### Audio Powerpoint

```Sound can make multimedia
presentations dynamic and interesting.
MM Topic 3 - Audio Data
1
Sound is normally captured using a
microphone.
The microphone converts the analogue sound
into an analogue electrical signal.
However the computer still needs to
understand this signal.
The sound card on a computer contains an Analogue to Digital Converter
(ADC). An ADC is required to turn that analogue signal into a digital signal
that can then be stored and manipulated by a computer.
This is known as digitising or sampling the sound.
Most sound cards also have the capability to take digital input from a CD
or DVD (in this case the data would not pass through the ADC)
MM Topic 3 - Audio Data
2
An example of an electrical sound
signal.
Remember the computer can only handle digital signals so the above
analogue signal must be turned into 1s and 0s. This is done by taking
measurements of the signal and we refer to this as sampling.
MM Topic 3 - Audio Data
3
The number of times each second that the signal is
sampled (measured) is called the sampling frequency.
4
2
3
5
In the figure 5 samples are
taken in 1 second so we say
the signal has been sampled
using a sampling frequency of
5 Hz (Hertz).
1
Example A sample frequency of 44100 Hz means that the sound has been
sampled (measured) 44,100 times per second.
(CD quality uses a sampling frequency of 44100 Hz(44.1 KHz))
MM Topic 3 - Audio Data
4
The number of bits used to record each measurement
is called the sampling depth. (sometimes called the
sampling resolution or sampling size)
In this example each
measurement is stored using
2 bits, giving us 4 possible
levels to record the signal.
The sampling depth
determines the dynamic
range of the recording.
Example Audio CDs use a
sample depth of 16 bits.
MM Topic 3 - Audio Data
5
The span from the lowest amplitude to the greatest
amplitude is known as the dynamic range.
The dynamic range of digital sound is determined by
the sampling depth. The bigger the sampling depth
then the bigger the dynamic range. For example, if you
use a sampling depth of 16 bits then you will have a
dynamic range of 65536 levels of signal. If you have a
sampling depth of 8 bits then you will have a dynamic
range of 256. Obviously the bigger the sampling depth
the better quality of sound.
MM Topic 3 - Audio Data
6
Remember the following
Increasing the sampling depth will
give a better sound but more memory
is required.
Recording in stereo increases sound
quality but will double the file size.
Increasing the sampling frequency will
give a clearer sound but more memory
will be used.
MM Topic 3 - Audio Data
7
One method of converting analogue into digital sound is called PCM (Pulse
Code Modulation).
The data can then be stored in a number of different ways
1 RAW files
RAW files are files which store the sound as it was during sampling. The
data has not been altered in any way and has not been compressed. The
file’s contents consist solely of a string of numeric data with no special
RAW files are so called because no compression has taken place. RAW files
can be extremely large.
MM Topic 3 - Audio Data
8
2 RIFF
RIFF is an example of a container file format. It can contain various types
of multimedia data. The first part of a RIFF file contains a header which
informs the type of data the file stores. If a RIFF file contains sound the
header is called WAVE. Wave files (on PCs with the suffix .wav) are RIFF
files containing digitised sound files.
WAV files can still be very large
3 ADPCM (Adaptive Delta Pulse Code Modulation) is a codec that is used
to compress sound data. This takes sound data that has been encoded into
its normal PCM values and compresses the data so that it requires less disk
space than its raw .WAV equivalent. It is a lossy method of compression. It
works by taking standard audio and only storing the difference between
each sample. These values only use 4 bits compared to 16 bits for the
absolute value. It therefore reduce the amount of disk space required by
MM Topic 3 - Audio Data
9
4 MP3
The most popular standard for general use is MP3 ( Moving Picture Expert
Group Layer 3). This is a lossy compression format. MP3 compresses in
the following ways:
1 An algorithm filters out sound beyond the frequencies of the human ear.
2 Then a technique called Huffman encoding is used. This technique looks
for repeating data and stores this only once. (the data does not have to
be continuous).
3 Next the sampling resolution is reduced and
4 Finally background noise or sounds that are drowned out are not stored.
MP3 can reduce file sizes to one twelfth of their original. MP3 is now a
widely used format used by most mobile music players. (not iPods – they
use a format called AAC (Advanced Audio Coding)).
MM Topic 3 - Audio Data
10
As sound is played digital signals are constantly having to be
converted to analogue in order for us to hear it. The bit rate is the
number of bits that are sent per second to transmit a sound file to
an output device like a speaker or headphone.
If the sound quality is high, then there will be a greater number of bits
as there will be a greater number of samples each second to
converted back to analogue.
The bit rate for sounds can be calculated as follows
Bit rate (bits per second) = sampling depth (bits) * sampling frequency (Hz)
MM Topic 3 - Audio Data
11
A captured signal may not use the full dynamic range that is
available. This means that the sound does not use the available
range of volumes.
