LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall...

31
LING 520 Introduction to Phonetics I LING 520 Introduction to Phonetics I Fall 2008 Fall 2008 Week Week 4 4 Basic acoustics Fourier transform and LTI systems Recording and sampling Oct. 1, 2008

Transcript of LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall...

Page 1: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics ILING 520 Introduction to Phonetics I

Fall 2008Fall 2008

WeekWeek 44

Basic acoustics

Fourier transform and LTI systems

Recording and sampling

Oct. 1, 2008

Page 2: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

2

02

2

=+ kxdt

xdm

mkùtAx =+= 00 ,)cos(

Page 3: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

3

Page 4: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

4

• displacement: momentary distance from restpoint B

• cycle: one complete oscillation

• amplitude: maximum displacement

• frequency: number of cycles per second (hertz or Hz)

• period: number of seconds per cycle

• phase: portion of a cycle through which a waveform hasadvanced relative to some arbitrary reference point

Page 5: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

5

Page 6: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

6

Page 7: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

7

Page 8: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

8

wave 1: 440Hz

wave 2: 330Hz

wave 1 + 2: ? Hz

Page 9: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

9

Page 10: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

10

25

0

2

00 N/m102 ,)log(10)log(20 === pppppSPL

212

00 rwatts/mete10 ),/(log10 == IIIIL

Page 11: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

11

Page 12: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

12

• Fourier's theorem:

Any periodic signal (must satisfy Dirichlet conditions) iscomposed of a superposition of pure sine waves, whosefrequencies are harmonics of the fundamental frequency of thesignal.

• The fundamental frequency F0 determines the perceived pitchof the sound (In reality, perceived pitch is determined by thespacing harmonics as much as by F0)

• The harmonic frequencies change the quality or timbre of thesound.

...3sin2sinsin

...3cos2coscos

)sincos()(

321

3210

1

0

++++

++++=

++==

xbxbxb

xaxaxaa

nxbnxaaxfn

nn

Page 13: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

13

• A waveform can be represented as a spectrum showing:

• Transmission of Component frequencies on the x-axis• Intensity (Amplitude) of each component on on the y-axis

Page 14: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

14

• A spectrogram displays:

• Time along the x-axis• Component frequencies along the y-axis• Intensity (Amplitude) using darkness

Page 15: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

15

H (T{.})x [n] y [n]

Page 16: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

16

]}[{]}[{]}[][{ 2121 nxbTnxaTnbxnaxT +=+

][][ 21 nbxnax + ][][ 21 nbynay +

][ 01 nny

][2 nx][1 ny ][2 ny

][ 01 nnx

][1 nx

][1 nx ][1 ny

Page 17: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

17

==

otherwise 0

0 if 1][

nn

)][][(][=

=k

knkxnx

][n ][nh

==k

knkxTny ][][][ { }=

=k

knTkxny ][][][

][*][][ nnxnx =

Page 18: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

18

Page 19: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

19

• The Fourier transform of a convolution is the product ofthe Fourier transforms

• The Fourier transform of a product is the convolution ofthe Fourier transforms

Page 20: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

20

Recording

Page 21: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

21

Recording

Page 22: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

22

Recording

Page 23: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

23

Recording

• Classroom recording (SNR 29 dB)

• Laptop recording (SNR 44 dB)

• Professional recording (SNR 90 dB)

[From: Chilin Shih]

Page 24: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

24

Recording

[From: Chilin Shih]

Page 25: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

25

Sampling

• In order to transform sound into a digital format, you must sample the

sound. The computer takes a snapshot of the sound level at small time

intervals while you are recording.

• The number of samples taken each second is called the sampling

rate. The more samples that are taken, the better sound quality. But we

also need more storage space for higher quality sound.

44100 Hz

22050 Hz

11025 Hz

8000 Hz

5000 Hz

Page 26: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

26

Page 27: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

27

Page 28: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

28

Page 29: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

29

Quantization

Page 30: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

30

Sound file formats

Page 31: LING 520 Introduction to Phonetics I · 2008. 9. 29. · LING 520 Introduction to Phonetics I, Fall 2008 12 • Fourier's theorem: Any periodic signal (must satisfy Dirichlet conditions)

LING 520 Introduction to Phonetics I, Fall 2008

31

Sound file formats

wav 550K

mp3 51K

wav->mp3->wav

[From: Chilin Shih]