[Mahalakshmi, 3(10): October, 2014] ISSN: 2277-9655

Scientific Journal Impact Factor: 3.449

(ISRA), Impact Factor: 2.114

IJESRT
INTERNATIONAL JOURNAL OF ENGINEERING SCIENCES & RESEARCH
TECHNOLOGY
Data Compression in Multimedia (Text,Image,Audio and Video)
R.Mahalakshmi*, S.K.Mahendran
*Department of Compter Science,Madurai Kamaraj University College, India
Department of Computer Science, Bharathiar University, India


Abstracts
Multimedia files are large and consume lots of hard disk space. The files size makes it time-consuming to


move them from place to place over school networks or to distribute over the Internet.


Compression shrinks files, making them smaller and more practical to store and share.
Compression works by removing repetitious or redundant information, effectively summarizing the contents
of a file in a way that preserves as much of the original meaning as possible.

Keywords: data compression algorithm, text, image, audio and video.

Introduction
Beginning in the 1980s, the capacity and speed of for multimedia has reached it next generation to provide
storage devices have been tremendously improving. smooth service even
Today, new kinds of cheaper and more efficient memory in unreliable network infrastructure.
devices are constantly emerging. Nevertheless, as files Data Compression shrinks down a file so that it
become too large, it is helpful if we could somehow takes up less space. This is desirable for data storage and
encode them into a smaller size and restore them back data communication. Storage space on disks is
when needed later. That way, we can avail more of our expensive so a file which occupies less disk space is
storage media like diskettes, hard disks, CDs, USB Flash cheaper than an uncompressed file. Smaller files are
Disks, and tapes. Most data that we store in our computer also desirable for data communication, because the
devices are digital each unit of information packed in smaller a file the faster it can be transferred.
binary form. This binary nature of the source is how
much information it really contains, and which better
way to represent that information in a smaller number of
binary digits, or bits. The compressed file is ultimately a
concatenation of thousands or even millions of bit
strings. Clearly, the cost of sending data over
communications networks is minimized if the files are
highly compressed.

Data compression
Data compression is the process of encoding data Figure 1: Examples of data compressed as digital like 1’s
using a representation that reduces the overall size of and 0.s to reduce the space
data. This reduction is possible when the original dataset . A compressed file appears to increase the speed of
contains some type of redundancy. Data data transfer over an uncompressed file. Data
compression, also called compaction, the process of compression involves a certain amount of raw data being
reducing the amount of data needed for the storage or processed and stored in a much more compact form. The
transmission of a given piece of information, typically processing is done by computer applications that use
by the use of encoding techniques. Multimedia specific compression algorithms. There are many of
compression is employing tools and techniques in order these, each with its own strengths and weaknesses and
to reduce the file size of various media formats. With the particular areas where they are used.
development of World Wide Web the importance of
compress algorithm was highlighted because it performs Compression and decompression
faster in networks due to its highly reduced file Compressing a file is the process by which a large
size. Furthermore with the popularity of voice and video file is converted to a smaller file, such as a 10 KB
conferencing over the internet, compression method (Kilobyte or 1000 bytes) file converted to a 1 KB file.
http: // www.ijesrt.com (C)International Journal of Engineering Sciences & Research Technology
[376]
[Mahalakshmi, 3(10): October, 2014] ISSN: 2277-9655
Scientific Journal Impact Factor: 3.449
(ISRA), Impact Factor: 2.114