Normalisation means taking a
file that wasn’t recorded at the
full volume it could have been
and making it as loud as possible
Sound recorded
(not used full dynamic range)
After normalisation
(full dynamic range used)
MM Topic 3 - Audio Data
12
We use the following formula to calculate
the size of a uncompressed sound file
File size (bits) =
Sampling Frequency (Hz) x Sound time (s) x Sampling Depth (bits) x
Channels(1 for mono, 2 for stereo)
MM Topic 3 - Audio Data
13
File size (bits) = Sampling Frequency (Hz) x Sound time (s) x Sampling Depth
(bits) x Channels(1 for mono, 2 for stereo)
Example What is the file size, in megabytes of a ten second stereo clip recorded at
44KHz and using 16 bit resolution?
= 44000(sampling frequency) x 10 (time) x 16 (resolution or depth) x 2 (stereo)
= 14080000 bits
Convert to bytes (/ 8)
= 1760000 bytes
Convert to kilobytes(/1024)
= 1718.75 kilobytes
Convert to megabytes(/1024)
= 1.68 megabytes
MM Topic 3 - Audio Data
14
File size (bits) = Sampling Frequency (Hz) x Sound time (s) x Sampling Depth
(bits) x Channels(1 for mono, 2 for stereo)
Example A stereo has to be recorded at CD quality. The song is 3 minutes and 12
seconds long. How much disc space would be required to store the captured file?
Hint: Settings for CD quality are Sampling Frequency 44100 Sampling Depth 16 bits
File size (bits)
= 44100 x 192 x 16 x 2
= 270950400 bits
Convert to bytes (/8)
= 33868800 bytes
Convert to kilobytes (/1024)
= 33075 kilobytes
Convert to megabytes (/1024)
= 32.3 megabytes
MM Topic 3 - Audio Data
15
Clipping
If a sound file does not sound too good – perhaps part of the sound
seems unclear or missing – the most probable cause of this is clipping.
Clipping can occur when a sound is recorded at too high a level and
part of the waveform is cut off. What happens is that the sound
goes out with the range of digital codes used to represent the sound
i.e. out with the dynamic range.
MM Topic 3 - Audio Data
16
Stereo
A sound recorded in stereo has been recorded over two channels. This is
done by using two microphones – one for the left channel and one for the
right channel. It will be of higher quality than a sound recorded in mono.
It will also be double the file size.
When editing a stereo sound in a waveform editor, both sounds are
displayed as separate waveforms.
Editing a stereo sound
Editing a mono sound
MM Topic 3 - Audio Data
17
Surround Sound
In cinemas or home entertainment systems a combination of speaker are
used to create an effect called surround sound. Surround sound make the
file more life like.
Surround sound is created by software in the sound card using special
algorithms to create the effect. The algorithms filter the original sound
into several different components based on how they will be heard from a
number of different directions. The filtered sounds are then sent to the
appropriate speaker.
Dolby Surround sound – 4 speakers two in front of the listener and two
behind the listener.
5.1 – 5 speakers, forward left and right, rear left and right and centre
forward speakers. The .1 refers to an extra device called a subwoofer
that plays low frequencies. Eg making the room rumble when there is an
explosion.
MM Topic 3 - Audio Data
18
This means to slowly reduce the sound so that it dies away slowly rather
than coming to a sudden stop. The rate at which the sound fades and its
duration can be changed by the user. Most sound editing software, like
Audacity, comes with fade settings. Fade settings can be referred to as
envelopes.
MM Topic 3 - Audio Data
19
A sound card is needed to generate sounds that can be fed to
DAC takes the digital data and changes it to an analogue signal which is
fed out through the sound line out socket on the sound card. (green).
DSP (Digital Signal Processor) is an integrated circuit designed for high
speed data manipulation. Main function to compress and decompress
sound files as well as to provide enhancements to sounds.
MM Topic 3 - Audio Data
20
MIDI (Musical Instrument Digital Interface) is an interface which
allows a PC to control an instrument or to communicate with it.
MIDI has also become a standard file type for musical files that
contain data that enable musical instruments to recreate music.
MIDI files do NOT contain any recording of sound, no digitised
sound is stored. All they contain is the data required by an
instrument to synthesise the sound.
MM Topic 3 - Audio Data
21
A MIDI file can be a text file. The file can be edited in text
editor like Notepad. Just as SVG graphics are defined a MIDI
text file has similarities.
MIDI instructions contain a description of each note in terms of
its pitch, timing, instrument type, duration, channel, tempo and
volume.
MIDI files are not as high quality as recordings and MIDI cannot
create some sounds such as speech.
MM Topic 3 - Audio Data
22