Respiratory system Phonatory system - Free

1 1 Speech Slide 1 2001 Christian Martyn JonesSpeech and Natural Language ProcessingSpeech what s so good about speech?1. Speech is a unique faculty to humans and one ofthe most important, requiring the precise controland co-ordination of over eighty different muscles,making speech the highest learned skill a humancan Speaking requires generally around 1,500 musclecommands every second but needs only, aschildren, a few years to Although a primary means of communication initself, speech can convey other messages throughaccents, tone, pitch, and ..but only limited achievements in machinespeech recognition and synthesisSpeech Slide 2 2001 Christian Martyn JonesSpeech and Natural Language ProcessingSpeech production The act of speech involves three major anatomicalsubsystems:1.

2 Respiratory systemincluding the lungs, rib cage, and diaphragm;2. Phonatory systemwhich includes the larynx;3. articulatory featuresthe lips, teeth, tongue, and Slide 3 2001 Christian Martyn JonesSpeech and Natural Language ProcessingRespiratory systemnasal cavityalveolar ridgeteethlipstongueglottislarynxlungsdi aphragmvelumepiglottisesophagusSpeech Slide 4 2001 Christian Martyn JonesSpeech and Natural Language ProcessingPhonatory systemSpeech Slide 5 2001 Christian Martyn JonesSpeech and Natural Language ProcessingArticulatory system ..considering in tern how we createconsonantsvowelsother the glottis, tongue, teeth, lips, and nasalcavitySpeech Slide 6 2001 Christian Martyn JonesSpeech and Natural Language ProcessingArticulation of consonants How we classify the production of consonantsinvolves1.

3 The place of articulation(the relative position of the lips, teeth, and tongue),2. the manner of articulation(how the air-stream from the lungs is obstructedstops, fricatives, affricates, nasals, liquids and glides),3. and whether the vocal cords are set to vibrate2 Speech Slide 7 2001 Christian Martyn JonesSpeech and Natural Language use American than British Englishbut why? my work most other researchers will also simplify not consider regional accentsSpeech Slide 8 2001 Christian Martyn JonesSpeech and Natural Language ProcessingAnnotation ASRtakes an acoustic waveform as input and produces asoutput a string of words TtStakes a sequence of text words to produce anacoustic Slide 9 2001 Christian Martyn JonesSpeech and Natural Language main alphabet standards 1.

4 International Phonetic Alphabet (IPA)standard originally developed by internationalphonetic association in 1888 with the idea totranscribe all human languagesit is more than just a set of symbols (eg one to onerelationship to sounds) and differ for differentlanguages 2. ARPA betusing ASCII characters rather than more non-standard charactersmakes it much easier to create phonetic dictionaries,syntactic and semantic rules, and build them intoASR we will Slide 11 2001 Christian Martyn JonesSpeech and Natural Language ProcessingPlace of articulation.

5 Refers to the relative positions of the lips, teeth andtongue. There are six distinct types of classification:bilabial,labiodental,inte rdental,alveolar,alveo-palatal,and Slide 12 2001 Christian Martyn JonesSpeech and Natural Language Processingnasal cavitysoft palateuvulahard palatepalatealveolar ridgelipstip of tongueblade of tongueback of tonguejawSpeech Slide 15 2001 Christian Martyn JonesSpeech and Natural Language Processing The six places of articulationbilabial, labiodental, interdental, alveolar, alveopalatal,and velardescribe the parts of the vocal tract which areresponsible for the obstruction of the air flow fromthe lungs the degree of obstruction the airstream incurs mustalso be is the manner of Slide 16 2001 Christian Martyn JonesSpeech and Natural Language ProcessingManner of articulation The manner of articulation , and to what degree, air from the lungs isobstructed Terms used.