Decompressing a file is the process by which a Data compression algorithms

compressed file is converted back to its standard state, The compression methods can be classified broadly
such as if our 1 KB file were converted back to a 10 KB into lossy or lossless compression. Lossy compression
file. The process by which files are compressed and can achieve a high compression ratio, 50:1 or higher,
decompressed are defined in algorithms, of which there since it allows some acceptable degradation. Yet it
are many. These algorithms are programmed into codecs cannot completely recover the original data. On the other
(COmpression/DECompression) and used by your hand, lossless compression can completely recover the
operating system or software programs to access these original data but this reduces the compression ratio to
files. The block diagram of compression and around 2:1.Lossy compression is most commonly used
decompression as show below. to compress multimedia data (audio, video, and still
images), especially in applications such as streaming
media and internet telephony. By contrast, lossless
compression is typically required for text and data files,
such as bank records and text articles.
A. Lossless Compression
Lossless compression is a class of data compression
algorithms that allows the exact original data to be
reconstructed from the compressed data. The term
Figure 2: Structure of compression and decompression lossless is in contrast to lossy data compression, which
Compression is the reduction in size of data in order only allows an approximation of the original data to be
to save space or transmission time. For data reconstructed, in exchange for better compression rates.
transmission, compression can be performed on just the Most lossless compression programs do two things in
data content or on the entire transmission unit depending sequence: the first step generates a statistical model for
on a number of factors. Content compression can be as the input data, and the second step uses this model to
simple as removing all extra space characters, inserting map input data to bit sequences in such a way that
a single repeat character to indicate a string of repeated "probable" data will produce shorter output than
characters, and substituting smaller bit strings for "improbable" data. Refers to data compression
frequently occurring characters. This kind of techniques in which no data is lost.
compression can reduce a text file to 50% of its original B. Lossy compression.
size. Compression is performed by a program that uses Lossy compression technologies attempt to
a formula or algorithm to determine how to compress or eliminate redundant or unnecessary information. Most
decompress data. video compression technologies, such as MPEG, use a
A simple characterization of data compression is lossy technique. Is a data encoding method which
that it involves transforming a string of characters in discards some of the data, in order to achieve its goal,
some representation such as ASCII into a new string of with the result that decompressing the data yields
bits, which contains the same information but whose content that is different from the original, though similar
length is as small as possible. Data compression has enough to be useful in some way. Lossy compression is
important application in the areas of data transmission most commonly used to compress multimedia data
and data storage. Many data processing applications (audio, video, still images), especially in applications
require storage of large volumes of data, and the number such as streaming media and internet telephony. Lossy
of such applications is constantly increasing as the use compression formats suffer from generation loss:
of computers extends to new disciplines. At the same repeatedly compressing and decompressing the file will
time, the proliferation of computer communication cause it to progressively lose quality. This is in contrast
networks is resulting in massive transfer of data over with lossless data compression.
communication links. Compressing data to be stored or C. Lossless Compression Algorithms
transmitted reduces storage and/or communication costs. There will be no data loss in this type of
When the amount of data to be transmitted is reduced, compression as it is defined by the name. Both original
the effect is that of increasing the capacity of the data and the compressed data are the same in this
communication channel. Similarly, compressing a file to compression. The algorithms for the compression and
half of its original size is equivalent to doubling the decompression are exact inverse of each other in the
capacity of the storage medium. It may then become Lossless Compression. The main mechanism in this
feasible to store the data at a higher, thus faster, level of compression is removing the redundant data in the
the storage hierarchy and reduce the load on the compression and adding them in the decompression. The
input/output channels of the computer system. methods are
http: // www.ijesrt.com (C)International Journal of Engineering Sciences & Research Technology
[377]
[Mahalakshmi, 3(10): October, 2014] ISSN: 2277-9655
Scientific Journal Impact Factor: 3.449
(ISRA), Impact Factor: 2.114

