The Voice Source in Speech Production: from Models to Applications

<!DOCTYPE html> <html lang="en" class="no-js"> <head> <meta charset="UTF-8"> <meta name="viewport" content="width=device-width, initial-scale=1.0, user-scalable=yes"> <title>The Voice Source in Speech Production: from Models to Applications</title> <meta id="meta-title" property="citation_title" content="The Voice Source in Speech Production: from Models to Applications"/> <meta id="og-title" property="og:title" content="The Voice Source in Speech Production: from Models to Applications"/> <meta name="twitter:widgets:autoload" content="off"/> <meta name="twitter:dnt" content="on"/> <meta name="twitter:widgets:csp" content="on"/> <meta name="google-site-verification" content="lQbRRf0vgPqMbnbCsgELjAjIIyJjiIWo917M7hBshvI"/> <meta id="meta-abstract" name="citation_abstract" content="The voice source contains important lexical and non-lexical information. The non-lexical information can convey, for example, prosodic events, emotional status, as well as cues pertaining to the uniqueness of the speaker's voice. A better understanding, and eventually a better model of the voice source, would benefit various speech applications, such as speech recognition, speech synthesis, speaker identification, age/gender classification, as well as clinical assessments.This dissertation has three main goals. The first is to better understand the voice source through analyzing images of the vocal folds using laryngeal high-speed videoendoscopy (HSV) recordings. A new automatic method is proposed to compactly summarize the overall spatial synchronization pattern of vocal fold vibration for the entire laryngeal area from HSV data. Additionally, a new measure is proposed to adequately capture perceptually-important variations in glottal area pulse shapes, which are extracted from HSV data.The second goal is to study the acoustic consequence of a physiological vocal-fold vibration pattern---the glottal gap effect, and apply our findings to a gender classification task of children's voices. Voice source related measures are found to improve classification accuracy, especially for younger (10-15 year old) speakers.The third goal is to propose new voice source models and evaluate them in different applications. In the first application, a new source model and a noise-robust automatic source estimation algorithm are proposed to estimate the voice source from speech signals. Results in both clean and noisy conditions show that the proposed model and algorithm are robust in accurately estimating the voice source signal. The second application is to use the proposed source model for vowel synthesis. Perceptual listening experiments show that the proposed model provides a better perceptual match to the target voice than do traditional models."/> <meta id="meta-description" name="description" content="Author(s): Chen, Gang | Advisor(s): ALWAN, ABEER | Abstract: The voice source contains important lexical and non-lexical information. The non-lexical information can convey, for example, prosodic events, emotional status, as well as cues pertaining to the uniqueness of the speaker's voice. A better understanding, and eventually a better model of the voice source, would benefit various speech applications, such as speech recognition, speech synthesis, speaker identification, age/gender classification, as well as clinical assessments.This dissertation has three main goals. The first is to better understand the voice source through analyzing images of the vocal folds using laryngeal high-speed videoendoscopy (HSV) recordings. A new automatic method is proposed to compactly summarize the overall spatial synchronization pattern of vocal fold vibration for the entire laryngeal area from HSV data. Additionally, a new measure is proposed to adequately capture perceptually-important variations in glottal area pulse shapes, which are extracted from HSV data.The second goal is to study the acoustic consequence of a physiological vocal-fold vibration pattern---the glottal gap effect, and apply our findings to a gender classification task of children's voices. Voice source related measures are found to improve classification accuracy, especially for younger (10-15 year old) speakers.The third goal is to propose new voice source models and evaluate them in different applications. In the first application, a new source model and a noise-robust automatic source estimation algorithm are proposed to estimate the voice source from speech signals. Results in both clean and noisy conditions show that the proposed model and algorithm are robust in accurately estimating the voice source signal. The second application is to use the proposed source model for vowel synthesis. Perceptual listening experiments show that the proposed model provides a better perceptual match to the target voice than do traditional models."