Speech Processing Introduction Syllabus u u ECE

Speech Processing Introduction

Syllabus u u ECE 5525 Speech Processing Contact Info: u CRN 81888 u u Instructor(s): Veton Këpuska Textbook(s): “Discrete-Time Speech Signal Processing: Principles and Practice”, Thomas F. Quatieri, Prentice Hall, 2002 u Reference Material: Këpuska, Veton Olin Engineering Building, Office # 353 Tel. (321) 674 -7183 E-mail: vkepuska@fit. edu http: //my. fit. edu/~vkepuska/ece 5525 n n n 3/15/2018 Subj ECE Crse 5525 Sec E 1 Credits Title 3. 00 Speech Processing “Digital Processing of Speech Signals”, L. R. Rabiner and R. W. Schafer, Prentice Hall, 1978 “Digital Signal Processing”, Alan V. Oppenheim and Ronald W. Schafer, Prentice Hall, 1975 “Digital Signal Processing A Practical Approach”, Emmanuel C. Ifeachor and Barrie W. Jervis, Second Edition, Prentice Hall, 2002 Veton Këpuska 2

Syllabus u Course Goals: Teach modern methods that are used to process speech signals. u Subject Area: Digital Speech Processing u Prerequisites by Topic: u Topics Covered: n n n n n u Discrete-Time Speech Signal Processing, Production and Classification of Speech Sounds, Acoustic Theory of Speech Production, Speech Perception, Speech Analysis, Speech Synthesis, Homomorphic Speech Processing, Short-Time Fourier Transform Analysis and Synthesis, Filter-Bank Analysis/Synthesis, Sinusoidal Analysis/Synthesis, Frequency Domain-Pitch Estimation, Nonlinear Measurement and Modeling Techniques Coding of Speech Signals, Speech Enhancement, Speaker Recognition, Methods for Speech Recognition, Digital Speech Processing for Man-Machine Communication by Voice Recommended Grading n n n u Digital Signal Processing Homework: Exams: Project(s): 20% 30% 50% MATLAB Exercises and Homework Problems 3/15/2018 Veton Këpuska 3

Course Information u http: //my. fit. edu/~vkepuska/web n or Directly u http: //my. fit. edu/~vkepuska/ece 5525/ 3/15/2018 Veton Këpuska 4

Discrete-Time Speech Signal Processing Introduction Veton Këpuska

Introduction u The topic of this study is as old as our human language and as new as the newest computer chip. u Goal of creation of more efficient and effective systems for: n human-to-human communication, as well as n More recently human-machine communication, is studied. 3/15/2018 Veton Këpuska 7

History 1960’s Digital Signal Processing (DSP): u Assumed a central role in speech communication studies: n It enabled development of a large number of applications n Advances in IC technology n DSP algorithms n Computer architecture 3/15/2018 Veton Këpuska 8

History u Created environment with virtually limitless opportunities for innovation in n n n 3/15/2018 Speech Processing Image Processing Video Processing Radar & Sonar Medical Diagnostics, and Consumer Electronics Veton Këpuska 9

History u Three levels of understanding are required to appreciate the technological advancement of DSP n Theoretical n Conceptual n Practical Practice Concepts Theory 3/15/2018 Veton Këpuska Ability to implement theory and concepts in working code (MATLAB, C, C++, JAVA) Basic understanding of how theory is applied Mathematics, derivations, signal processing 10

Speech Technology u Theoretical: n Acoustic Theory of Speech Production n The basic Mathematics of Speech Signal representation. n Derivation of Various Properties of Speech associated with each representation n Basic Signal Processing mathematics: u u 3/15/2018 Speech signal Sampling Aliasing Filtering and other DSP methods Veton Këpuska 11

Speech Technology u Conceptual n Hos speech Processing theory is applied in order to make various speech measurements and to estimate and quantify various attributes of the speech signal 3/15/2018 Veton Këpuska 12

Speech Technology u Practical n n For technology to realize its full potential, it is essential to be able to convert theory and conceptual understanding to practice. This process involves constraints: u Knowledge of the goals of the application u Knowledge of the engering tradeoffs and judgments, and u Ability to provide implementations in working computer code (e. g. , MATLAB, C/C++ or java), u Specialized code running in real-time signal-processing chips u Specialized languages and technologies such as: n n n 3/15/2018 ASICs FPGAs (Field Programmable Gate Arrays) DSP chips Veton Këpuska 13

