The Speech Chain (Denes Pinson, 1993)

Preview:

DESCRIPTION

What information is embedded within the speech acoustic signal? Phonetic information Affective information Personal information Transmittal information Diagnostic Information Tasko SPPA 6010 Advanced Speech Science

Citation preview

Tasko SPPA 6010 Advanced Speech Science

The Speech Chain (Denes & Pinson, 1993)

Tasko SPPA 6010 Advanced Speech Science

What information is embedded within the speech acoustic signal?

Phonetic information Affective information Personal information Transmittal information Diagnostic Information

Tasko SPPA 6010 Advanced Speech Science

Branches of science employed to understand speech communicationPhysics Acoustics Aerodynamics Kinematics Dynamics

Biology Anatomy

Gross anatomy Microscopic anatomy Molecular biology Neuroimaging

Physiology Electrophysiology

Tasko SPPA 6010 Advanced Speech Science

Physical Quantities Basic vs. Derived Scalar vs. Vector Area Volume Displacement Velocity Acceleration Force

Pressure Work Power Intensity Resistance

Ohm’s Law (V=IR)

Tasko SPPA 6010 Advanced Speech Science

Speech anatomy as “tubes” and “valves”

Speech production is achieved through the systematic regulation of air pressures and flows within the lungs and vocal tract.

Tasko SPPA 6010 Advanced Speech Science

Source-Filter Theory of Speech Production

The sounds we hear as speech is the product of a sound source that has undergone filtering by the vocal tract

source and the filter may be considered to be independent of each other

Tasko SPPA 6010 Advanced Speech Science

Source-Filter Theory

Tasko SPPA 6010 Advanced Speech Science

Source-Filter Theory

Tasko SPPA 6010 Advanced Speech Science

Sound: Acoustics review What is sound? Graphic representation of sound Classifying sounds Filters Resonance The decibel

Tasko SPPA 6010 Advanced Speech Science

What is sound? It may be defined as the propagation of a

pressure wave in space and time. propagates through a medium

Tasko SPPA 6010 Advanced Speech Science

What is sound? Mass-spring model

Tasko SPPA 6010 Advanced Speech Science

Wave action of molecular motionTime

1

2

3

4

5

Distance

Tasko SPPA 6010 Advanced Speech Science

Amplitude waveform

Position

Time

Tasko SPPA 6010 Advanced Speech Science

Amplitude waveform

Amplitude

Time

Question: How long will this last?

Tasko SPPA 6010 Advanced Speech Science

Model of air molecule vibrationTime

1

2

3

4

5

Distance

a b c d

Tasko SPPA 6010 Advanced Speech Science

Simple Harmonic Motion: Sine Wave

Features Amplitude Period Frequency

Hz octave

Phase

Pres

sure

Time

Tasko SPPA 6010 Advanced Speech Science

Graphic representation of sound Time domain

Called a waveform Amplitude v. time

Frequency domain Called a spectrum Amplitude spectrum

amplitude vs. frequency Phase spectrum

phase vs. frequency May be measured using a

variety of “window” sizes

Spectrogram frequency v. amplitude v. time

Tasko SPPA 6010 Advanced Speech Science

Same sound, different graphs

Time domain

Frequency domain

From Hillenbrand

Tasko SPPA 6010 Advanced Speech Science

Are all sound waves simply sinusoids?NO! Waves can be summed Simple waves can combine to produce complex waves Fourier: French Mathematician:

Any complex waveform may be formed by summing sinusoids of various frequency, amplitude and phase

Fourier Analysis Provides a unique (only one) solution for a given sound signal Is reflected in the amplitude and phase spectrum of the signal Reveals the building blocks of complex waves, which are sinusoids