/> <meta id="og-description" name="og:description" content="Author(s): Chen, Gang | Advisor(s): ALWAN, ABEER | Abstract: The voice source contains important lexical and non-lexical information. The non-lexical information can convey, for example, prosodic events, emotional status, as well as cues pertaining to the uniqueness of the speaker's voice. A better understanding, and eventually a better model of the voice source, would benefit various speech applications, such as speech recognition, speech synthesis, speaker identification, age/gender classification, as well as clinical assessments.This dissertation has three main goals. The first is to better understand the voice source through analyzing images of the vocal folds using laryngeal high-speed videoendoscopy (HSV) recordings. A new automatic method is proposed to compactly summarize the overall spatial synchronization pattern of vocal fold vibration for the entire laryngeal area from HSV data. Additionally, a new measure is proposed to adequately capture perceptually-important variations in glottal area pulse shapes, which are extracted from HSV data.The second goal is to study the acoustic consequence of a physiological vocal-fold vibration pattern---the glottal gap effect, and apply our findings to a gender classification task of children's voices. Voice source related measures are found to improve classification accuracy, especially for younger (10-15 year old) speakers.The third goal is to propose new voice source models and evaluate them in different applications. In the first application, a new source model and a noise-robust automatic source estimation algorithm are proposed to estimate the voice source from speech signals. Results in both clean and noisy conditions show that the proposed model and algorithm are robust in accurately estimating the voice source signal. The second application is to use the proposed source model for vowel synthesis. Perceptual listening experiments show that the proposed model provides a better perceptual match to the target voice than do traditional models."/> <meta id="og-image" property="og:image" content="https://escholarship.org/images/escholarship-facebook2.jpg"/> <meta id="og-image-width" property="og:image:width" content="1242"/> <meta id="og-image-height" property="og:image:height" content="1242"/> <meta id="meta-author0" name="citation_author" content="Chen, Gang"/> <meta id="meta-publication_date" name="citation_publication_date" content="2014"/> <meta id="meta-dissertation_institution" name="citation_dissertation_institution" content="UCLA"/> <meta id="meta-online_date" name="citation_online_date" content="2014-10-23"/> <meta id="meta-pdf_url" name="citation_pdf_url" content="https://escholarship.org/content/qt87p8p8p9/qt87p8p8p9.pdf?t=ndw6ls"/> <link rel="canonical" href="https://escholarship.org/uc/item/87p8p8p9"/> <link rel="stylesheet" href="/css/main-fa87d2fbdbef47d2.css"> <link rel="resource" type="application/l10n" href="/node_modules/pdfjs-embed2/dist/locale/locale.properties"> <noscript><style> .jsonly { display: none } </style></noscript>   <script type="text/plain" data-type="application/javascript" data-name="matomo"> var _paq = window._paq = window._paq || []; /* tracker methods like "setCustomDimension" should be called before "trackPageView" */ _paq.push(['trackPageView']); _paq.push(['enableLinkTracking']); (function() { var u="//matomo.cdlib.org/"; _paq.push(['setTrackerUrl', u+'matomo.php']); _paq.push(['setSiteId', '7']); var d=document, g=d.createElement('script'), s=d.getElementsByTagName('script')[0]; g.async=true; g.src=u+'matomo.js'; s.parentNode.insertBefore(g,s); console.log('*** MATOMO LOADED ***'); })(); </script>  </head> <body> <div id="main"><div data-reactroot=""><div class="body"><a href="#maincontent" class="c-skipnav">Skip to main content</a><div><div><header id="#top" class="c-header"><a class="c-header__logo2" href="/"><picture><source srcSet="/images/logo_eschol-small.svg" media="(min-width: 870px)"/><img src="/images/logo_eschol-mobile.svg" alt="eScholarship"/></picture><div class="c-header__logo2-tagline">Open Access Publications from the University of California</div></a><div class="c-header__search"><form class="c-search2"><div class="c-search2__inputs"><div class="c-search2__form"><label