DOC PREVIEW
CMU CS 15492 - Speech Synthesis Overview

This preview shows page 1-2-24-25 out of 25 pages.

Save
View full document
View full document
Premium Document
Do you want full access? Go Premium and unlock all 25 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 25 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 25 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 25 pages.
Access to all documents
Download any document
Ad free experience
Premium Document
Do you want full access? Go Premium and unlock all 25 pages.
Access to all documents
Download any document
Ad free experience

Unformatted text preview:

Speech Processing 15-492/18-492Speech SynthesisOverviewText processingSpeech SynthesisFrom text to speechFrom text to speechText AnalysisText AnalysisStrings of characters to wordsStrings of characters to wordsLinguistic AnalysisLinguistic AnalysisFrom words to pronunciations and prosodyFrom words to pronunciations and prosodyWaveform SynthesisWaveform SynthesisFrom pronunciations to waveformsFrom pronunciations to waveformsText AnalysisThis is a pen.This is a pen.My cat who lives dangerously has nine lives.My cat who lives dangerously has nine lives.He stole $100 from the bank.He stole $100 from the bank.He stole 1996 cattle on 25 Nov 1996.He stole 1996 cattle on 25 Nov 1996.He stole $100 million from the bank.He stole $100 million from the bank.It's 13 St. Andrew St. near the bank.It's 13 St. Andrew St. near the bank.Its a PIII 1.5Ghz, 512MB RAM, 160Gb SATA, Its a PIII 1.5Ghz, 512MB RAM, 160Gb SATA, (no IDE) 24x (no IDE) 24x cdromcdromand 19" LCD.and 19" LCD.My home My home pgaepgaeis is http://http://www.geocities.com/awbwww.geocities.com/awb/./.Emailfrom from [email protected]@cstr.ed.ac.uk("Alan W Black") on Thu 23 Nov 15:30:45:("Alan W Black") on Thu 23 Nov 15:30:45:>>> ... but, *I* wont make it :> ... but, *I* wont make it :--) Can you tell me who's going?) Can you tell me who's going?>>IMHO I think you should go, but I think the IMHO I think you should go, but I think the followignfollowignare goingare goingGeorge BushGeorge BushBill ClintonBill Clintonand that other guyand that other guyBobBob----___ _ ___ _ ------------------++------------------------------------------------------------------------------------------------------+ |+ |\\\\//|//|| Bob Beck E| Bob Beck E--mail mail [email protected]@beck.demon.co.uk| | | | \\\\// |// |++------------------------------------------------------------------------------------------------------+ | > < |+ | > < || // | // \\\\||Alba Alba gugubrathbrath|//___|//___\\\\||----------------Text Analysis TasksCharacter encodings:Character encodings:LatinLatin--1, iso1, iso--88598859--1, utf1, utf--8 (or special)8 (or special)Find tokensFind tokensWhite space separatedWhite space separatedChunk into reasonably sized chunksChunk into reasonably sized chunksSort of sentencesSort of sentencesMap tokens to wordsMap tokens to wordsDisambiguate token typesDisambiguate token typesNumbersNumbersChunkingMaking reasonable sized sectionsMaking reasonable sized sectionsSomething to do with full stops …Something to do with full stops …Hi Alan,Hi Alan,I went to the conference. They listed you as Mr. Black when weI went to the conference. They listed you as Mr. Black when weknow you should be Dr. Black days ahead for their research.know you should be Dr. Black days ahead for their research.Next month I'll be in the U.S.A. I'll try to drop by C.M.U.Next month I'll be in the U.S.A. I'll try to drop by C.M.U.if I have time.if I have time.byebyeDorothyDorothyInstitute of XYZInstitute of XYZUniversity of Foreign PlaceUniversity of Foreign Placeemail: [email protected]: [email protected] analysisNormal wordsNormal wordsHomographs, Homographs, OOVsOOVsNumbersNumbersYears, quantities, digits, addressesYears, quantities, digits, addressesOther standard formsOther standard formsDates, times, moneyDates, times, moneyAbbreviations and Letter SequencesAbbreviations and Letter SequencesNASA, CIA, SATA, IDENASA, CIA, SATA, IDESpelling errors (choices)Spelling errors (choices)SoooooSooooo, … , … colourcolour, , collorcollorPunctuationPunctuation::--) quotes, dashes, ) quotes, dashes, asciiasciiart, art, Text layoutText layoutFinding WordsWhite space separated tokensWhite space separated tokensButBut------if I may interjectif I may interject------not all not all word(sword(s) are like ) are like thatthatWeanWean--HallHall--like architecturelike architectureSome languages don’t use spacesSome languages don’t use spacesChinese, Japanese, ThaiChinese, Japanese, ThaiSome languages use lots of compoundingSome languages use lots of compoundingunspacedmultiwordsunspacedmultiwordsHomographsHomographsHomographsSame writing, different pronunciationSame writing, different pronunciation(Homophones: same pronunciation different writing. “to” (Homophones: same pronunciation different writing. “to” “two” “write” “right”)“two” “write” “right”)English: not many:English: not many:Stress shift (Noun/Verb)Stress shift (Noun/Verb)Segment, project, convictSegment, project, convictSemanticSemanticBass, read, Begin, bathing, lives, Celtic, wind, Reading, sun, Bass, read, Begin, bathing, lives, Celtic, wind, Reading, sun, wed, …wed, …Roman NumeralsRoman NumeralsNon-standard Words (NSW)• Words not in the lexicon20.1%20.1%IMIM27.9%27.9%ClassifiedsClassifieds13.7%13.7%RecipesRecipes10.7%10.7%EmailEmail4.9%4.9%Press wirePress wire1.5%1.5%NovelsNovels%NSW%NSWText TypeText TypeDistribution of NSW• 3yrs News text, 2.2M tokens 120K NSWs2%2%As AbbrevAs Abbrev12%12%As lettersAs letters30%30%As wordAs wordAlphabeticAlphabetic3%3%OrdinalOrdinal7%7%YearYear26%26%NumberNumberNumericNumeric% of NSW% of NSWMinor typeMinor typeMajor typeMajor typeProcessing NSWsHow hard are they?How hard are they?Finding themFinding themIdentifying themIdentifying themExpanding themExpanding themCurrent processing techniquesCurrent processing techniquesIgnoredIgnoredLexical lookupLexical lookupHackyHackyhandhand--written ruleswritten rules(not so) (not so) HackyHackyhandhand--written ruleswritten rulesStatistically train models (and Statistically train models (and hackyhackyhand written rules)hand written rules)Homograph Disambiguation (Yarowsky)Same tokens in different contextsSame tokens in different contextsIdentify target homographIdentify target homographE.g. numbers, roman numerals, “St”E.g. numbers, roman numerals, “St”Find instances in large text corporaFind instances in large text corporaHand label them with correct answerHand label them with correct answerTrain a decision tree to


View Full Document
Download Speech Synthesis Overview
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Speech Synthesis Overview and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Speech Synthesis Overview 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?