Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Large-scale analysis of Spanish /s/-lenition using audiobooks Neville Ryant 1 and Mark Liberman 2 Linguistic Data Consortium, USA 1 nryant@gmail.com , 2 markyliberman@gmail.com September 5, 2016
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Outline Introduction 1
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Outline Introduction 1 Audiobooks 2
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Outline Introduction 1 Audiobooks 2 Acoustic measurements 3
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Outline Introduction 1 Audiobooks 2 Acoustic measurements 3 Nontraditional acoustic measurement 4
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Outline Introduction 1 Audiobooks 2 Acoustic measurements 3 Nontraditional acoustic measurement 4 Future directions 5
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Outline Introduction 1 Audiobooks 2 Acoustic measurements 3 Nontraditional acoustic measurement 4 Future directions 5
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Spanish /s/-lenition Definition Spanish /s/-lenition is the weakening of /s/ in syllable-final position to one of the following variants: [s] [h] [z] (before a voiced stop) deletion
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Example: [s] (Venezuelan)
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Example: [h] (Chilean)
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Example: [z] (Mexican)
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Example: Deletion (Venezuelan)
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Prevalence sC s#C s# { sil } s#V1 s#V0 Argentina 88 89 22 7 6 Chile 93 96 36 10 22 Cuba 97 98 39 53 90 Dominican Republic 91 98 65 49 85 El Salvador 45 90 14 37 72 Honduras 41 89 28 15 39 Nicaragua 87 98 64 72 93 Panama 95 95 66 53 80 Paraguay 86 98 17 53 85 Puerto Rico 94 96 54 55 84 Venezuela 95 49 98 41 89 Percent aspiration/deletion in conversational speech (Lipski, 1983; Lipski, 1985)
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Prior work Overview Conditioning factors One of the most widely-studied Sex sociolinguistic variables Age Wide range of conditioning factors Social class (phonetic and sociolinguistic) Speech style examined Phonetic context Grammatical category Lexical frequency ...
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Prior work: Limitations Methodological Transcription is inherently subjective Transcriber error and biases In reality /s/-lenition is a gradient process → any partitioning of the space of outcomes is wrong Logistical Segmentation and measurement typically manual → expensive and slow Datasets typically number only in hundreds or thousands of measurement But, cross product of linguistic factors of interest is large and requires correspondingly large datasets of tens or even hundreds of thousands of observations
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Scaling via audiobooks Pro Con Easy to obtain Currently, number of LibriVox (free) distinct speakers is limited Audible (pay) → very in-depth data for Large-scale: single book may yield > 40,000 single speakers observations Limited to single speech style: read speech Doesn’t require expensive transcriptions → cheap segmentations
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Audiobooks provide scale: LibriVox Works Hours Works (2015) Hours (2015) English 8,516 50,591 914 5,170 German 482 2,805 35 197 Dutch 180 2,100 9 102 French 163 1,057 5 33 Spanish 103 638 15 112
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions LibriVox Works Hours Works (2015) Hours (2015) English 8,516 50,591 914 5,170 German 482 2,805 35 197 Dutch 180 2,100 9 102 French 163 1,057 5 33 Spanish 103 638 15 112
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Outline Introduction 1 Audiobooks 2 Acoustic measurements 3 Nontraditional acoustic measurement 4 Future directions 5
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Corpora Overview Seven audiobooks encompassing four varieties of Spanish Eight speakers Audio from LibriVox or Audible Peninsular Spanish Los Pazos de Ulloa by Emilia Pardo Baz´ an Historietas Nacionales by Pedro Antonio de Alarc´ on y Ariza El 19 de Marzo y el 2 de Mayo by Benito P´ erez Gals´ o
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Corpora Argentinian Cien A˜ nos de Soledad by Gabriel Garc´ ıa M´ arquez La Isla Del Tesoro by Robert Louis Stevenson (translated by Manuel Caballero) Chilean La Casa de los Esperitus by Isabel Allende Mexican interior Angelina by Rafael Delgado
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Corpora Hours Speakers Words /s/ Peninsular 24.24 3 204,448 85,645 Chilean 16.98 2 165,620 66,726 Argentinian 28.72 2 226,489 88,246 Mexican 10.15 1 92,811 37,493 TOTAL 80.09 8 689,368 278,110
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Alignment MFCCs (5-1-5 Frames) Acoustic features 13 MFCCs + deltas + delta-deltas 512 Rectified Linear Units Per-utterance cesptral mean-variance normalization 512 Rectified Linear Units 10 ms step, 25 ms analysis window 11-frame context window 512 Rectified Linear Units HMM topology speech: 3-state Bakis 512 Rectified Linear Units non-speech: 5-state w/ skips boundary: 1-state SoftMax
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Alignment MFCCs (5-1-5 Frames) Training 512 Rectified Linear Units All turns from West Point-Heroico corpus of Mexican Spanish 512 Rectified Linear Units CALLHOME Spanish pronunciation dictionary OOV pronunciations generated via 512 Rectified Linear Units grapheme-to-phoneme transducer trained on CALLHOME 512 Rectified Linear Units Vowel stress differentiated SoftMax
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Outline Introduction 1 Audiobooks 2 Acoustic measurements 3 Nontraditional acoustic measurement 4 Future directions 5
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Acoustic measurements Approach Contexts Extract acoustic features every 5 ms word-final before pause Average across frames within each word-final before vowel /s/ segment before voiced stop Compare segment-level averages before nasal across phonetic contexts before voiceless stop Differences across contexts expected to be more pronounced in dialects known for being leniting (Argentinian and Chilean)
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Acoustic measurements Spectral centroid Center of mass of decibel power spectrum, viewed as density Excluded frequencies below 1 kHz POV Derived from output of Kadi pitch tracker Random spot checks done to verify sanity on these materials Duration Duration of /s/ segment (seconds) Derived from forced alignment boundaries of /s/
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Expected patterns Spectral centroid Highest in environments encouraging retention (word-final before vowel or pause) Lowest in environments typically associated with weakening; in particular, before voiced consonant POV Lowest in environments encouraging retention (word-final before vowel or pause) Highest in environments typically associated with weakening to [z] (before voiced stop or nasal)
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Expected patterns Duration Highest before pause and, to lesser degree, word-final before vowel Lowest in environments typically associated with weakening to [z] (before voiced stop or nasal)
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Spectral Centroid
Introduction Audiobooks Acoustic measurements Nontraditional acoustic measurement Future directions Probability-of-voicing
Recommend
More recommend