class="c-search2__label" for="global-search">Search eScholarship</label><input type="search" name="q" id="global-search" class="c-search2__field" placeholder="Search" autoComplete="off" autoCapitalize="off"/></div><fieldset><legend>Refine Search</legend><div class="c-search2__refine"><input type="radio" id="c-search2__refine-eschol" name="searchType" value="eScholarship" checked=""/><label for="c-search2__refine-eschol">All of eScholarship</label><input type="radio" id="c-search2__refine-campus" name="searchType" value="ucla_etd"/><label for="c-search2__refine-campus">This Series</label><input type="hidden" name="searchUnitType" value="series"/></div></fieldset></div><button type="submit" class="c-search2__submit-button" aria-label="search"></button><button type="button" class="c-search2__search-close-button" aria-label="close search field"></button></form></div><button class="c-header__search-open-button" aria-label="open search field"></button></header></div><div class=" c-subheader has-banner is-wide has-campus-label is-black "><a class="c-subheader__title" href="/uc/ucla"><h1>UCLA</h1></a><a class="c-subheader__banner" href="/uc/ucla"><img src="/cms-assets/c16e7a631956312b6b0ca3c58ac21c61c80a2176d91f50236b191969f594f0b1" alt="UCLA Electronic Theses and Dissertations banner"/></a><a class="c-subheader__campus" href="/uc/ucla">UCLA</a><div id="wizardModalBase" class="c-subheader__buttons"><button id="wizardlyDeposit" class="o-button__3">Deposit</button><div class="c-modal"></div><button id="wizardlyManage" class="o-button__3">Manage<span class="c-subheader__button-fragment">Submissions</span></button><div class="c-modal"></div></div></div><div class="c-navbar"><nav class="c-nav"><details open="" class="c-nav__main"><summary class="c-nav__main-button">Menu</summary><ul class="c-nav__main-items"><li><a href="/uc/ucla/about">About eScholarship</a></li><li><a href="/uc/ucla/ucoapolicies">UC Open Access Policies</a></li><li><a href="/ucla/journals">Journals</a></li><li><a href="/ucla/units">Academic Units</a></li></ul></details></nav><div class="c-socialicons"><a href="/rss/unit/ucla_etd"><img src="/images/logo_rss-circle-black.svg" alt="RSS"/></a></div></div><nav class="c-breadcrumb"><ul><li><a href="/">eScholarship</a></li><li><a href="/uc/ucla">UCLA</a></li><li><a class="c-breadcrumb-link--active" href="/uc/ucla_etd">UCLA Electronic Theses and Dissertations</a></li></ul></nav><div class="c-columns--sticky-sidebar"><main id="maincontent"><div><div class="c-itemactions"><div class="o-download"><a href="/content/qt87p8p8p9/qt87p8p8p9.pdf?t=ndw6ls" class="o-download__button" download="eScholarship UC item 87p8p8p9.pdf">Download PDF</a><details class="o-download__formats"><summary aria-label="formats"></summary><div><ul class="o-download__nested-menu"><li class="o-download__nested-list1">Main<ul><li><a href="/content/qt87p8p8p9/qt87p8p8p9.pdf?t=ndw6ls" download="eScholarship UC item 87p8p8p9.pdf">PDF</a></li></ul></li></ul></div></details></div><details class="c-share"><summary><span>Share</span></summary><div><div class="c-share__list"><a class="c-share__email" href="">Email</a><a class="c-share__facebook" href="">Facebook</a></div></div></details></div></div><h2 class="c-tabcontent__main-heading" tabindex="-1"><div class="c-clientmarkup">The Voice Source in Speech Production: from Models to Applications</div></h2><div class="c-authorlist"><time class="c-authorlist__year">2014</time><ul class="c-authorlist__list"><li class="c-authorlist__begin"><a href="/search/?q=author%3AChen%2C%20Gang">Chen, Gang</a> </li><li class="c-authorlist__begin"><span class="c-authorlist__heading">Advisor(s):</span> <a href="/search/?q=author%3AALWAN%2C%20ABEER">ALWAN, ABEER</a> </li></ul></div><div class="c-tabs"><div class="c-tabs__tabs"><button class="c-tabs__button-more" aria-label="Show all tabs">...</button><button class="c-tabs__button--active">Main Content</button><button class="c-tabs__button">Metrics</button><button class="c-tabs__button">Author & Article Info</button></div><div class="c-tabs__content"><div class="c-tabcontent"><a name="article_abstract"></a><details class="c-togglecontent" open=""><summary>Abstract</summary><div class="c-clientmarkup"><p>The voice source contains important lexical and non-lexical information. The non-lexical information can convey, for example, prosodic events, emotional status, as well as cues pertaining to the uniqueness of the speaker's voice. A better understanding, and eventually a better model of the voice source, would benefit various speech applications, such as speech recognition, speech synthesis, speaker identification, age/gender classification, as well as clinical assessments.