Speech Technology u What is the nature of speech signal? u How do DPS techniques play a role in learning about the speech signal? u What are the basic digital representations of speech signal and how are they used in algorithms for speech processing? u What are the important applications that are enabled by digital speech processing methods? 3/15/2018 Veton Këpuska 14

Discrete-Time Speech Signal Processing u Speech has evolved as a primary form of communication between humans. u Technological advancement has enhanced our ability to communicate: n One early case is the transduction by a telephone handset of the continuously-varying speech pressure signal at the lips output to continuously-varying (analog) electric voltage signal. n Digital technology has brought communication to a new level. This technology requires additional transduction of the signals: u Analog-to-digital (A/D) and Digital-to-Analog converter (D/A), vs u Continuous Analog Signal 3/15/2018 Veton Këpuska 15

Speech Communication Chain 3/15/2018 Veton Këpuska 16

Discrete-Time Speech Signal Processing u The topic of this course can be loosely defined as: n The manipulation of sampled speech signals by a digital processor to obtain a new signal with some desired properties. n Example: Changing a speaker’s rate of articulation with the use of digital computer. u Modification of articulation rate (referred to as timescale modification of speech) has as an objective to: n n 3/15/2018 generate a new speech waveform that corresponds to a person talking faster or slower than the original rate, maintain the character of the speaker’s voice (i. e. , there should be little change in the pitch and spectrum of the original utterance). Veton Këpuska 17

Time-Scale Modification Example u Useful applications: n Fast Scanning of a long recording in a message playback system. n Slowing Down difficult to understand speech. 3/15/2018 Veton Këpuska 18

The Speech-Communication Pathway u The linguistic level of communication: n Idea is first formed in the mind of the speaker. n This idea is then transformed to words, phrases, and sentences according to the grammatical rules of the language. u The physiological level of communication: n The brain creates electric signals that move along the motor nerves these electric signals activate muscles in the vocal tract and vocal cords. u The acoustic level in speech communication pathway: n This vocal tract and vocal cord movement results in pressure changes within the vocal tract, and, in particular, at the lips initiating a sound wave that propagates in space. n Pressure changes at the ear canal cause vibrations at the ear drum of the listener. 3/15/2018 Veton Këpuska 19

The Speech-Communication Pathway u The physiological level of communication pathway: n Eardrum vibrations induce electric signals that move along the sensory nerves to the brain. u The linguistic level of the listener: n The brain performs speech recognition and understanding. u The linguistic and physiological activity of: n The speaker n The listener 3/15/2018 => => “transmitter” “receiver” Veton Këpuska 20

The Speech-Communication Pathway u Transmitter and Receiver of the speechcommunication system have other functions besides basic communication: n Monitoring and correction of one’s own speech via the feedback through speakers ear (importance of the feedback studied in the speech of the deaf). u Control of articulation rate u Adaptation of speech production to mimic voices, etc. n Receiver performs voice recognition: u Robust to noise and other interferences u Able to focus on a single low-volume speaker in a room full with louder interfering multiple speakers (cocktail party effect). 3/15/2018 Veton Këpuska 21

The Speech-Communication Pathway u Significant advances in reproducing parts of this communication system by synthetic means. u Far from emulating the human communication system. 3/15/2018 Veton Këpuska 22

Analysis/Synthesis Based on Speech Production and Perception u This class does not cover entire speech communication pathway: n Signal measurements of the acoustic waveform. u From these measurements and current understanding of speech of how the vocal tract and vocal cords produce sound waves production models are build. u Starting from analog representations which are then transformed to discrete-time representations (A/D). n From the receiver side the signal processing of the ear and higher auditory levels are covered however to significantly lesser extend only to account for the effect of speech processing on perception. 3/15/2018 Veton Këpuska 23

Analysis/Synthesis Based on Speech Production and Perception u Preview of a speech model. Figure in the next slide shows a model of vowel production. u In a vowel production: n Air is forced from the lungs (by contraction of the muscles around the lung cavity). n Air then flows pass the vocal cord/folds (two masses of flesh) causing periodic vibration of the cords. n The rate of vibration of the cords determines the pitch of the sound. n Periodic puffs caused by vibration of cords act as an excitation input, or source, to the vocal tract. n The vocal tract is the cavity between the vocal cords and the lips: n Vocal tract acts as a resonator that spectrally shapes the periodic input (much like the cavity of the musical wind instrument). 3/15/2018 Veton Këpuska 24

