Sound 2

Autumn 2016 CSCU9N5 - Sound Slide 1

Physical characteristics of sound (I)

Sound (to a physicist)

– is a pressure wave which travels in air at about 331m/s
• (at 0 degrees: at 343m/s at 20 degrees C)
– with a frequency between 20 and 20,000 Hz
(variations/second)
To a Psychologist...
– Sound is a perceptual effect caused by a pressure wave of
between 20 and 20,000Hz being detected at the ear.

Autumn 2016 CSCU9N5 - Sound Slide 2

Page 1
1
 Physical characteristics of sound (II)

The pressure wave has two physical characteristics:

Amplitude
– the size of the pressure wave – strength of the rarefactions
and compressions in the pressure wave
Frequency
– the number of compressions (or rarefactions) per second
– related to
• the period of the sound, 1/frequency, and
• the wavelength of the sound, speed/frequency

Autumn 2016 CSCU9N5 - Sound Slide 3

Characteristics of real sounds

Sound waveform:
plucked guitar

Frequency
spectrum

Autumn 2016 CSCU9N5 - Sound Slide 4

Page 2
2
 Devices for sound generation and
transduction
For input to a computer, the pressure wave is
– converted to an analogue electrical signal (transduced)
– converted to a digital signal (digitised)
For output from a computer, the digitised signal is
– converted to an analogue signal
– converted to a pressure wave
Microphone Loudspeaker

Computer
ADC System DAC

Autumn 2016 CSCU9N5 - Sound Slide 5

Psychological characteristics of sound

From the perspective of sound being what we hear, sound has

three defining characteristics:
– loudness: how intense the sound is perceived
– pitch: the sense of the sound having a tone
– timbre: the nature of the sound

As befits psychological descriptions, these are inexact.

All sounds have a loudness,

– but many have no pitch
Timbre is often used as a catch-all term to describe those
aspects of the sound not captured by loudness and pitch.
Autumn 2016 CSCU9N5 - Sound Slide 6

Page 3
3
 Some more real sounds

Spanish guitar

Saxophone

Autumn 2016 CSCU9N5 - Sound Slide 7

Pitch and Loudness

Pitch perception is complex

– Complex tones (many frequency components) often have a lower
pitch than a pure tone of the same mean frequency
– Indeed, a low pitched tone may consist entirely of energy at
high frequencies.

Apparent loudness of a sound depends on the frequency as well

as the amplitude of the sound
– human ear responds differently to different frequencies
– young people can often hear higher frequencies than older
people.

Autumn 2016 CSCU9N5 - Sound Slide 8

Page 4
4
 Measuring Loudness

Our ears have (essentially) a logarithmic response

– loudness depends on power:
• proportional to (amplitude * amplitude)
– doubling the power of a sound does not make it twice as loud
– actually, (real, perceptual) loudness is difficult to compute
Decibels
– ratio of the power of two signals is measured in decibels (dB)
– this is a logarithmic scale
– if signal 1 has power P1, and signal 2 has power P2, then
– P2 is 10 log10(P2/P1) dB louder than P1
– e.g. If P2 has 100 times the power of P1, it is 20dB louder

Autumn 2016 CSCU9N5 - Sound Slide 9

Measuring Loudness

Again: 10 log10(P2/P1) dB

0 dB is threshold for a human to hear a sound of 1000Hz (P1)

20dB whisper
90dB loud music
100dB risking damage
140dB aeroplane engine at close range

Autumn 2016 CSCU9N5 - Sound Slide 10

Page 5
5
 Perceived Sound Source Direction
Sound from a single source appears to come from that source
– whether it’s a musical instrument or a single loudspeaker
– even although it gets reflected off walls etc.

The real sound field comes from many sources

– but human auditory scene analysis allows us to detect multiple
sources, and to concentrate on one of them at a time

Autumn 2016 CSCU9N5 - Sound Slide 11

How does this happen?

We appear to use
– information in the fine time structure of monaural sounds to
group sounds together
– the differences in timing and spectral intensity between the
two ears to allow the listener to analyse the auditory scene

Physical correlates of direction are

– for horizontal direction
• IID: interaural intensity difference
• ITD: interaural time difference Earlier &
louder
– for front/back, elevation: spectral shape
– for distance: spectral shape, reflections

Autumn 2016 CSCU9N5 - Sound Slide 12

Page 6
6
 Synthesising Directional Sounds

Synthesised sound can be made to appear to come from

particular directions
– original sound is modified, and different sounds played to each
ear to create this illusion.
– The way in which a particular sound gets modified by the head
on its way to the ear (or eardrum) is called the Head-related
Transfer Function …and there’s one for each ear

Autumn 2016 CSCU9N5 - Sound Slide 13

Binaural Sound

Sound recorded using a synthetic head

Autumn 2016 CSCU9N5 - Sound Slide 14

Page 7
7
 Head related transfer functions (HRTF)

When a sound is played, the stimulus received depends on the

angle of the stimulus …and it is different at each ear

HRTF for each ear:

