This document discusses audio compression techniques. It begins by defining audio and compression. There are two main types of audio compression: lossy and lossless. Lossy compression reduces file sizes but results in some quality loss, while lossless compression decompresses the file back to its original quality. Common lossy audio compression methods are discussed, including those based on psychoacoustics involving how humans perceive sound. MPEG layers are then introduced as a standard for audio compression, with Layer I being highest quality but also highest bitrate, and Layer III providing greater compression but still high quality at lower bitrates like 64kbps. Effectiveness is shown to increase with each newer layer.

1. Audio Compression
Prepared by : Darshansinh Joddha

2. Outline
1.Introduction
2.Audio compression
3.Types of compression
4.Audio Compression Methods
5.Psychoacoustics
6.Algorithm
7.MPEG layers
8.Effectiveness of MPEG audio

3. Introduction
➢ What is audio?
● Sound is a sequence of naturally analog signals that are converted
to digital signals by the audio card, using a microchip called an
analog-to-digital converter (ADC).
● An audio file is a record of captured sound that can be played
back.
● When sound is played, the digital signals are sent to the speakers
where they are converted back to analog signals that generate
varied sound.

4. Introduction
➢ What do compression mean?
● Compression is the reduction in size of data in order to save space or
transmission time.
● Compression can be used to:
✔ Reduce file size
✔ Save disk space
✔ Reduce transmission time
● Compression is performed by a program that user an algorithm to
determine how to compress or decompress data.

5. Audio Compression
➢ What is audio compression?
● Audio compression is form of data compression designed to reduce the
size of audio file.
● There is a conditions on this definition:
✔ The audio file must still be playable after compression, without
decompressing it to original size when you want to play it (for
example with WinRAR).
✔ If the file is compressed 'too much' there will be loss of quality.
✔ The compression is done with a thing called a codec. This is an
aggregation of the words: compressor and decompressor.
✔ This codec is a special algorithm to reduce the size.

6. Type of Compression
➢ Lossy compression
● A compression technique that does not decompress data back to 100%
of the original.
● Lossy methods provide high degrees of compression and result in
smaller compressed files, but there is a certain amount of visual loss
when restored.
➢ Losless compression
● A compression technique that decompresses data back to its original
form without any loss.
● The decompressed file and the original are identical. For example, the
ZIP archiving technology (WinZip...) is the most widely used lossless
method.

7. Audio Compression Methods
➢ The following are some of the Lossy methods applied to audio
compression:
● Silence Compression - detect the "silence", similar to run-length
coding.
● Adaptive Differential Pulse Code Modulation (ADPCM)
a) Encodes the difference between two consecutive signals,
b) Adapts at quantization so fewer bits are used when the value is
smaller.
● Linear Predictive Coding (LPC) fits signal to speech model and then
transmits parameters of model. Sounds like a computer talking, 2.4
kbits/sec.
● Code Excited Linear Predictor (CELP) does LPC, but also transmits
error term - audio conferencing quality at 4.8 kbits/sec.

8. Psychoacoustics
➢ These methods are related to how humans actually hear sounds:
● Human hearing and voice
➢ Range is about 20 Hz to 20 kHz, most sensitive at 2 to 4 KHz.
➢ Dynamic range (quietest to loudest) is about 96 dB.
➢ Normal voice range is about 500 Hz to 2 kHz.
● Low frequencies are vowels and bass
● High frequencies are consonants

9. Psychoacoustics
● Frequency Masking
➢ A frequency component can be partly or fully masked by
another component that is close to it in frequency.
➢ A lower tone can effectively mask higher tone
● Temporal masking
➢ A quieter sound can be masked by a louder sound if they
are temporally close sounds that occur both(shortly) before
and after volume increase can be masked.

10. Psychoacoustics
●
if two sound events occur within milliseconds of each other, we're
only going to be able to focus on the loudest one. It's how we've
been evolutionarily primed to react. Our ears and minds can't
separate events that close in time.

11. Algorithm
1. Use convolution filters to divide the audio signal (e.g., 48 kHz sound)
into frequency sub bands that approximate the 32 critical bands ->
sub-band filtering.
Audio
Samples
Sub-band filter 0
Sub-band filter 1
Sub-band filter 2
Sub-band filter 31
.
.
.
12
samples
12
samples
12
samples
12
samples
12
samples
12
samples
Layer I
Frame
Layer II, III
Frame
.
.
.
.
.
.

12. Algorithm
2. Determine amount of masking for each band caused by nearby band
using the results shown above (this is called the psychoacoustic
model).
4. If the power in a band is below the masking threshold, don't encode it.
3. Otherwise, determine number of bits needed to represent the
coefficient such that noise introduced by quantization is below the
masking effect.
4. Format bitstream.

13. Algorithm
Filter into
Critical Bands
(Sub-band filtering
Compute
Masking
(Psychoacoustic
Model)
Allocate bits
(Quantization)
Format
BitStream
Input Output

14. MPEG layers
➢ MPEG defines 3 layers for audio. Basic model is same, but codec
complexity increases with each layer.
➢ MPEG layer I
● Filter is applied one frame (12x32 = 384 samples) at a time. At 48
kHz, each frame carries 8ms of sound.
● Psychoacoustics model only uses frequency masking.
● Highest quality is achieved with a bit rate of 384Kbps.
● Typical applications : Digital recording on taps, hard disks, or
magneto-optical disks, which can tolerate the high bit rate.

15. MPEG layers
➢ MPEG layer II
● Use three frames in filter (before, current, next, a total of 1152
samples). At 48 kHz, each frame carries 24 ms of sound.
● Models a little bit of the temporal masking.
● Highest quality is achieved with a bit rate of 256k bps.
● Typical applications: Audio Broadcasting, Television, Consumer
and Professional Recording, and Multimedia.

16. MPEG layers
➢ MPEG layer III
● Better critical band filter is used
● Uses non-equal frequency bands
● Psychoacoustic model includes temporal masking effects, takes
into account stereo redundancy, and uses Huffman coder.
● MP3 stands for MPEG Layer III

17. Effectiveness of MPEG audio
Layer Target bit-
rate
Ratio Quality* at
64 kbps
Quality at
128 kbps
Layer I 192 kbps 4:1 -- --
Layer II 128 kbps 6:1 2.1 to 2.6 4+
Layer III 64 kbps 12:1 3.6 to 3.8 4+
*Quality factor:
– 5 – perfect
– 4 - just noticeable
– 3 - slightly annoying
– 2 – annoying
– 1 - very annoying

18. Thank You