Analysis/Synthesis Based on Speech Production and Perception 3/15/2018 Veton Këpuska 25

Analysis/Synthesis Based on Speech Production and Perception u From this basic understanding of the speech production mechanism a simple engineering model can be build, referred to source/filter model. u In this model the following is assumed: n Vocal tract is a liner time-invariant system (or filter), n This linear time-invariant system is driven by a periodic impulse-like input. n Those assumptions imply that the output at the lips that is itself periodic. 3/15/2018 Veton Këpuska 26

Analysis/Synthesis Based on Speech Production and Perception u Example of a vowel is “a” as in the word “father”. u Vowel “a” is one of many basic sounds of a language called phonemes. u For each phoneme a different production model is built. u Typical speech utterance consists of a string of vowel and consonant phonemes whose temporal and spectral characteristics change with time. u This change corresponds to the changes in n n excitation source and vocal tract system. n n A time-varying source and system, Furthermore, in realty there is a complex non-linear interaction of both. u This fact implies: u Therefore, even though a simple linear time-invariant model seems plausible, it does not always represent well the real system. 3/15/2018 Veton Këpuska 27

Analysis/Synthesis Based on Speech Production and Perception u Using discrete-time modeling of speech production the course will cover the design of speech analysis/synthesis systems as depicted in Figure 1. 3. u Analysis part extracts underlying parameters of time-varying model from speech waveform. u Synthesis takes extracted parameters and models to put back together the speech waveform. u An objective in this development is to achieve an identity system for which the output equals to input when no manipulation is performed. u A number of other analysis/synthesis methods can be derived based on various useful mathematical representations in time or frequency. u These analysis/synthesis methods are the backbone for applications that transform the speech waveform into some desirable form. 3/15/2018 Veton Këpuska 28

Applications (Project Areas) u Speech Modification: n time-scale manipulations: u Fitting the speech waveform n u In Radio and TV commercials into an allocated time slot and the synchronization of audio and video presentation. Speeding up speech – n n u Message playback Voice mail Reading machines and books for the blind Learning a foreign language Slowing down speech – n Voice transformations using Pitch and spectral changes of speech signal: u Voice disguise u Entertainment u Speech synthesis n Spectral change of frequency compression and expansion: u may be useful in transforming speech as an aid to the partially deaf. n Many methods can be applied to music and special effects. 3/15/2018 Veton Këpuska 29

Applications (Project Areas) u Speech Coding n Goal is to reduce the information rate measured in bits per second while maintaining the quality of the original waveform. u Waveform coders: n Represent the speech waveform directly and do not rely on a speech production model. Operate in a high range of 16 -64 kbps n u Vocoders: n Largely are speech model-based and rely on a small set of model parameters. n Operate at the low bit range of 1. 2 -4. 8 kbps n Lower quality then waveform coders. u Hybrid coders: n Partly waveform based and partly speech model-based n Operate in the 4. 8 – 16 kbps range 3/15/2018 Veton Këpuska 30

Applications (Project Areas) n Applications of speech coders include: u Digital telephony over constrained bandwidth channels n n n 3/15/2018 Cellular Satellite Voice over IP (Internet) Video phones Storage of Voice messages for computer voice mail applications. Veton Këpuska 31

Applications (Project Areas) u Speech Enhancement n Goal is to improve the quality of degraded speech. u Preprocess speech before is degraded: n Increasing the broadcast range of transmitters constrained by a peak power transmission limits (e. g. , AM radio and TV transmissions). u Enhancing the speech waveform after it is degraded. n n 3/15/2018 Reduction of additive noise in u (Digital) telephony u Vehicle and aircraft communications Reduction of interfering backgrounds and speakers for the hearing impaired, Removal of unwanted convolutional channel distortion and reverberation Restoration of old phonograph recordings degraded by: u Acoustic horns u Impulse-like scratches from age and wear Veton Këpuska 32

Applications (Project Areas) u Speaker Recognition n Speech signal processing exploits the variability of speech model parameters across speakers. u Verifying a person’s identity (Biometrics) u Voice identification in forensic investigation. n Understanding of the speech model features that cue a person’s identity is also important in speech modification where model parameters can be transformed for the study of specific voice characteristics: u Speech modification and speaker recognition can be developed synergistically. u Speech (Voice) Recognition is covered in ECE 5526 3/15/2018 Veton Këpuska 33

End