6 Stops,fricatives,affricates,nasals,later al,retroflex,and glidesSpeech Slide 19 2001 Christian Martyn JonesSpeech and Natural Language ProcessingVoicing Voicingvibration of the vocal cords in order to change thecharacteristics of the airstream through the mouthor nose and the overall acoustic nature of thephone..sounds that are generated with the vocal cordsvibrating are voiced, and conversely those soundrequiring static vocal cords are Slide 21 2001 Christian Martyn JonesSpeech and Natural Language ProcessingArticulation of vowels The articulation of American English vowels is not asdefined as consonants and can vary a great deal fromspeaker to speaker, especially due to dialectvariations.

7 Vowels generally present little obstruction and requirewide open mouth positions. The description of vowels will consider the motion ofthe tongue and Slide 22 2001 Christian Martyn JonesSpeech and Natural Language ProcessingFeatures to consider What we look for are:1. tongue elevation,2. part of tongue involved,3. tongue muscle tension,4. mouth Slide 23 2001 Christian Martyn JonesSpeech and Natural Language ProcessingElevation of the tongue normally categorised in terms of simply:high, mid, or low frontmid frontlow frontSpeech Slide 24 2001 Christian Martyn JonesSpeech and Natural Language ProcessingExamples /iy/ in beet and /ih/ in bit /ey/ in bait and /eh/ in bet /ae/ in bat /uw/ in boot and /uh/ in book /ow/ in boat and /ao/ in bought /aa/ in bott4 Speech Slide 25 2001 Christian Martyn JonesSpeech and Natural Language ProcessingRegion of the tongue We classify the region of the tongue as either.

8 Front, central, or backwe have already seen front vowels:'beet , bit , bait , bet , 'bat and back vowels such as: boot , book , boat , bought , and bott central vowels such as:/ah/ in but and schwa vowel in machine Speech Slide 26 2001 Christian Martyn JonesSpeech and Natural Language ProcessingTongue tenseness vowels requiring above the normal level of muscletension are termed tense whilst those in which this degree of tension is notneeded are simply relaxed. Consider list beat and bit, bait and bet , and boot and book - the first phone in each group is tense, with the tongue slightlyhigher in the mouth than the relaxed second phone- in addition, front vowels which are tense are articulated with thetongue slightly ahead of a similarly relax phone, whilst backtense vowels will be pronounced with the tongue further Slide 27 2001 Christian Martyn JonesSpeech and Natural Language ProcessingMouth shape a range of lip positions for all phones however vowels possess some degree ofgeneralisationvowels.

9 'beet , bit , bait , bet , 'bat , machine , and but whilst the back vowels of: boot , book , boat , bought :Speech Slide 29 2001 Christian Martyn JonesSpeech and Natural Language ProcessingOther phonetic articulations these require more dynamic movement of the tongueand lips than can be described within the constraintsof consonants and vowels, and include:Off-glides,diphthongs,and r-colourisationconsonant /h/Speech Slide 30 2001 Christian Martyn JonesSpeech and Natural Language ProcessingOff-glides Already considered are the high, front, tense /iy/ ('beet'), the mid,front, tense /ey/ ('bait'), the high, back, tense /uw/ ('boot'), and themid, back, tense /ow/ ('boat') vowels, however the articulation ofeach is more than these simple descriptors can suggest.

10 The front vowels (/iy/ and /ey/) are composed of a pure vowel /i/(idealised position) and /e/ followed immediately by the glide /y/(blade of tongue near hard palate), and similarly the back vowelsinvolve vowels /u/ and /o/ followed by /w/. Each of the constituent phones are pronounced and thesevowels are known as off-glides indicating the increased motion ofthe tongue and Slide 31 2001 Christian Martyn JonesSpeech and Natural Language ProcessingDiphthongs Similar in composition to off-glides, diphthongs are complexvowels consisting of a vowel sound followed by the glide /y/ or/w/. However, what separates diphthongs from off-glides is thatdiphthongs involve considerably greater tongue motion The diphthongs of American English are /aw/ in bout , /ay/ in bite , and /oy/ in boy.