Stimulus straight ahead
X-axis is frequency
Intensity
Y axis intensity

Frequency
Autumn 2016 CSCU9N5 - Sound Slide 15

Head related transfer functions (2)

HRTF for each ear:

Stimulus 30 degrees Closest ear
X-axis is frequency
Y axis intensity

Head shadow reduces

intensity and can
Furthest ear
alter frequency
spectrum

Autumn 2016 CSCU9N5 - Sound Slide 16

Page 8
8
 Head related transfer functions (3)

HRTF for each ear:

Stimulus 60 degrees
X-axis is frequency
Y axis intensity

Autumn 2016 CSCU9N5 - Sound Slide 17

Head related transfer functions (4)

HRTF for each ear:

Stimulus 90degrees
X-axis is frequency
Y axis intensity.
Note how the main
difference is above
1000 Hz.

By modifying the original sound to mimic the HRTF, a sound can

be made to appear to come from a particular direction.
Autumn 2016 CSCU9N5 - Sound Slide 18

Page 9
9
 Sound Transduction

Whenever sound is transduced, digitised, or reconverted to

analogue, the original signal is altered in some way. When high
quality reproduction is required, we need to keep this alteration
to a minimum.
Transduction:
– Microphones and loudspeakers have a limited frequency response
• they are more sensitive to sounds with certain frequencies
• we would like a flat frequency response from 20 to 20KHz
– They also have a limited dynamic range
• they cannot deal with sounds from the quietest up to the loudest
• the range in energy of everyday sounds is huge
For some applications, we may sacrifice quality
– e.g. telephony: we care really only about comprehensibility
Autumn 2016 CSCU9N5 - Sound Slide 19

Digitising Sound

• Sound is digitised using an analogue to digital converter

(ADC)
• Sound is converted back to analogue using a digital to
analogue converter (DAC)
• Both forms of conversion can introduce alterations in the
sound
– but the ADC is the more problematic.

Analogue to digital conversion has two parameters:

– sampling rate
– sample size

Autumn 2016 CSCU9N5 - Sound Slide 20

Page 10
10
 Sampling Rate

Sampling rate describes how frequently the analogue signal is

converted
– Normally measured in samples/second
• conversion is done regularly, at a fixed number of samples/second
– sampling rate must be at least twice the highest frequency of
interest (Nyquist sampling theorem) otherwise aliasing can
occur
AMPLITUDE

TIME
Autumn 2016 CSCU9N5 - Sound Slide 21

Signal Reconstruction

Quantization
AMPLITUDE

Autumn 2016 CSCU9N5 - Sound Slide 22

Page 11
11
 Signal Reconstruction

Sample and hold reconstruction

AMPLITUDE

Autumn 2016 CSCU9N5 - Sound Slide 23

Aliasing

Aliasing occurs if a sound is sampled too slowly

AMPLITUDE

Better...
AMPLITUDE

Autumn 2016 CSCU9N5 - Sound Slide 24

Page 12
12
 Sample size (I)

Sample size refers to the characteristics of the sample value

taken each sample time, e.g. amplitude
Samples have a fixed length
– 8-bit, (16-bit or 32-bit) which means each sample is a 2’s
complement 8-bit (16-bit or 32-bit) integer
– e.g. range -128 to +127 for 8-bit; -32768 to +32767 for 16-bit
AMPLITUDE

TIME
Autumn 2016 CSCU9N5 - Sound Slide 25

Sample Size (II)

Sampling may be linear or logarithmic

– linear: for sample value x, actual value is (x/maximum)* K for
some K
– logarithmic: provides more resolution at lower levels
• mu-law (µ-law) or A-law
• a form of data compression
LOGARITHMIC
LINEAR
SAMPLE
SAMPLE

SIGNAL SIGNAL

Autumn 2016 CSCU9N5 - Sound Slide 26

Page 13
13
 Sample Size (III)

Major concern for storage of a sampled sound is the total

amount of data collected. Data length is proportional to sample
rate * sample size
– 1 second of sound sampled at 44,100 16 bit samples/second
uses
44,100 * 2 = 88,200 bytes/second
– that is just 1 channel: stereo takes 176,400 bytes/second
• about 10.5Mbytes/minute
– this is CD-audio quality

Data can be compressed

– but decompression must take place in real time
– more on sound data compression in the next sound lecture.
Autumn 2016 CSCU9N5 - Sound Slide 27

Power and Loudness: Dynamic Range

Dynamic range
– loudest measurable signal compared to quietest signal

Measured using decibels

– if signal has power P1, and signal 2 has power P2, then
– P2 is 10 log10(P2/P1) dB louder than P1

For example: 16-bit linear sampling

– Maximum amplitude approx 32000 (and so power is
32000*32000)
– Minimum amplitude is 1 (and so power is 1*1)
– Dynamic range is 10 log10(32000*32000) = 90dB

Autumn 2016 CSCU9N5 - Sound Slide 28

Page 14
14
End of Lecture

Page 15
15