 Run Length Encoding

 Huffman Coding
In this process, images are converted to the
mathematical functions. Discreet Cosine
 Lempel Ziv Encoding Transformation uses series of cosine functions to
Advantages: The original format of the data remains approximate image. This technique is used with JPEG,
even it is compressed. MPEG1 and MPEG 2 formats. A wavelet function is
Disadvantages: Reduction of the size of the data is a used to approximate the image. This can be used with
small. Sometimes the size can increase instead of the JPEG 2000 and MPEG 4 formats.
decrease. ii. JPEG Encoding
i. Run-length Encoding In this encoding, an image is represented by a two
Run-length encoding can be named as the simplest dimensional array of pixels. A Grayscale picture of
method of encoding. Any combination of symbols can 307*200 pixels is represented by 2,457,600 bits and a
be compressed by this encoding technique. The concept color picture is represented by 7,372,800 bits. Due to the
of this algorithm is removing continuous sequence of number of calculations to be had in a JPEG format of a
symbols using another shorten symbol which represent grayscale picture, it is divided into blocks of 8*8 pixels.
it. Practically what is does is to replace the number of The number of the units’ id equal to the number of
occurrences with the numeric value and the symbol mathematical equations of each picture. The whole idea
itself. For example, a text like ‘AAAACBBBBB’ is of JPEG is to change the picture into a linear set of
replace by ‘A04C01B05’. This method is very much numbers that reveals the redundancies. In addition to
efficient when there are only two symbols. those techniques, MPEG is also a Lossy Compression
ii. Huffman Coding technique. It is a way to encode the moving images and
This is a little bit complicated than the previous audio included in it. It supports many video formats from
one. First, the frequency of each appearing character is mobile phone to HD TV.
taken. Then the characters that have least frequencies iii. H.261, H.263, H.264
being taken to give values ‘1’ for the higher one and ‘0’ H.261 is designed for video telephony and video
for the lower one and there wise it is continued. Then the conferencing applications. It was developed in 1988-
each available character is represented by a separated 1990. Data rate is a multiplication of a 64 kb/s. H.263 is
code included with 1s and 0s. a video coding technique for low bit rate
iii. Lempel Ziv Encoding communication. In addition, a 30% of bit saving can be
This system does the encoding as two processes. done by this technique when it is compared to the
They are indexing and compressing a string of symbols. MPEG-1. H.264 is a joint project of ITU-Ts Video
When there is a set of characters, first step is to check Experts Group and the ISO/IEC MPEG group. All those
whether an index have the minimum available character three methods use different methods of reducing
combination that appears. If available then put the index redundant data. There for the output differs from bit rate,
number in the string. If not, it will be added to the index. quality and latency.
This compression technique is more effective as it has a
two way compression method. Text compression
D. Lossy Compression Algorithms Text is a very big part of most files that digital
It is the compression technique which will lose data technology users create. For example, these files could
in the original source while trying to keep the visible be: Word or PDF documents, emails, cell phone texts
quality at the almost same amount. The compression SMS format or web pages. Therefore being able to
ratio will be very high. Most probably the ratio will be a compress text for storage or transmission is extremely
value near 10. It reduces non sensitive information to the important. Fortunately files containing mainly text can
human eyes and the compressed media will not be the be significantly compressed. Like image compression
media that was available before compression. The there are many algorithms or methods that have been

 Mathematical
methods of Lossy compressions are devised to do this. There is one important point to note
and Wavelet about text compression and that is it needs to use a
lossless method. This means the method must not
 JPEG Encoding
Transformation
discard any data when it compresses the data. If this was
 H.x so, the data when it is uncompressed would be
Advantages: Can reduce the file size more than in the incomplete. Some techniques used by general purpose
Lossless Compression compressors such as zip, gzip, bzip2, 7zip, etc, and some

 Static
Disadvantages: The original file cannot be taken after types of models of text compression are,

 Semiadaptive or Semistatic
the decompression
i. Mathematical And Wavelet Transformation
http: // www.ijesrt.com (C)International Journal of Engineering Sciences & Research Technology
[378]
[Mahalakshmi, 3(10): October, 2014] ISSN: 2277-9655
Scientific Journal Impact Factor: 3.449
(ISRA), Impact Factor: 2.114

 Adaptive. 200 by 200 (pixels) saved in the bitmap format (using a

A static model is a fixed model that is known by 24 bits per pixel storage system) can take up
both the compressor and the decompressor and does not 200x200x24 = 960,000 bits or 960,000/8 =
depend on the data that is being compressed. For 120,000Bytes = 117.1975Kilobytes (120000/1024 =