</p><p>This dissertation has three main goals. The first is to better understand the voice source through analyzing images of the vocal folds using laryngeal high-speed videoendoscopy (HSV) recordings. A new automatic method is proposed to compactly summarize the overall spatial synchronization pattern of vocal fold vibration for the entire laryngeal area from HSV data. Additionally, a new measure is proposed to adequately capture perceptually-important variations in glottal area pulse shapes, which are extracted from HSV data.</p><p>The second goal is to study the acoustic consequence of a physiological vocal-fold vibration pattern---the glottal gap effect, and apply our findings to a gender classification task of children's voices. Voice source related measures are found to improve classification accuracy, especially for younger (10-15 year old) speakers.</p><p>The third goal is to propose new voice source models and evaluate them in different applications. In the first application, a new source model and a noise-robust automatic source estimation algorithm are proposed to estimate the voice source from speech signals. Results in both clean and noisy conditions show that the proposed model and algorithm are robust in accurately estimating the voice source signal. The second application is to use the proposed source model for vowel synthesis. Perceptual listening experiments show that the proposed model provides a better perceptual match to the target voice than do traditional models.</p></div></details><details class="c-togglecontent" open=""><a name="article_main"></a><summary>Main Content</summary><div class="c-pdfview"><button class="c-pdfview__button-download">Download PDF to View</button><button class="c-pdfview__button-view">View Larger</button></div><div class="c-pdfview__accessibility">For improved accessibility of PDF content, <a href="/content/qt87p8p8p9/qt87p8p8p9.pdf?t=ndw6ls">download the file</a> to your device.</div><div class="c-pdfview__viewer"><div id="pdfjs-cdl-wrapper"> <div id="pdfjs-viewer" className="jsonly"> <noscript> <embed src=/content/qt87p8p8p9/qt87p8p8p9_noSplash_932f1d236bc6b557ca57567d83a63964.pdf?t=ndw6ls type='application/pdf' /> </noscript> <div id="outerContainer"> <div id="sidebarContainer"> <div id="toolbarSidebar"> <div class="splitToolbarButton toggled"> <button id="viewThumbnail" class="toolbarButton toggled" title="Show Thumbnails" tabindex="2" data-l10n-id="thumbs"> <span data-l10n-id="thumbs_label">Thumbnails</span> </button> <button id="viewOutline" class="toolbarButton" title="Show Document Outline (double-click to expand/collapse all items)" tabindex="3" data-l10n-id="document_outline"> <span data-l10n-id="document_outline_label">Document Outline</span> </button> <button id="viewAttachments" class="toolbarButton" title="Show Attachments" tabindex="4" data-l10n-id="attachments"> <span data-l10n-id="attachments_label">Attachments</span> </button> </div> </div> <div id="sidebarContent"> <div id="thumbnailView"> </div> <div id="outlineView" class="hidden"> </div> <div id="attachmentsView" class="hidden"> </div> </div> <div id="sidebarResizer" class="hidden"></div> </div>  <div id="mainContainer"> <div class="findbar hidden doorHanger" id="findbar"> <div id="findbarInputContainer"> <input id="findInput" class="toolbarField" title="Find" placeholder="Find in document…" tabindex="91" data-l10n-id="find_input"> <div class="splitToolbarButton"> <button id="findPrevious" class="toolbarButton findPrevious" title="Find the previous occurrence of the phrase" tabindex="92" data-l10n-id="find_previous"> <span data-l10n-id="find_previous_label">Previous</span> </button> <div class="splitToolbarButtonSeparator"></div> <button id="findNext" class="toolbarButton findNext" title="Find the next occurrence of the phrase" tabindex="93" data-l10n-id="find_next"> <span data-l10n-id="find_next_label">Next</span> </button> </div> </div> <div id="findbarOptionsOneContainer"> <input type="checkbox" id="findHighlightAll" class="toolbarField" tabindex="94"> <label for="findHighlightAll" class="toolbarLabel" data-l10n-id="find_highlight">Highlight all</label> <input type="checkbox" id="findMatchCase" class="toolbarField" tabindex="95"> <label for="findMatchCase" class="toolbarLabel" data-l10n-id="find_match_case_label">Match case</label> </div> <div id="findbarOptionsTwoContainer"> <input type="checkbox" id="findEntireWord" class="toolbarField" tabindex="96"> <label for="findEntireWord" class="toolbarLabel" data-l10n-id="find_entire_word_label">Whole words</label> <span