Audio

F.
4 CIT – Sound page 1
Introduction
Sound is formed by the vibration of particles and transmitted in the form of sound
wave. For example, when we talk, the sound is produced by the vibration of air
particles. The number of vibration of a single tone in 1 second is called the frequency
of the tone, and this is measured in Hertz (1Hz = 1s-1). The sound that we always hear
is composed of waves of different frequencies.
Sound is described by its pitch (frequency) and loudness (amplitude). The human
audible frequency range is from 20Hz to 20kHz.
P ay a t ten t ion. This is a t e s t.
Figure 1. A typical waveform of a speech
Figure 2. Waveform of music played by a piano (~10s) (Music extracted from “2046” by Eric Kwok)
Figure 3. Detail of a short period (~0.04s) of the waveform in Figure 2
Digitization of Sound
Sound information must be converted to an electrical signal before it can be digitized.
This step can be done by a microphone. The electrical signal is then transmitted to and
digitized in the sound card of a computer. In electrical string instruments, the vibration
of the strings produces electrical signal which is then digitized immediately.
The electrical signal has to be sampled and quantized in the digitization step.
1. Sampling
The electrical signal has to be sampled first. The number of samples
produced in 1 second is called the sampling rate or sampling frequency and
it is measured in Hz (or more commonly in kHz).
F.4 CIT – Sound page 2
A higher sampling frequency produces higher sound quality.

Quality Sampling Rate Example of Application
CD or near CD 44.1kHz Audio CD Production
FM Radio 22.05kHz Audio Streaming
AM Radio 11.025kHz Voice Transmission
Table 1. The quality and application of different sampling rate

2. Quantization
The number of bits for storing 1 sample is the sample size of the audio. A
sample size n will give 2n quantization levels.
The most common choice of sample size is 16 bits, as used in CD audio.
This gives 216 or 65536 quantization levels. The minimum acceptable
sample size is 8 bit, which is normally used for applications like voice
communication, where the distortion can be tolerated.
3. Channel
The number of channels determines whether a recording produces 1
waveform (mono) or 2 waveforms (stereo). Stereo sound can provide more
information and give a richer experience, but will double the file size of
mono sound.
4. Bit rate
Bit rate is the amount of information (in bits) transferred in a second. Bit
rate is measured in bps (bits per second). In sampling audio signal, the bit
rate is the number of bits used to stored samples in 1 second.
No. of samples of 1 channel in 1s = Sampling Frequency [Hz]
Total no. of samples in 1s = Sampling Frequency * No. of channels
Bit Rate = Sample Size * Total no. of samples in 1s
Bit Rate = Sampling Freq. * Sample Size * No. of Channels
e.g. Conventional CD recording uses 16 bits to store each sample taken at

44.1kHz. But Super Audio CD uses another new technology which is
known as DSD (Direct Stream Digital). The sampling rate of DSD is 64
times the conventional one and the sample size is reduced to 1 bit.
Calculate the bit rates of conventional CD recording and that of DSD. Give
your answers in bps.
(Information from http://www.sel.sony.com/SEL/consumer/dsd/dsd.pdf)
5. Wave File Size (Uncompressed)
The uncompressed file size of an audio is determined by the number of
samples stored. Samples are stored in the file as an array. This will be a 1
dimensional array for mono sound and 2 dimensional array for stereo sound.
Calculation of File Size:

No. of samples of 1 channel in 1s = Sampling Frequency [Hz]
Total no. of samples in 1s = Sampling Frequency * No. of channels
Total no. of samples in the file = Total no. of samples in 1s * Duration [s]
Size of all samples [bits] = Total no. of samples * Sample Size [bits]
Size of File [bytes] = Size of all samples [bits] / 8
Sampling Freq. ∗ Sample Size ∗ No. of Channels ∗ Duration
File Size =
8
e.g. Calculate at most how many songs of 3.5mins can be stored in a 650MB
CD using the CD audio coding standard.
Common Audio File Types

1. Wave Format (*.wav)
Wave is the standard form for uncompressed audio on a PC. Since a wave file is
uncompressed data - as close a copy to the original analogue data as possible - it
is therefore much larger than the same file would be in a compressed format such
as mp3 or RealAudio. Audio CDs store their audio in, essentially, the wave
format. Your audio will need to be in this format in order to be edited using a
wave editor, or burned to an audio CD that will play in your home stereo.
2. MIDI (*.mid or *.midi)
MIDI stands for Musical Instrument Digital Interface. MIDI is a standard way
for communicating information about music between different electronics
devices, such as computers and sound synthesizers. A MIDI file is generally
much smaller than a wave file because it stores the codes about the instruments
and notes instead of the actual digitized sounds.
Since in the MIDI standard, the ways of producing sounds are not specified, one
file can be played differently in different computers. In order to solve this
problem, General MIDI was produced. This gives a standard which specifies 128
standard voices.
1 Acoustic Grand Piano 44 Contrabass 87 Synth Lead~7
2 Bright Acoustic Piano 45 Tremolo Strings 88 Synth Lead~8
3 Electric Grand Piano 46 Pizzicato Strings 89 Synth Pad~1
… … …
Table 2. General MIDI voice numbers

3. Compressed Audio
By compressing audio signal, the file size can be greatly reduced. Moreover,
most of the compressed audio format supports streaming. This means the
audience can play an audio file through network while the audio data is still
transmitting to the audience. However, since different formats use different
compression algorithms, specific codec (coder-decoder) should be installed

before playing the audio stream.
MP3 (*.mp3)
MP3 stands for MPEG 1 - Audio Layer 3. It is a standard format that
enables audio files to be compressed into small files, by losing sound signal
in the human insensitive range. It should be noted that the more
compression that is used, the more the quality loss will be noticeable.
However, for voice data (such as a recording of a lecture) very high
compression can be used, and the speaker’s voice will remain recognisable.
Typical file size for MP3 is 1Mb/min for near CD quality audio.
MP3 files can be played by using Winamp (www.winamp.com) or Windows
Media Player.
Real Audio (*.ra)
Real Audio is developed by Real Networks (www.realnetworks.com). Like
MP3, Real Audio is also a lossy audio and it supports streaming. It can be
played only with codec developed by Real Networks. The most common
player is Real Player.
Sound Recording and Editing

Wavepad (Freeware): http://www.nch.com.au/wavepad/index.html
Exercise: 1. Record a 20s voice or trim a music to 20s.
2. Try different playback setting (sampling frequency, sample size
and no. of channels).
3. Fade in or fade out the sound.
4. Mix the sound with other audio files.
Figure 4. Screenshot of Wavepad

Audio

Uploaded by

Copyright:

Available Formats

Audio

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Audio

Uploaded by

Copyright:

Available Formats

F.

4 CIT – Sound page 1

P ay a t ten t ion. This is a t e s t.

Figure 1. A typical waveform of a speech

Figure 3. Detail of a short period (~0.04s) of the waveform in Figure 2

A higher sampling frequency produces higher sound quality.

Table 1. The quality and application of different sampling rate

Bit Rate = Sampling Freq. * Sample Size * No. of Channels

e.g. Conventional CD recording uses 16 bits to store each sample taken at

Calculation of File Size:

Common Audio File Types

Table 2. General MIDI voice numbers

compression algorithms, specific codec (coder-decoder) should be installed

Sound Recording and Editing

Figure 4. Screenshot of Wavepad

You might also like