Text Processing ke Multimedia Processing
Zainal A. Hasibuan
Acknowledgment
This slide is mainly taken from Multimedia System Tutorial at Hong Kong University by Romeo CK YIP email: [email protected]
Covered Topics
Fundamentals of Multimedia
What is a Multimedia System? Data types in multimedia system Quality of Service (QoS) Lossless and Lossy Compression
Multimedia System
There is no clear definition of the term "Multimedia System". Multimedia versus Multi Media Common characteristics of a Multimedia System
Digital Interactive Networking Integration of text, graphics, sound, image, and video
What is Multimedia?
media that uses multiple forms of information content and information processing to inform or entertain the (user) audience.[Wikipedia]
Text
Audio
Image
Animation
Video
Interactivity
What is multimedia?
any combination of two or more media, represented in a digital form, sufficiently well integrated to be presented via a single interface, or manipulated by a single computer program. [Chapman]
What is Multimedia?
Key properties of multimedia [Steinmetz] : Discrete and continuous media Independent media Computer-controlled systems Integration
Discrete and Continuous Media
Discrete media composed of timeindependent information items.
Example: text, graphics, or images.
Continuous media consist of timedependent information items.
Example: sound or video.
Discrete and Continuous Media
Continous
Natural Sound
Video
Synthetic Sound
Animation
Discrete
Still Images Text Graphics
Captured (from real world)
Synthesized (by computer)
Software
Applications for different media types Images: image editing, painting and drawing Text: editor, layout programs Video: editing and post production Animation: drawing Sound: editing and effects
Digital Media
Analog Data The signals, which we send each other to communicate, are data. Our daily data have many forms: sound, letters, numbers, and other characters (handwritten or printed), photos, graphics, film. All this data is in its nature analog, which means that it varies in type.
Digital Data A digital system is an electronic unit. Therefore, it can only deal with data, which are associated with electricity. That is accomplished using electric switches, which are either off or on. It can be illustrated with regular household switches. If the switch is off, it reads numeral 0. If it is on, it is read as numeral 1.
Digital Media
Bits Each 0 or 1 is called a bit. Bit is an abbreviation of the expression BInary digit, which is derived from the binary number system.
0 1 0110 01101011 1 bit 1 bit 4 bit 8 bit
Table 1.1 Example of Bits
Digital Media
Binary Number System The binary number system is made up of digits. Compare with decimal system (10 digit system), decimal system uses digits 0 through 9, the binary system only uses digits 0 and 1.
Number in Decimal System 0 2 3 8 Representation in Binary Number System 0 10 11 1000
Table 1.2 Decimal System and Binary System
Digital Media
Units
1 byte = 8 bits 1 kilobytes(kB) = 1024 bytes
1 megabytes(MB) = 1000 kB = 1024000 bytes
Digital Media
Examples of storage media Hard disk, MO, Compact Disc
Pros and Cons of Digital Data Pros: Digital can reproduce, transmit, or copy without introducing loss in quality. Cons: Consume a lot more bandwidth during transmission than analog signal.
Data Types in MM system
Text Data
ASCII stands for American Standard Code for Information Interchange. Computers can only understand numbers, so an ASCII code is the numerical representation of a character.
Text Data
Figure 2.1 ASCII table and description (copyright: www.asciitable.com)
Text Data
ASCII uses 7 bits to represent a character. As a result only 127 characters are defined as standard ASCII characters. Characters 128-255 are called extended ASCII characters (Figure 2.2).
Figure 2.2 Extended ASCII codes (copyright: www.asciitable.com)
Text Data
EBCDIC
EBCDIC (Extended Binary Coded Decimal Interchange Code) is a character set used on early IBM computers. EBCDIC was first introduced in 1965, it was the new character-coding scheme came with IBM System 360 series. EBCDIC uses 8 bits to represent a character.
Text Data
UNICODE
The Unicode character uses 16 bits to represent a character, thus more than 65000 characters can be represented. While 65000 characters are sufficient for encoding most of the many thousands of characters used in major languages of the world.
Sound Data
Sound Data
A typical compact disc can hold up to 74 minutes of 16 bit, 44.1 kHz audio that is uncompressed about 650 megabytes.
Sound Data
Table 2.1 shows how the size of a file is affected by the sampling rate and bit length. The file is a one-minute sound clip, recorded and saved in various forms in the Microsoft Windows WAV file format.
Quality CD Sampling Rate 44 kHz 44 kHz FM Radio 22 kHz 22 kHz AM Radio 11 kHz 11 kHz Resolution 16 bit Stereo 8 bit Stereo 16 bit Stereo 8 bit Stereo 16 bit Stereo 8 bit Stereo File Size 10.3 MB 5.18 MB 5.18 MB 2.59 MB 2.59 MB 1.29 MB
Table 2.1 Variation of file size and sampling rate (60 seconds audio clip in MS WAV format)
Image Data
Image Data
Images, or pictures, are two-dimensional arrays of data called bitmaps, with each element is called pixel.
Image Data
Units Dpi - Dots Per Inch Bit Depth - The number of bits used to hold a pixel. Also called color depth and pixel depth, the bit depth determines the number of colors that can be displayed at one time.
Color Depth
4 bits
Number of Colors
16 256 65,536 16,777,216
Table 2.2 Color Depths
8 bits 16 bits 24 bits
Video Data
Video Data
Video, or moving images, is a sequence of images. To create a sense of continuity, video must be played at a rate of at least 25 frames per second (fps).
Nature of Digital Document
Text (.doc, .txt, .pdf, etc), Image (.jpg, .png, .gif, .bmp, etc), Audio (.wav, .au, .mp3, etc), Video (.avi, .asf, .mov, etc)
Why compression?
Keep more stuff in memory (increases speed) Increase data transfer from disk to memory
[read
compressed data and decompress] is faster than [read uncompressed data]
Premise: Decompression algorithms are fast
True
of the decompression algorithms we
use
Lossless & Lossy Compression
Lossless Compression The original data obtained from lossless compression can be recovered from its compressed data e.g. Huffman Coding
Lossy Compression Compressed data of lossy compression cannot be recovered to the original data. e.g. JPEG, MPEG
Sample Word Frequency Data
(from B. Croft, UMass)
Zipf and Term Weighting
Luhn (1958) suggested that both extremely common and extremely uncommon words were not very useful for indexing.
Summary
Common characteristics of multimedia systems: digital, interactive, distributed, and involving a wide range of data types, such as video and audio.
QoS as an alternative to best effort approach in networking.