id="findResultsCount" class="toolbarLabel hidden"></span> </div> <div id="findbarMessageContainer"> <span id="findMsg" class="toolbarLabel"></span> </div> </div>  <div id="secondaryToolbar" class="secondaryToolbar hidden doorHangerRight"> <div id="secondaryToolbarButtonContainer"> <button id="secondaryPresentationMode" class="secondaryToolbarButton presentationMode visibleLargeView" title="Switch to Presentation Mode" tabindex="51" data-l10n-id="presentation_mode"> <span data-l10n-id="presentation_mode_label">Presentation Mode</span> </button> <button id="secondaryOpenFile" class="secondaryToolbarButton openFile visibleLargeView" title="Open File" tabindex="52" data-l10n-id="open_file"> <span data-l10n-id="open_file_label">Open</span> </button> <button id="secondaryPrint" class="secondaryToolbarButton print visibleMediumView" title="Print" tabindex="53" data-l10n-id="print"> <span data-l10n-id="print_label">Print</span> </button> <button id="secondaryDownload" class="secondaryToolbarButton download visibleMediumView" title="Download" tabindex="54" data-l10n-id="download"> <span data-l10n-id="download_label">Download</span> </button> <a href="#" id="secondaryViewBookmark" class="secondaryToolbarButton bookmark visibleSmallView" title="Current view (copy or open in new window)" tabindex="55" data-l10n-id="bookmark"> <span data-l10n-id="bookmark_label">Current View</span> </a> <div class="horizontalToolbarSeparator visibleLargeView"></div> <button id="firstPage" class="secondaryToolbarButton firstPage" title="Go to First Page" tabindex="56" data-l10n-id="first_page"> <span data-l10n-id="first_page_label">Go to First Page</span> </button> <button id="lastPage" class="secondaryToolbarButton lastPage" title="Go to Last Page" tabindex="57" data-l10n-id="last_page"> <span data-l10n-id="last_page_label">Go to Last Page</span> </button> <div class="horizontalToolbarSeparator"></div> <button id="pageRotateCw" class="secondaryToolbarButton rotateCw" title="Rotate Clockwise" tabindex="58" data-l10n-id="page_rotate_cw"> <span data-l10n-id="page_rotate_cw_label">Rotate Clockwise</span> </button> <button id="pageRotateCcw" class="secondaryToolbarButton rotateCcw" title="Rotate Counterclockwise" tabindex="59" data-l10n-id="page_rotate_ccw"> <span data-l10n-id="page_rotate_ccw_label">Rotate Counterclockwise</span> </button> <div class="horizontalToolbarSeparator"></div> <button id="cursorSelectTool" class="secondaryToolbarButton selectTool toggled" title="Enable Text Selection Tool" tabindex="60" data-l10n-id="cursor_text_select_tool"> <span data-l10n-id="cursor_text_select_tool_label">Text Selection Tool</span> </button> <button id="cursorHandTool" class="secondaryToolbarButton handTool" title="Enable Hand Tool" tabindex="61" data-l10n-id="cursor_hand_tool"> <span data-l10n-id="cursor_hand_tool_label">Hand Tool</span> </button> <div class="horizontalToolbarSeparator"></div> <button id="scrollVertical" class="secondaryToolbarButton scrollModeButtons scrollVertical toggled" title="Use Vertical Scrolling" tabindex="62" data-l10n-id="scroll_vertical"> <span data-l10n-id="scroll_vertical_label">Vertical Scrolling</span> </button> <button id="scrollHorizontal" class="secondaryToolbarButton scrollModeButtons scrollHorizontal" title="Use Horizontal Scrolling" tabindex="63" data-l10n-id="scroll_horizontal"> <span data-l10n-id="scroll_horizontal_label">Horizontal Scrolling</span> </button> <button id="scrollWrapped" class="secondaryToolbarButton scrollModeButtons scrollWrapped" title="Use Wrapped Scrolling" tabindex="64" data-l10n-id="scroll_wrapped"> <span data-l10n-id="scroll_wrapped_label">Wrapped Scrolling</span> </button> <div class="horizontalToolbarSeparator scrollModeButtons"></div> <button id="spreadNone" class="secondaryToolbarButton spreadModeButtons spreadNone toggled" title="Do not join page spreads" tabindex="65" data-l10n-id="spread_none"> <span data-l10n-id="spread_none_label">No Spreads</span> </button> <button id="spreadOdd" class="secondaryToolbarButton spreadModeButtons spreadOdd" title="Join page spreads starting with odd-numbered pages" tabindex="66" data-l10n-id="spread_odd"> <span data-l10n-id="spread_odd_label">Odd Spreads</span> </button> <button id="spreadEven" class="secondaryToolbarButton spreadModeButtons spreadEven" title="Join page spreads starting with even-numbered pages" tabindex="67" data-l10n-id="spread_even"> <span data-l10n-id="spread_even_label">Even Spreads</span> </button> <div class="horizontalToolbarSeparator spreadModeButtons"></div> <button id="documentProperties" class="secondaryToolbarButton documentProperties" title="Document