 JPEG
example, the frequencies of symbols in English language 117.1875). Most of compression techniques are,

 GIF
computed from a large corpus of English texts could be

 PNG
used as the model.
A semiadaptive or semistatic model is a fixed
model that is constructed from the data to be Joint Photographic Experts Group created the
compressed. or example, the symbol frequencies standard. The standard specifies the algorithm used for
computed from the text to be compressed can be used as compression and decompression (codec). Lossy
the model. This model has to be included as a part of the compression - some visual quality is lost. The jpg format
compressed data. is good with photographs. It is not so good with high
An adaptive model changes during the contrast pictures like screenshots or computer art. It
compression. At a given point in compression, the model relies on the eye not being sensitive to tiny changes of
is a function of the previously compressed part of the color. Compression can be varied when the image is
data. Since that part of the data is available to the saved. This format recompresses each time it is saved
decompressor at the corresponding point in and repeated saving may lose quality. You should
decompression, there is no need to store the model. For always work with uncompressed formats before saving
example, we could start compressing using a uniform in the target format.
distribution of symbols but then adjust that distribution The Graphics Interchange Format is based on
towards the symbol frequencies in the already processed limiting the colors used in the image. Typically up to 256

 Text compression predates most work on

part of the text. colors are used to make the palette - a table assigning up
to 256 colors to the number 0 - 255. The pixel data for

 Text compression is a kind of data

general data compression. the image is then stored using the 8 bit number that
represents the color’s position in the table.
compression optimized for text (i.e., based Portable Network Graphics uses lossless data

 Text compression can be faster or

on a language and a language model). compression. PNG is an open-source format that was
created to improve upon the GIF. GIF was not open
simpler than general data compression, source and there were licensing costs to developers using
because of assumptions made about the format. This was also a motivating factor in the

 Text compression assumes a language and

the data. uptake of the PNG
Bitmap Jpeg (8KB) Gif (27KB) Png (25KB)

 Text compression is effective when the

language model (118KB)

assumptions are met;

Example 1:
Uncompressed text:
"I am dumb and because I am dumb, I can't
even tell you that I am dumb.
”
Compressed text: Figure 3: various format of image
"$1 and because $1, I can't even tell you that $1. $1=[I The above figure of image is saved in
am dumb]" the bitmap, gif, png and jpeg formats. The last three are
popular web page formats. As can be seen the most
Image compression efficient format is jpeg at 8KB file size because it is
Image compression is an application of data designed to store information on millions of colours that
compression that encodes the original image with few may blend into each other e.g. a photograph or
bits. The objective of image compression is to reduce the complexly coloured image. The least efficient is bitmap,
redundancy of the image and to store or transmit data in and that is why this type of image is not commonly found
an efficient form. Images make up a very big part of the on web pages . Both gif and png are designed to store
(web page traffic) data. The image contains a lot of images that have relatively distinct areas of plain
information compared to text. Even a small image sized colours.
http: // www.ijesrt.com (C)International Journal of Engineering Sciences & Research Technology
[379]
[Mahalakshmi, 3(10): October, 2014] ISSN: 2277-9655
Scientific Journal Impact Factor: 3.449
(ISRA), Impact Factor: 2.114

Example 2: Audio compression is a companion to video

compression, but the techniques are very different. First,
the audio signal requires much less bandwidth than the
video signal. For example, a stereo audio pair sampled
at 48 KHz and using 16-bit samples requires 1.536
Mbps. It would be quite possible to send uncompressed
audio. Moreover, audio compression cannot achieve the
same compression ratio as video; a typical rate for the
stereo audio pair is 192 Kbps, an 8:1 compression ratio.

