as italic text style
|
|
3765
|
+
Set interval text: .tier, .interval, .label$
|
|
3766
|
+
endfor
|
|
3767
|
+
endproc
|
|
3768
|
+
|
|
3769
|
+
|
|
3770
|
+
procedure tier_merge_intervals_except: .gridID, .tier, .except$
|
|
3771
|
+
; Merge adjacent intervals other than those with label <.except$>
|
|
3772
|
+
selectObject: .gridID
|
|
3773
|
+
.n = Get number of intervals: .tier
|
|
3774
|
+
.i = 1
|
|
3775
|
+
while (.i < .n)
|
|
3776
|
+
.label1$ = Get label of interval: .tier, .i
|
|
3777
|
+
.label2$ = Get label of interval: .tier, .i+1
|
|
3778
|
+
if (.label1$ <> .except$ and .label2$ <> .except$)
|
|
3779
|
+
Remove right boundary: .tier, .i
|
|
3780
|
+
; Set interval text: .tier, .i, ""
|
|
3781
|
+
.n -= 1
|
|
3782
|
+
else
|
|
3783
|
+
.i += 1
|
|
3784
|
+
endif
|
|
3785
|
+
endwhile
|
|
3786
|
+
endproc
|
|
3787
|
+
|