Tasko SPPA 6010 Advanced Speech Science

Classification of sounds Number of frequency components

Simple Complex

Relationship of frequency components Periodic Aperiodic

Duration Continuous Transient

Tasko SPPA 6010 Advanced Speech Science

Complex periodic sounds: Graphic appearance

From Hillenbrand

Tasko SPPA 6010 Advanced Speech Science

Complex periodic sounds: Graphic appearance

Tasko SPPA 6010 Advanced Speech Science

Brief Digression

Tasko SPPA 6010 Advanced Speech Science

Amplitude vs. Phase Spectrum

Amplitude spectrum: different

Phase spectrum: same

Tasko SPPA 6010 Advanced Speech Science

Amplitude vs. Phase Spectrum

Amplitude spectrum: same

Phase spectrum: different

Tasko SPPA 6010 Advanced Speech Science

Digression concluded

Tasko SPPA 6010 Advanced Speech Science

Aperiodic sounds: Graphic appearance

From Hillenbrand

Tasko SPPA 6010 Advanced Speech Science

What “class” of sound is speech?

Tasko SPPA 6010 Advanced Speech Science

The “envelope” of a sound wave Amplitude envelope Spectrum envelope

Tasko SPPA 6010 Advanced Speech Science

Amplitude envelope

From Hillenbrand

Tasko SPPA 6010 Advanced Speech Science

Spectrum envelope

From Hillenbrand

Tasko SPPA 6010 Advanced Speech Science

Amplitude Spectrum: Window Size “instantaneous” amplitude spectrum (long term) average amplitude spectrum

Tasko SPPA 6010 Advanced Speech Science

“Instantaneous” Amplitude Spectra

Tasko SPPA 6010 Advanced Speech Science

(Long Term) Average Amplitude Spectrum

Tasko SPPA 6010 Advanced Speech Science

Tasko SPPA 6010 Advanced Speech Science

The Spectrogram

Tasko SPPA 6010 Advanced Speech Science

Rotate90 degrees

F

A F

A

Tasko SPPA 6010 Advanced Speech Science

Rotate it so thatThe amplitude isComing out of thepage

F

AThis is really narrow because it is a slice in time

F

Time

Tasko SPPA 6010 Advanced Speech Science

Dark bands= amplitudePeaks

Time

F

Tasko SPPA 6010 Advanced Speech Science

Two main types of spectrograms Wide-band spectrograms

Akin to spectrum envelopes “lined up” Frequency resolution not so sharp

Narrow-band spectrograms Akin to amplitude spectrums “lined up” Frequency resolution is really sharp

Tasko SPPA 6010 Advanced Speech Science

Highlights harmonic structure

Highlights spectrum envelope

Tasko SPPA 6010 Advanced Speech Science

Filters What is a filter? How are they relevant to speech? Frequency response curve Representing filter operation Types of filters

Tasko SPPA 6010 Advanced Speech Science

Frequency Response Curve (FRC)

Frequencylow high

Gai

n

+

-

Center frequency

lower cutofffrequency

upper cutoff frequency

passband

3 dB

Tasko SPPA 6010 Advanced Speech Science

Operation of a filter on a signal

NOTE: Amplitude spectrum describes a soundFrequency response curve describes a filter

Tasko SPPA 6010 Advanced Speech Science

Source-Filter Theory revisited

Tasko SPPA 6010 Advanced Speech Science

Some frequency selective filtersLow-pass filtersHigh-pass filtersBand-pass filters

Tasko SPPA 6010 Advanced Speech Science

Resonance What is resonance? Free vibration Forced vibration Acoustic resonators Resonance and speech Resonators as frequency selective filters

Tasko SPPA 6010 Advanced Speech Science

Resonance and Speech

Tasko SPPA 6010 Advanced Speech Science

Resonators as frequency selective filters

Tasko SPPA 6010 Advanced Speech Science

Measuring signal amplitude Amplitude vs. loudness Sound intensity vs. sound pressure Decibel scale

Linear vs. logarithmic Absolute vs. relative Reference values Deriving the equations

Recommended