 MPEG
There are two leading audio compression,

 Dolby AC-3
Figure 4 : Compression of image at 20%
The original size of the image is 3.09MB, it
compress at the level of 20% to reduce the file size as MPEG-1 Layer 2 audio compression, also known
1.30MB. in this to save and share 1.78MB space. It as Musicam, is specified in ISO/IEC IS 13818-3. MPEG
compress at the level of 100% to reduce the file size as audio compression delivers near–CD quality audio using
119.45KB. in this to save 2.97MB space. a technique called sub-band encoding. MPEG audio
compression is used mainly in Europe and is used by
most direct-to-home satellite providers in the United
States. MPEG audio compression is a two-channel
system, but it can encode a Dolby Pro-Logic signal,
which includes two additional channels for rear and
center speakers. (This is analogous to the way in which
a Dolby Pro-Logic signal is carried by a BTSC
[Broadcast Television System Committee] signal as part
of an NTSC transmission).
Dolby AC-3 is a more advanced system than MPEG
Figure 5: Compression of image at 100% audio compression. AC-3 was selected as the audio
compression system for digital television in North
Audio compression America and is specified by ATSC Standard A/52. AC-
Audio compression the amount of data in a 3 encodes up to six discrete channels: left, right, center,
recorded waveform is reduced for transmission. This is and left-rear, right-rear, and sub-woofer speakers. The
used in CD and MP3 encoding, internet radio, and the sub-woofer channel carries only low frequencies and is
like. Audio level compression is the dynamic range commonly referred to as a 0.1 channel because it has
difference between loud and quiet of an audio waveform such a limited frequency range. Thus, AC-3 5.1 gets its
is reduced. This is used in guitar effects racks, recording name from five full channels and a 0.1 channel. In
studios, etc. Audio compression is a form of data addition, AC-3 has a two-channel mode that can be used
compression designed to reduce the size of audio files. to carry stereo and Dolby Pro-Logic encoded signals.
Audio compression algorithms are implemented in
computer software as audio codecs. Generic data Video compression
compression algorithms work with audio data but the Video compression refers to reducing the quantity
compression ratios are low (around 50-60% of original of data used to represent video images. Compressed
size) and they do not work in real time and are therefore video can effectively reduce the bandwidth required to
not practical. Specific audio "lossless" and "lossy" transmit digital video via terrestrial broadcast, via cable,
algorithms have been created. Lossy algorithms provide or via satellite services. It is the process of converting
far greater compression ratios and are used in digital video into a format that takes up less storage
mainstream consumer audio devices. space or transmission bandwidth. It is an essential
technology for applications such as digital television
(terrestrial, cable or satellite transmission), optical
storage/reproduction, mobile television,
videoconferencing, and Internet video streaming. Video
compression standards make it possible for products
from different manufacturers (e.g. encoders, decoders,
and storage media) to interoperate. An encoder converts
Figure 6: Audio compression wave video into a compressed format and a decoder convert’s
http: // www.ijesrt.com (C)International Journal of Engineering Sciences & Research Technology
[380]
[Mahalakshmi, 3(10): October, 2014] ISSN: 2277-9655
Scientific Journal Impact Factor: 3.449
(ISRA), Impact Factor: 2.114

compressed video back into an uncompressed format. plainly impractical. The most popular internet video

 MPEG-1
One of the big advantages of digital video is that it can compression standards are,

