Transcript Document

Multimedia Data Compression
Lecturer
Dr. Aree Ali Mohammed
2012-2013
[email protected]
18-Jul-15
School of Science \ Computer
Science Dept.
1
Overview










Definition
Why Compression
Motivation for Compression
Goals of Compression
Coding Requirements
Compression Types
Text Compression
Image Compression
Audio Compression
Video Compression
18-Jul-15
School of Science \ Computer
Science Dept.
2
Definition

Compression: the process of coding that will
effectively reduce the total number of bits needed to
represent certain information.
18-Jul-15
School of Science \ Computer
Science Dept.
3
Definition



How is compression possible?
 Redundancy in digital audio, image, and
video data
 Properties of human perception
Digital audio is a series of sample values;
image is a rectangular array of pixel values;
video is a sequence of images played out at a
certain rate
Neighboring sample values are correlated
18-Jul-15
School of Science \ Computer
Science Dept.
4
Redundancy



Adjacent audio samples are similar (predictive
encoding); samples corresponding to silence
(silence removal)
In digital image, neighboring samples on a scanning
line are normally similar (spatial redundancy)
In digital video, in addition to spatial redundancy,
neighboring images in a video sequence may be
similar (temporal redundancy)
18-Jul-15
School of Science \ Computer
Science Dept.
5
Human Perception Factors



Compressed version of digital audio, image, video
need not represent the original information exactly
Perception sensitivities are different for different
signal patterns
Human eye is less sensitive to the higher spatial
frequency components than the lower frequencies
(transform coding)
18-Jul-15
School of Science \ Computer
Science Dept.
6
Definition (contd.)

If the compression and decompression
processes induce no information loss, then the
compression scheme is lossless; otherwise it
is lossy.
18-Jul-15
School of Science \ Computer
Science Dept.
7
Why Compression?



Video and audio have much higher storage
requirements than text
Data transmission rates (in terms of bandwidth
requirements) for sending continuous media are
considerably higher than text
Efficient compression of audio and video data,
including some compression standards, will be
considered in this lesson
18-Jul-15
School of Science \ Computer
Science Dept.
8
Motivation for Compression: Discrete Data



Text

Assuming 2 bytes are used for every 8 x 8 pixel character,
 Character per screen page = ...
 Storage required per screen page = ...
Vector Image

Assuming that a typical image consists of 500 lines, each of which is
defined by its coordinates in the x direction and the y direction, and an
8-bit attribute field
 Coordinates in the x direction require ...
 Coordinates in the y direction require ...
 Bits per line = ...
 Storage required per screen page
Bitmap Image

Assuming using 256 colors requiring a single byte per pixel
 Storage required per screen page = ...
18-Jul-15
School of Science \ Computer
Science Dept.
9
Motivation for Compression: Continuous Data


Uncompressed speech of telephone quality
 Assuming being sampled at 8 kHz and quantized
using 8 bit per sample yielding a data stream of 64
Kbit/second
 Storage space required per second = ...
Uncompressed stereo audio signal of CD quality
 Assuming
being sampled at 44.1 kHz and
quantized using 16 bits
 Data rate = ...
 Storage space required per second = ...
18-Jul-15
School of Science \ Computer
Science Dept.
10
Motivation for Compression: Continuous Data

Video sequence
 Assuming
25 full frames per second, luminance and
chrominance of each pixel are coded using 3 bytes,
luminance sampled at 13.5 MHz while chrominance (R-Y and
B-Y) is sampled at 6.75 MHz, each, and samples are
uniformly coded using 8 bits.
 Bandwidth = ...
 Data Rate = ...
 Storage space required per second = ...
18-Jul-15
School of Science \ Computer
Science Dept.
11
Can Multimedia Data be Significantly Compressed?




Redundancy can be exploited to do compression
Spatial redundancy
 correlation between neighboring pixels in
image/video
Spectral redundancy
 correlation among colors
Psycho-visual redundancy
 Perceptual properties of human visual system
18-Jul-15
School of Science \ Computer
Science Dept.
12
History of Bandwidth in Computer Networks
18-Jul-15
School of Science \ Computer
Science Dept.
13
Goals of Compression
18-Jul-15
School of Science \ Computer
Science Dept.
14
Architecture of a Multimedia PC
18-Jul-15
School of Science \ Computer
Science Dept.
15
What Makes “Good” Compression




Quality of compressed and decompressed data
should be as good as possible
Compression/decompression process should be as
simple as possible
Decompression time must not exceed certain
thresholds
[De]/Compression requirements can be divided into
 Dialogue mode (video conferencing)
 Retrieval mode (digital libraries)
 Both
18-Jul-15
School of Science \ Computer
Science Dept.
16
Coding Requirements: Dialogue Mode

