MULTIPLEXING OF TUTORIALS IN DISTANCE EDUCATION

USING TV BROADCAST NETWORK

Arindan Saha

, Aniruddha Sinha

, Arpan Pal

and Anupam Basu

Innovation Lab, Tata Consultancy Services Ltd., Kolkata, India

Indian Institute of Technology, Kharagpur, India

Keywords: Distance Education, Satellite Broadcast, Metadata in Analog Video, Spatial and Temporal Multiplexing.

Abstract: The demand of distance education in rural areas of developing countries like India is increasing due to

unavailability of skilled teachers in rural India and limited or lack of infrastructural support. Moreover, in

developing countries, the penetration of digital TV receivers is around 10% and remaining 90% uses analog

TV receiver. Hence we propose a distance education solution which not only uses the exiting satellite

broadcasting Television (TV) network but also maximizes the utilization of available bandwidth. In this

paper we present a novel way of multiplexing of multiple tutorials simultaneously into one broadcast

channel which is retrievable by the analog TV receiver. This multiplexing leads to a technical challenge of

embedding multiple audio streams in a single analog channel which is overcome by treating them as a class

of metadata.

1 INTRODUCTION

The poor quality of delivery of educational services

in rural areas of the developing world (Akamai

Technologies, 2010) can be attributed to the paucity

of good teachers in these areas. As a consequence,

the demand for distance education is on the rise in

these areas.

Most of the existing distance education solutions

are based on the broadband internet protocol (IP)

connection (Basu, 2007). The average connection

speed in India for example is 256 kbps (Akamai

Technologies, 2010) and that is also mostly

available in cities and urban areas. Although

recently, wireless broadband connection is peaking

up, the situation in rural areas remains pale due to

the lack of infrastructure. Thus, the available

solutions for distant learning (Franco, 2008),

(Wattamwar, 2011), (Arger, 1990), (Basu, 2007)

cannot be applied.

On the other hand, the local cable operators in

developing countries like India occupy 90% of the

market share in pay TV services (Murray, 2011).

This indicates the dominance of analog client in

households. There are a few solutions proposed

(Vasantha, 2011), (Galajda, 2009) based on satellite

connections, which either need costly specialized

hardware setup or are less suitable for developing

countries.

With the above motivation, we present an

alternate approach for distance education solution

based on the existing television broadcast network

which uses a low cost Home Infotainment Platform

(HIP) (Arpan, 2010) client box capable of taking

analog video-audio as input and providing analog

output to the TV.

Figure 1: HIP based client configuration for distance

education using TV network.

On the client end, the standard digital set-top box

(STB) receives the satellite signal and extracts the

analog video-audio (VA) as shown in Figure 1. The

analog VA is then sent to the HIP client device

which acts as pass-through for the normal TV

channels and extracts the tutorial video, audio and

metadata information for the tutorials channel. In the

absence of the digital STB, the analog signal from

RF cable operators are given to the HIP device via a

327

Saha A., Sinha A., Pal A. and Basu A..

MULTIPLEXING OF TUTORIALS IN DISTANCE EDUCATION USING TV BROADCAST NETWORK.

DOI: 10.5220/0003918003270330

In Proceedings of the 4th International Conference on Computer Supported Education (CSEDU-2012), pages 327-330

ISBN: 978-989-8565-06-8

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

RF demodulator. The desired tutorial is selected by

the user which gets automatically recorded in HIP

upon detection of the broadcast time for the same.

The student can then consume the content in off-line

mode.

In the country like India, it becomes a great

challenge to support multiple states, each having

their own curriculum and language, using a limited

number of broadcast channels. Considering 6 classes

per state, 6 subjects per class, 2 languages per

tutorial and more than 22 states, it comes to 1584

tutorials; if each tutorial is of one hour then it

requires 66 broadcast channels to broadcast 24

hours. This is an infeasible proposition. Hence there

is a need to design a system which will reuse the

standard broadcast infrastructure and standard

receivers to support the above requirement within a

few broadcast channels.

The novelty of the current paper lies in the

architecture for the transmission of multiple tutorials

(video, audio) simultaneously along with the

metadata (questions and their associated answers) in

a single TV channel using the existing infrastructure

of digital broadcast. The robustness of the method of

embedding metadata in analog video frame is

analyzed in (Basu, 2012).