 MPEG-4
be compressed for reduced bandwidth applications
including transmission over satellite, cable TV and
Internet-based networks. Compressed video is  DV
particularly useful for reducing storage requirements  H.261 AND H.263
especially in the broadcast and government markets.  DIVX
 FLASH VIDEO
MPEG-1 is a compression standard for audio
and moving pictures, with support for bit rates up to 1.5
Mbit/sec. This is a popular standard for streaming videos
as .mpg files over the internet. The ubiquitous MP3
format, standing for MPEG 1 Audio Layer 3, is a famed
standard for audio compression.
MPEG-4 is a popular object-based
compression standard for multimedia compression.
Objects in a frame are tracked independently and
Figure 7: Various format of video compressed together to form an MPEG4 stream. The
result is an efficient compression standard flexible over
Uncompressed video that we record from a a vast range of bit rates. Its interactivity provides
video camera(e.g. movies) occupies huge amount of developers to control independent objects in a scene with
data. For example, a video clip recorded with a ease.
resolution of 720x576(PAL), with a refresh rate of 25fps DV is a high-resolution digital video format
and 8-bit color depth takes: which employs lossy compression where certain
redundant information in a file is permanently deleted,
720 x 576 x 25 x 8 + 2 x (360 x 576 x 25 x 8) = 1.66 so that even when the file is uncompressed, only a part
Mb/s (luminance + chrominance) of the original information is still there.
H.261 is standard for duplex communication
If we take HDTV(High definition television) which uses over ISDN lines and supports data rates in multiples of
a resolution of 1920x1080 64Kbit/s. The algorithm uses intraframe and interframe
compression. H.263 is an enhanced version of the H.261
1920 x 1080 x 60 x 8 + 2 x (960 x 1080 x 60 x 8) = 1.99 codec that improves quality of video streaming over
Gb/s modems.
DivX Compression is an application that
(Note: In YUV color space, each pixel has one employs MPEG-4 compression standards to facilitate
brightness(luminance) value and two fast downloads over DSL/cable modem, without
color(chrominance) compromising on video quality.
Values) The most popular compression format for
videos on the internet is .flv or Flash Video. FLV and
If no compression was used, a video with F4V are container file formats which play videos on the
dimensions of 640 x 480 pixels (with each pixel using internet through Adobe Flash Player. Encoding of audio
24bits to store its color information) would need and video data for FLV files are same as those for SWF
640x480x24 = 7,372,800bits of memory storage per files. This format is the favorite choice for embedded
frame. If the video was played at a frame rate of 24fps video on the web as it compresses video to low bitrates
then one second's playback would consume and maintains optimum quality.
7,372,800x24 = 176,947,200bits or 22,118,400Bytes or
21,600KB. One minute's playback would need Conclusion
21,600x60 = 1,296,000KB or 1,265.625MB or 1.27GB. Data compression is the art of finding short
This is about 25% the size of a standard DVD disc. So, descriptions for long strings. Every compression
a standard DVD disc could hold about 4 minutes of algorithm can be decomposed into zero or more
uncompressed 640x480 video. Note: this example video transforms, a model, and a coder. Lossless compressors
contains no audio track. If it did more storage would be ignore meaningless data in selecting contexts.
required. So, storing video in an uncompressed format is Meaningless or random data has no predictive value and
is itself not compressible. A lossy compressor not only
http: // www.ijesrt.com (C)International Journal of Engineering Sciences & Research Technology
[381]
[Mahalakshmi, 3(10): October, 2014] ISSN: 2277-9655
Scientific Journal Impact Factor: 3.449
(ISRA), Impact Factor: 2.114

ignores the meaningless data, but also discards it

completely. Data compression is most consideration
thing of the recent world with multimedia. We have to
send a huge amount of data like text file, image audio
and video are in a limited bandwidth. That is why data
has to be compressed using the standard models and its
related techniques. Each and every data gets some type
and different compression ratio(file size).

References
1. S. Bozoki, R.L. Lagendijk, Scene adaptive rate
control in a distributed parallel MPEG video
encoder, in: International Conference on Image
Processing, vol. 2, 1997, pp. 780783.
2. O. Cantineau, J.D. Legat, client parallelization
of an MPEG-2 codec on a TMS320C80 video
processor, in: 1998 International Conference on
Image Processing, vol. 3, 1998, pp. 977980.
3. K. Sayood, Introduction to Data Compression.
4. G. Held and T. R. Marshall, Data and Image
Compression: Tools and Techniques.
5. B. Furht, S. W. Smoliar, H-J. Zang, “Video and
Image Processing in Multimedia Systems”,
Kluwer Academic Pub, 1996.
6. CLARKE, R. J. Digital compression of still
images and video. London: Academic press.
1995, pp. 285-299
7. http://www.newmediarepublic.com/dvideo/co
mpression/adv08.html (3. Feb. 2006)

http: // www.ijesrt.com (C)International Journal of Engineering Sciences & Research Technology

[382]