End-to-end delay does not exceed 150 ms for
compression and decompression alone.
 Ideally, compression and decompression should not
exceed 50ms in order to ensure natural dialogue.
 In addition
 delay in the network,
 communications protocol processing in the end
system,
 data transfer to and from the respective input
and output devices.
18-Jul-15
School of Science \ Computer
Science Dept.
17
Coding Requirements: Retrieval Mode



Fast forward and fast rewind with simultaneous display
(or playback) of the data should be possible
Random access to single images or audio passages in a
data stream should be possible in less than 0.5 s.
 Maintains interaction aspects in retrieval systems
Decompression of images, video or audio passages
should be possible without interpreting all preceding
data.
 Allows random access and editing
18-Jul-15
School of Science \ Computer
Science Dept.
18
Coding Requirements: Both Modes





Support display of the same data in different systems
 Formats have to be independent of frame size and video
frame rate
Audio and video compression should support different data
rates at different qualities
Precisely synchronize audio and video
Support for economical solution
 Software
 Few VLSI chips
Enable cooperation of different systems
 Data generated on a multimedia system can be reproduced
on another system (e.g. course materials).
18-Jul-15
School of Science \ Computer
Science Dept.
19
Compression Types





Physical versus logical Compression
 Physical
 Performed on data regardless of what information
it contains
 Translates a series of bits to another series of bits
 Logical
 Knowledge-based
 e.g. United Kingdom to UK
Spatial Compression – 2D or single image
Temporal Compression – 3D or video
Codec – Compression / Decompression
Color / intensity … same thing
18-Jul-15
School of Science \ Computer
Science Dept.
20
Compression Types


Symmetric

Compression and decompression roughly use the same techniques and take
just as long

Data transmission which requires compression and decompression on-thefly will require these types of algorithms
Asymmetric

Most common is where compression takes a lot more time than
decompression
 In an image database, each image will be compressed once and
decompressed many times

Less common is where decompression takes a lot more time than
compression
 Creating many backup files which will hardly ever be read
18-Jul-15
School of Science \ Computer
Science Dept.
21
Compression Types


Non-adaptive
 Contain
a static dictionary of predefined
substrings to encode which are known to occur
with high frequency
Adaptive
 Dictionary is built from scratch
18-Jul-15
School of Science \ Computer
Science Dept.
22
Compression Types

Lossless
 decompress(compress(data)) = data
 Used for computer data, medical images, etc.
18-Jul-15
School of Science \ Computer
Science Dept.
23
Compression Types

Lossy

decompress(compress(data))  data

Some distortion

A small change in pixel values may be invisible

Suited for audio and video
18-Jul-15
School of Science \ Computer
Science Dept.
24
Text Compression

Lossless Algorithms



Run Length Encoding
Huffman Coding
Lempel-Ziv Coding
18-Jul-15
School of Science \ Computer
Science Dept.
25
Image Compression

Lossy Compression
18-Jul-15
School of Science \ Computer
Science Dept.
26
Image Compression

Lossy Compression
18-Jul-15
School of Science \ Computer
Science Dept.
27
Audio Compression






Original (64000 bps) This is the original speech signal sampled at 8000
samples/second and u-law quantized at 8 bits/sample. Approximately 4
seconds of speech.
ADPCM (32000 bps) This is speech compressed using the Adaptive
Differential Pulse Coded Modulation (ADPCM) scheme. The bit rate is 4
bits/sample (compression ratio of 2:1).
LD-CELP (16000 bps) This is speech compressed using the Low-Delay
Code Excited Linear Prediction (LD-CELP) scheme. The bit rate is 2
bits/sample (compression ratio of 4:1).
S-ACELP (8000 bps) This is speech compressed using the ConjugateStructured Algebraic Code Excited Linear Prediction (CS-ACELP) scheme.
The bit rate is 1 bit/sample (compression ratio of 8:1).
CELP (4800 bps) This is speech compressed using the Code Excited Linear
Prediction (CELP) scheme. The bit rate is 0.6 bits/sample (compression
ratio of 13.3:1).
LPC10 (2400 bps) This is speech compressed using the Linear Predictive
Coding (LPC10) scheme. The bit rate is 0.3 bits/sample (compression ratio
of 26.6:1).
18-Jul-15
School of Science \ Computer
Science Dept.
28
Video Compression

Lossy Compression (MPEG)






MPEG1
MPEG2
MPEG4
MPEG7
MPEG21
……
18-Jul-15
School of Science \ Computer
Science Dept.
29
Group Discussion
18-Jul-15
School of Science \ Computer
Science Dept.
30