2 CREATION OF BROADCAST

CONTENT

The video, audio and QA are multiplexed to create

the broadcast content. The detail of the multiplexing

is shown in Figure 2. A special video frame is

multiplexed once in every 10 second along with the

tutorial video. The special frame contains the tutorial

information, QA of each tutorial and Electronic

Program Guide (EPG) for the tutorials in the form of

metadata (Basu, 2012). Due to the multiplexing of

the tutorials, the video-audio related metadata are

also inserted as new information in the special frame

which is detailed in section 2.1. The encoded audio

is also treated as metadata and embedded in normal

analog video frame. The video of the tutorials are

multiplexed in time and space in order to achieve

compression. The multiplexed video is encoded

using a standard video encoder (MPEG2) to generate

the elementary streams.

The video of the tutorials are multiplexed in time

and space in order to achieve compression. First

level of compression can be achieved by reducing

the resolution of each tutorial. Normal broadcaster

will send SD (720x576) quality channels, but

tutorial video does not need such high resolution

video as they are of low motion in nature. The idea

is to reduce the resolution to QVGA (320x240) so

we can multiplex 4 channels in spatial domain as a

2x2 array. The resolution of each frame becomes

640x480, which is still less than the SD resolution.

Now the remaining place in each frame is used for

insertion of audio. The encoded audio is treated as

metadata and is inserted as pixel data. The

description of insertion is given in (Basu, 2012).

Second level of compression can be achieved by

reducing the frame rate to half of the normal. If we

consider PAL system, the frame rate of each tutorial

is reduce to 12.5 fps and this reduced frame rate

gives us a chance to multiplex 2 channels in time

domain. In Figure 2 the video encoded to QVGA

resolution is encoded to 12.5 fps streams and finally

two such streams multiplexed together to make final

stream of 25 fps.

Figure 2: Content creation in broadcast end with spatial

and temporal multiplexing.

So totally 4*2 = 8 tutorials can be

simultaneously multiplexed within one channel as

shown in Figure 2 and encoded with a standard

video encoder. This multiplexed stream has one

special frame in every 10 second.

Temporal multiplexing of spatially multiplexed

tutorial sets is done within a Group of Pictures

(GOP) as shown in Figure 3. The normal video

frame in a GOP contains the pilot pattern, frame ID,

audio and its language ID, encoded audio data and

video data as shown in Figure 4. The tutorials are

taken from MIT online courses on Linear Algebra

by Professor W. Gilbert Strang.

CSEDU2012-4thInternationalConferenceonComputerSupportedEducation

328

Figure 3: Temporal multiplexing of tutorials.

Figure 4: Layout of the normal video frame.

2.1 Video Audio related Metadata

The information required for extracting the

multiplexed video and audio is present in the special

frame in every 10 sec. The information encompassed

the following:

 Number of tutorial in the current GOP (8 bits)

 For all the tutorials

o Tutorial ID (32 bits)

o Start offset in frame no. for video from

special frame (8 bits)

o Length of video in frame no. in the

GOP (8 bits)

o X-Y start co-ordinate of video (16+16 =

32 bits)

o Width-height for video (16+16 = 32

bits)

o Indication of end or not (1 bit) for the

tutorial.

o Fixed Marker bits (7 bits)

o Start offset in frame no. for audio from

special frame (8 bits)

 Pilot bit start location X, Y (16+16 = 32 bits).

This is for video frames only.

 It is assumed that every line will start at X point

in the video frame.

 Length of the data inserted in a line from point

X (16 bits)

Thus assuming 8 tutorials are multiplexed, total

required bits for all tutorials are (8 +

(32+8+8+32+32+1+7+8)*8+48) = 1080 bits.

3 EXTRACTION OF TUTORIAL

A standard STB (e.g. TataSky) will able to receive

and decode the channel and generate the analog

video output. The HIP connected to the output of

STB, receives the analog data in video input. The

required tutorial is extracted by de-multiplexing the

input video and recorded as selected by the user. The

metadata is first extracted from the special frame in

HIP box. This metadata provides the QA

information.

Figure 5: Content extraction in HIP client box for satellite

broadcast.

The content extraction and recording process is done

by the following steps and also as shown in Figure 5.

 Detects the special video frame.

 Extracts the EPG from the special video

frame.

 Extracts the required tutorial video based

on user selected input and the tutorial time

information present in EPG.

 Extracts the QA.

 Extracts the NB-AMR encoded audio.

 Encodes the extracted video with H.264

encoder.