Properties…" tabindex="68" data-l10n-id="document_properties"> <span data-l10n-id="document_properties_label">Document Properties…</span> </button> </div> </div>  <div class="toolbar"> <div id="toolbarContainer"> <div id="toolbarViewer"> <div id="toolbarViewerLeft"> <button id="sidebarToggle" class="toolbarButton" title="Toggle Sidebar" tabindex="11" data-l10n-id="toggle_sidebar"> <span data-l10n-id="toggle_sidebar_label">Toggle Sidebar</span> </button> <div class="toolbarButtonSpacer"></div> <button id="viewFind" class="toolbarButton" title="Find in Document" tabindex="12" data-l10n-id="findbar"> <span data-l10n-id="findbar_label">Find</span> </button> <div class="splitToolbarButton hiddenSmallView"> <button class="toolbarButton pageUp" title="Previous Page" id="previous" tabindex="13" data-l10n-id="previous"> <span data-l10n-id="previous_label">Previous</span> </button> <div class="splitToolbarButtonSeparator"></div> <button class="toolbarButton pageDown" title="Next Page" id="next" tabindex="14" data-l10n-id="next"> <span data-l10n-id="next_label">Next</span> </button> </div> <input type="number" id="pageNumber" class="toolbarField pageNumber" title="Page" value="1" size="4" min="1" tabindex="15" data-l10n-id="page"> <span id="numPages" class="toolbarLabel"></span> </div> <div id="toolbarViewerRight"> <button id="presentationMode" class="toolbarButton presentationMode hiddenLargeView" title="Switch to Presentation Mode" tabindex="31" data-l10n-id="presentation_mode"> <span data-l10n-id="presentation_mode_label">Presentation Mode</span> </button> <button id="openFile" class="toolbarButton openFile hiddenLargeView" title="Open File" tabindex="32" data-l10n-id="open_file"> <span data-l10n-id="open_file_label">Open</span> </button> <button id="print" class="toolbarButton print hiddenMediumView" title="Print" tabindex="33" data-l10n-id="print"> <span data-l10n-id="print_label">Print</span> </button> <button id="download" class="toolbarButton download hiddenMediumView" title="Download" tabindex="34" data-l10n-id="download"> <span data-l10n-id="download_label">Download</span> </button> <a href="#" id="viewBookmark" class="toolbarButton bookmark hiddenSmallView" title="Current view (copy or open in new window)" tabindex="35" data-l10n-id="bookmark"> <span data-l10n-id="bookmark_label">Current View</span> </a> <div class="verticalToolbarSeparator hiddenSmallView"></div> <button id="secondaryToolbarToggle" class="toolbarButton" title="Tools" tabindex="36" data-l10n-id="tools"> <span data-l10n-id="tools_label">Tools</span> </button> </div> <div id="toolbarViewerMiddle"> <div class="splitToolbarButton"> <button id="zoomOut" class="toolbarButton zoomOut" title="Zoom Out" tabindex="21" data-l10n-id="zoom_out"> <span data-l10n-id="zoom_out_label">Zoom Out</span> </button> <div class="splitToolbarButtonSeparator"></div> <button id="zoomIn" class="toolbarButton zoomIn" title="Zoom In" tabindex="22" data-l10n-id="zoom_in"> <span data-l10n-id="zoom_in_label">Zoom In</span> </button> </div> <span id="scaleSelectContainer" class="dropdownToolbarButton"> <select id="scaleSelect" title="Zoom" tabindex="23" data-l10n-id="zoom"> <option id="pageAutoOption" title="" value="auto" selected="selected" data-l10n-id="page_scale_auto">Automatic Zoom</option> <option id="pageActualOption" title="" value="page-actual" data-l10n-id="page_scale_actual">Actual Size</option> <option id="pageFitOption" title="" value="page-fit" data-l10n-id="page_scale_fit">Page Fit</option> <option id="pageWidthOption" title="" value="page-width" data-l10n-id="page_scale_width">Page Width</option> <option id="customScaleOption" title="" value="custom" disabled="disabled" hidden="true"></option> <option title="" value="0.5" data-l10n-id="page_scale_percent" data-l10n-args='{ "scale": 50 }'>50%</option> <option title="" value="0.75" data-l10n-id="page_scale_percent" data-l10n-args='{ "scale": 75 }'>75%</option> <option title="" value="1" data-l10n-id="page_scale_percent" data-l10n-args='{ "scale": 100 }'>100%</option> <option title="" value="1.25" data-l10n-id="page_scale_percent" data-l10n-args='{ "scale": 125 }'>125%</option> <option title="" value="1.5" data-l10n-id="page_scale_percent" data-l10n-args='{ "scale": 150 }'>150%</option> <option title="" value="2" data-l10n-id="page_scale_percent" data-l10n-args='{ "scale": 200 }'>200%</option> <option title="" value="3" data-l10n-id="page_scale_percent" data-l10n-args='{ "scale": 300 }'>300%</option> <option title="" value="4" data-l10n-id="page_scale_percent" data-l10n-args='{ "scale": 400 }'>400%</option> </select> </span> </div> </div> <div