Last Week Phonetics Smoothing algorithms redistribute probability CS 341: Natural Language Processing Prof. Heather Pon-Barry www.mtholyoke.edu/courses/ponbarry/cs341.html N-gram language models are used in many applications, including speech recognition
Today Sound Categories ARPAbet transcription Articulatory phonetics Acoustic phonetics Spectrograms HH EH L OW W ER L D Phone: basic speech sound of a language A minimal sound difference between two words Phoneme: class of speech sounds Phoneme may include several phones /p/ in potato, leopard
ARPAbet An alphabet for transcribing American English phonetic sounds. Multiple (ascii) characters for each sound!! HH EH L OW W ER L D http://www.speech.cs.cmu.edu/cgi-bin/cmudict canonical pronunciations actual speech may vary Phonetic Symbol Sets ARPAbet is for American English only What about other languages? International Phonetic Alphabet (IPA) Single (unique) character for each sound Represents all sounds of the world s languages Large, requires a special (non-ascii) font http://en.wikipedia.org/wiki/international_phonetic_alphabet
English Orthography A single sound can be written in many different ways [iy] [s] [uw] [ay] sea, see, scene, receive, thief cereal, same, miss true, few, choose, lieu, do lie, prime, pry, buy,
Exercise Find and correct the mistakes in the ARPAbet transcriptions of the following words: three [dh r iy] sing [s ih n g] eyes [ay s] stutter [s t uh t er] though [th ow] http://en.wikipedia.org/wiki/arpabet
Exercise Answers Exercise Find and correct the mistakes in the ARPAbet transcriptions of the following words: three [dh r iy] [th r iy] sing [s ih n g] [s iy ng] eyes [ay s] [ay z] Write your first name in ARPAbet Write first name of person to your right in ARPAbet stutter [s t uh t er] [s t ah dx er] though [th ow] [dh ow]
Today Speech Production Process ARPAbet transcription Articulatory phonetics Acoustic phonetics Spectrograms Respiration We (normally) speak while breathing out. Respiration provides airflow. Pulmonic egressive airstream Phonation Airstream sets vocal folds in motion. Vibration of vocal folds produces sounds. Sound is then modulated by: Articulation and Resonance Shape of vocal tract, characterized by: Oral tract Teeth, soft palate (velum), hard palate Tongue, lips, uvula Nasal tract Content from Dan Jurafsky
Sagittal section of the vocal tract (Techmer 1880) Nasal Cavity Pharynx Vocal Folds (within the Larynx) Trachea Lungs Text copyright J. J. Ohala, Sept 2001, from Sharon Rose slide Video from Shri Narayanan USC SAIL Lab
Voicing Air comes up from lungs Forces its way through vocal cords, pushing open (2,3,4) This causes air pressure in glottis to fall, since: when gas runs through constricted passage, its velocity increases (Venturi tube effect) increase in velocity results in a drop in pressure (Bernoulli principle) Because of drop in pressure, vocal cords snap together again (6-10) Single cycle: ~1/100 of a second. Video from Shri Narayanan USC SAIL Lab Figure & text from John Coleman s web site
Voicelessness Consonants and Vowels When vocal cords are open, air passes through unobstructed Voiceless sounds: p/t/k/s/f/sh/th/ch If the air moves very quickly, the turbulence causes a different kind of phonation: whisper Consonants: phonetically, sounds with audible noise produced by a constriction Vowels: phonetically, sounds with no audible noise produced by a constriction
Consonants Place of Articulation How is the airflow restricted Stop (or plosive): [p], [t], [g], Airflow is completely blocked (closure) and released (release) Nasal: air is released through the nose [m], [ng] Fricative: [s], [z], [f] air is forced through a narrow channel, leading to turbulent airflow Affricates: [t ] begin as stops, but the release is fricative
Vowels All voiced Vowel height How high is the tongue? High or low? Where is its highest point? Front or back? How rounded are the lips? Video from Shri Narayanan USC SAIL Lab
Vowels American English Vowel Space HIGH iy uw How many vowels does English have? FRONT ih ey eh ix ax ah ux oy uh aw ow ao BACK ay ae aa LOW Figure from Jennifer Venditti
Today Audio Processing energy ARPAbet transcription Articulatory phonetics Acoustic phonetics time Spectrograms 25 ms
Lungs Vocal cords Vocal tract Spectrum 40 + 20 0 25 millisecond window 0 5000 Frequency (Hz) Fourier Transform
Spectrogram Spectrograms of 3 Vowels formants: dark horizontal bands time
Summary Next Class ARPAbet alphabet for expressing American English phones Articulatory phonetics: how different sounds are produced Spectrum: energy at different frequencies Wednesday: meet in Clapp 202 Bring headphones Spectrogram: spectrum over time