MULTIPLEXINGOFTUTORIALSINDISTANCEEDUCATIONUSINGTVBROADCASTNETWORK

329

 Generates the multiplexed AVI file.

The audio and video extraction process from the

normal video frames is shown in Figure 6 whose

steps are given below:

 Synchronization to pilot symbols

 Extraction of frame ID

 Extraction of audio ID and language ID

 Extraction of encoded audio bits if the audio

corresponds to the desired tutorial selected by

the user

 Extraction of the video pixels if that

corresponds to the desired tutorial selected by

the user

 Encoding of the video to H.264

Encoded audio and video is multiplexed to store

the same in local hard-disk for future consumption

by the user.

Figure 6: Extraction process of metadata, audio and video

in normal video frame.

4 CONCLUSIONS

In this paper we propose multiplexing multiple

tutorials in spatial and temporal domain to support

lot more tutorials in a single TV channel to support

distance education using a low cost client HIP box.

The degree of multiplexing will depend on the cost

of each tutorial to broadcast. To extract and play

back the required tutorial it requires some metadata

information and these metadata are inserted as part

of the analog video frame. This metadata includes

pilot pattern, tutorial information and QA. The

encoded audio for multiple tutorials are also inserted

as metadata into analog video frame. The

acceptability of multiplexing multiple tutorials in

spatial and temporal domain depends on the trade-

off between bandwidth and video quality which

needs to be further analyzed.

REFERENCES

Akamai Technologies, Q3 2010, Report on “State of the

Internet”, http://www.akamai.com/stateoftheinternet/

Anupam Basu, Aniruddha Sinha, Arindam Saha, Arpan

Pal, Jan 2012, “Novel Approach in Distance Education

using Satellite Broadcast”, 3rd International

Conference on e-Education, e-Business, e-

Management and e-Learning, IPEDR vol.27 (2012)

IACSIT Press, Singapore

Arpan Pal, C. Bhaumik, M. Prashant, A. Ghose, June

2010, “Home Infotainment Platform”, Proc. of

International Conf. on Ubiquitous Computing and

Multimedia Applications, (UCMA2010), Miyazaki,

Japan

B. B. De Franco and H. C. Oliveira, 2008: “A

Communication Architecture For E-Learning And T-

Learning Integration”, IADIS International

Conference, pp.- 334-338, ISBN 978-972-8924-68-3

G. Arger. 1990: Distance education in the third world:

Critical analysis on the promise and reality. Open

Learning, 5(2):9–18

N. G.Vasantha Kumar; Mohanchur Sarkar; Vishal

Agarwal; B. P. Chaniara; S. V. Mehta; V. S. Palsule;

K. S. Dasgupta, 2011: “Digital Video Broadcast

Return Channel via Satellite (DVB-RCS) Hub for

Satellite Based ELearning, International Journal of

Multimedia & Its Applications”, pp. 133-143, vol. 3,

no. 1

P. Basu, A. H. Thamrin, S. Mikawa, K. Okawa, and J.

Murai, 2007: Internet technologies and infrastructure

for asia-wide distance education. In Proceedings of the

2007 International Symposium on Applications and

the Internet (SAINT’07)

Pavol Galajda; Stanislav Marchevsky; Jn Gamec; Mria

Gamcova; Slavomr Pillar, 2009: “Infrastructure For

Packet Based E-Learning Services Provided Via

Satellite”, Acta Electrotechnica et Informatica, pp.74-

80,Vol. 9, No. 1

Simon Murray, 2011, "India digital TV forecasts", Digital

TV Research http://www.ekmpowershop4.com/ekm

ps/shops/broadbandtv/india-digital-tv-forecasts-1-5-

users-132-p.asp

Sujal Wattamwar, Tavleen Oberoi, Anubha Jindal,

Hiranmay Ghosh, Kingshuk Chakravorty, 2011: "A

Novel Framework for Distance Education using

Asynchronous Interaction", accepted in ACM MM

MTDL

Y-F. Ou, Z. Ma, and Y. Wang, January 2009: "A Novel

Quality Metric for Compressed Video Considering

both Frame Rate and Quantization Artifacts" in Proc.

of VPQM'09, Scottsdale, AZ, USA

ITU-R Rec. BT.500-11, 2002, “Methodology for the

subjective assessment of the quality of television

pictures”

CSEDU2012-4thInternationalConferenceonComputerSupportedEducation

330