id="loadingBar"> <div class="progress indeterminate" style="width:100%;"> <div class="glimmer"> </div> </div> </div> </div> </div> <menu type="context" id="viewerContextMenu"> <menuitem id="contextFirstPage" label="First Page" data-l10n-id="first_page"></menuitem> <menuitem id="contextLastPage" label="Last Page" data-l10n-id="last_page"></menuitem> <menuitem id="contextPageRotateCw" label="Rotate Clockwise" data-l10n-id="page_rotate_cw"></menuitem> <menuitem id="contextPageRotateCcw" label="Rotate Counter-Clockwise" data-l10n-id="page_rotate_ccw"></menuitem> </menu> <div id="viewerContainer" tabindex="0"> <div id="viewer" class="pdfViewer"></div> </div> <div id="errorWrapper" hidden='true'> <div id="errorMessageLeft"> <span id="errorMessage"></span> <button id="errorShowMore" data-l10n-id="error_more_info"> More Information </button> <button id="errorShowLess" data-l10n-id="error_less_info" hidden='true'> Less Information </button> </div> <div id="errorMessageRight"> <button id="errorClose" data-l10n-id="error_close"> Close </button> </div> <div class="clearBoth"></div> <textarea id="errorMoreInfo" hidden='true' readonly="readonly"></textarea> </div> </div>  <div id="overlayContainer" class="hidden"> <div id="passwordOverlay" class="container hidden"> <div class="dialog"> <div class="row"> <p id="passwordText" data-l10n-id="password_label">Enter the password to open this PDF file:</p> </div> <div class="row"> <input type="password" id="password" class="toolbarField"> </div> <div class="buttonRow"> <button id="passwordCancel" class="overlayButton"><span data-l10n-id="password_cancel">Cancel</span></button> <button id="passwordSubmit" class="overlayButton"><span data-l10n-id="password_ok">OK</span></button> </div> </div> </div> <div id="documentPropertiesOverlay" class="container hidden"> <div class="dialog"> <div class="row"> <span data-l10n-id="document_properties_file_name">File name:</span> <p id="fileNameField">-</p> </div> <div class="row"> <span data-l10n-id="document_properties_file_size">File size:</span> <p id="fileSizeField">-</p> </div> <div class="separator"></div> <div class="row"> <span data-l10n-id="document_properties_title">Title:</span> <p id="titleField">-</p> </div> <div class="row"> <span data-l10n-id="document_properties_author">Author:</span> <p id="authorField">-</p> </div> <div class="row"> <span data-l10n-id="document_properties_subject">Subject:</span> <p id="subjectField">-</p> </div> <div class="row"> <span data-l10n-id="document_properties_keywords">Keywords:</span> <p id="keywordsField">-</p> </div> <div class="row"> <span data-l10n-id="document_properties_creation_date">Creation Date:</span> <p id="creationDateField">-</p> </div> <div class="row"> <span data-l10n-id="document_properties_modification_date">Modification Date:</span> <p id="modificationDateField">-</p> </div> <div class="row"> <span data-l10n-id="document_properties_creator">Creator:</span> <p id="creatorField">-</p> </div> <div class="separator"></div> <div class="row"> <span data-l10n-id="document_properties_producer">PDF Producer:</span> <p id="producerField">-</p> </div> <div class="row"> <span data-l10n-id="document_properties_version">PDF Version:</span> <p id="versionField">-</p> </div> <div class="row"> <span data-l10n-id="document_properties_page_count">Page Count:</span> <p id="pageCountField">-</p> </div> <div class="row"> <span data-l10n-id="document_properties_page_size">Page Size:</span> <p id="pageSizeField">-</p> </div> <div class="separator"></div> <div class="row"> <span data-l10n-id="document_properties_linearized">Fast Web View:</span> <p id="linearizedField">-</p> </div> <div class="buttonRow"> <button id="documentPropertiesClose" class="overlayButton"><span data-l10n-id="document_properties_close">Close</span></button> </div> </div> </div>  <div id="printServiceOverlay" class="container hidden"> <div class="dialog"> <div class="row"> <span data-l10n-id="print_progress_message">Preparing document for printing…</span> </div> <div class="row"> <progress value="0" max="100"></progress> <span data-l10n-id="print_progress_percent" data-l10n-args='{ "progress": 0 }' class="relative-progress">0%</span> </div> <div class="buttonRow"> <button id="printCancel" class="overlayButton"><span data-l10n-id="print_progress_close">Cancel</span></button> </div> </div> </div>     </div>  </div>  <div id="printContainer"></div> </div>  </div></div></details></div><p><br/></p></div></div></main><aside><section class="o-columnbox1"><header><h2>Jump To</h2></header><div class="c-jump"><a href="/uc/item/87p8p8p9">Article</a><ul class="c-jump__tree" aria-labelledby="c-jump__label"><li><a href="/uc/item/87p8p8p9">Abstract</a></li><li><a href="/uc/item/87p8p8p9">Main Content</a></li></ul><ul class="c-jump__siblings"><li><a href="/uc/item/87p8p8p9">Metrics</a></li><li><a href="/uc/item/87p8p8p9">Author & Article Info</a></li></ul></div></section><div><section class="o-columnbox1"><header><h2>Related Items</h2></header><ul class="c-relateditems"><li><h3><a class="o-textlink__secondary" href="/uc/item/7nm121ws"><div class="c-clientmarkup">Sampling is All You Might Need</div></a></h3><div class="c-authorlist"><ul class="c-authorlist__list"><li class="c-authorlist__begin"><a href="/search/?q=author%3AChan%2C%20Jeffrey">Chan, Jeffrey</a> </li><li class="c-authorlist__begin"><span class="c-authorlist__heading">Advisor(s):</span> <a href="/search/?q=author%3AGuido%20F.%2C%20Mont%C3%BAfar">Guido F., Montúfar</a> </li></ul></div></li><li><h3><a class="o-textlink__secondary" href="/uc/item/7qm8v5mb"><div class="c-clientmarkup">Non-spatial hippocampal behavioral timescale synaptic plasticity during working memory is gated by entorhinal inputs</div></a></h3><div class="c-authorlist"><ul class="c-authorlist__list"><li class="c-authorlist__begin"><a href="/search/?q=author%3ADorian%2C%20Conor%20C">Dorian, Conor C</a> </li><li class="c-authorlist__begin"><span class="c-authorlist__heading">Advisor(s):</span> <a href="/search/?q=author%3AGolshani%2C%20Peyman">Golshani, Peyman</a> </li></ul></div></li><li><h3><a class="o-textlink__secondary" href="/uc/item/8c7759n4"><div class="c-clientmarkup">Prevalence and Clinical Significance of Incidental Findings in Orthodontic Large Field of View Cone-Beam Computed Tomography Scans</div></a></h3><div class="c-authorlist"><ul class="c-authorlist__list"><li class="c-authorlist__begin"><a href="/search/?q=author%3AMurillo%2C%20Guiselle">Murillo, Guiselle</a> </li><li class="c-authorlist__begin"><span class="c-authorlist__heading">Advisor(s):</span> <a href="/search/?q=author%3AMallya%2C%20Sanjay%20M">Mallya, Sanjay M</a> </li></ul></div></li><li><h3><a class="o-textlink__secondary" href="/uc/item/8fr0m7mq"><div class="c-clientmarkup">Risk, Reward, and Correlation: Bitcoin and Traditional Assets in a Volatile Market Environment</div></a></h3><div class="c-authorlist"><ul class="c-authorlist__list"><li class="c-authorlist__begin"><a href="/search/?q=author%3AHUANG%2C%20YI">HUANG, YI</a> </li><li class="c-authorlist__begin"><span class="c-authorlist__heading">Advisor(s):</span> <a href="/search/?q=author%3AWu%2C%20Yingnian">Wu, Yingnian</a> </li></ul></div></li><li><h3><a class="o-textlink__secondary" href="/uc/item/8zf0b26n"><div class="c-clientmarkup">Causes and consequences of competition in spatially variable environments for plant coexistence and distributions: a study with California annual plants</div></a></h3><div class="c-authorlist"><ul class="c-authorlist__list"><li class="c-authorlist__begin"><a href="/search/?q=author%3AHayashi%2C%20Kenji%20Thomas">Hayashi, Kenji Thomas</a> </li><li class="c-authorlist__begin"><span class="c-authorlist__heading">Advisor(s):</span> <a href="/search/?q=author%3AKraft%2C%20Nathan">Kraft, Nathan</a> </li></ul></div></li></ul></section></div></aside></div></div><div><div class="c-toplink"><a href="javascript:window.scrollTo(0, 0)">Top</a></div><footer class="c-footer"><nav class="c-footer__nav"><ul><li><a href="/">Home</a></li><li><a href="/aboutEschol">About eScholarship</a></li><li><a href="/campuses">Campus Sites</a></li><li><a href="/ucoapolicies">UC Open Access Policy</a></li><li><a href="/publishing">eScholarship Publishing</a></li><li><a href="https://www.cdlib.org/about/accessibility.html">Accessibility</a></li><li><a href="/privacypolicy">Privacy Statement</a></li><li><a href="/policies">Site Policies</a></li><li><a href="/terms">Terms of Use</a></li><li><a href="/login"><strong>Admin Login</strong></a></li><li><a href="https://help.escholarship.org"><strong>Help</strong></a></li></ul></nav><div class="c-footer__logo"><a href="/"><img class="c-lazyimage" data-src="/images/logo_footer-eschol.svg" alt="eScholarship, University of California"/></a></div><div class="c-footer__copyright">Powered by the<br/><a href="http://www.cdlib.org">California Digital Library</a><br/>Copyright © 2017<br/>The Regents of the University of California</div></footer></div></div></div></div> <script src="/js/vendors~app-bundle-2aefc956e545366a5d4e.js"></script> <script src="/js/app-bundle-4477d7630fb8c6f70662.js"></script> </body> </html>

CINXE.COM

The Voice Source in Speech Production: from Models to Applications