R&D White Paper - BBCdownloads.bbc.co.uk/rd/pubs/whp/whp-pdf-files/WHP109.pdf · Research &...

R&D White Paper

WHP 109

March 2005

Digital Radio Mondiale: DRM, digital radio on long, medium and short waves

– another radio revolution?

J.H. Stott

Research & Development BRITISH BROADCASTING CORPORATION

© BBC 2005. All rights reserved.

BBC Research & Development White Paper WHP 109

Digital Radio Mondiale: DRM, digital radio on long, medium and short waves – another radio revolution?

Abstract

This document publishes the slides used in a lecture given to the IEE Surrey Branch on 17 November 2004.

It gives an overview of DRM: what it is, why it is needed, and who has done it. It goes on to explain some details of the system and how it works (both audio source coding and the COFDM modulation scheme) with, in the case of the lecture itself, some demonstrations not reproducible here.

This is followed by a description of receiver architectures, including some examples of receivers designed or part-designed by the BBC R&D team, and the necessary transmission infrastructure needed by a broadcaster.

Finally an account is given of a range of field trial experiments, with examples drawn from BBC R&D experience.

© BBC 2005. All rights reserved. Except as provided below, no part of this document may be reproduced in any material form (including photocopying or storing it in any medium by electronic means) without the prior written permission of BBC Research & Development except in accordance with the provisions of the (UK) Copyright, Designs and Patents Act 1988.

The BBC grants permission to individuals and organisations to make copies of the entire document (including this copyright notice) for their own internal use. No copies of this document may be published, distributed or made available to third parties whether by paper, electronic or other means without the BBC's prior written permission. Where necessary, third parties should be directed to the relevant page on BBC's website at http://www.bbc.co.uk/rd/pubs/whp for a copy of this document.

White Papers are distributed freely on request.

Authorisation of the Chief Scientist is required for publication.

Research & Development

Kingswood Warren

Welcome to

1


Jonathan Stott

17 November 2004

Kingswood Warren

IEE Surrey Branch

Digital Radio Mondiale

DRM, digital radio on long, medium and

short waves – another radio revolution?

2

Research & Development 3

Contents

DRM: what, who, when & whySome system details – the ‘how’

audio coding; COFDMand how it sounds in practice

ReceiversTransmission infrastructureField trialsConclusions

3


DRM = Digital Radio Mondiale

A consortium with the mission:

“to create a universal, digital system for the AM broadcasting bands below 30MHz”

The digital broadcasting system they have created to fulfil this mission

DRM™ – what is it?

4


Commercial Radio Australia (Australia); Nautel Ltd., Radio Canada International/CBC (Canada); Academy of Broadcasting Science of China, Communications University of China (China); RIZ Transmitters (Croatia); HFCC (Czech Republic); ESPOL, HCJB World Radio (Ecuador); Digita Oy, Kymenlaakso Polytechnik (Finland); CCETT, DRF, Radio France, Radio France Internationale, TDF, Thales Broadcast & Multimedia (France); ADDX, Ahead Software AG, APR, Atmel Germany GmbH, Coding Technologies GmbH, Deutsche Welle, DeutschlandRadio, DLM, Sender Europa 1, Fraunhofer IIS, Georg-Simon-Ohm – University of Applied Sciences Nuremberg, IZT, IRT, Medienanstalt Sachsen-Anhalt/Digitaler Rundfunk Sachsen-Anhalt, Micronas GmbH, Robert Bosch GmbH,Sony International Europe, SWR Südwestrundfunk, TELEFUNKEN SenderSysteme Berlin AG, T-Systems International GmbH, University of Applied Sciences - FH Merseburg, University of Hannover, University of Ulm, VPRT (Germany); Antenna Hungaria, National Communications Authority Hungary (Hungary); Basamad College,Tehran (Iran); Hitachi Kokusai Electric Ltd., NHK (Japan); Libyan Jamahiriya Broadcasting (Libya); Broadcasting Centre Europe, RTL Group (Luxembourg); Asia Pacific Broadcasting Union (Malaysia); Agentschap Telecom, Nozema, Radio Netherlands, Technical University Delft (Netherlands); Radio New Zealand International (New Zealand);Telenor/Norkring (Norway); Radiodifusao Portuguesa (Portugal); RTRN/Voice of Russia (Russia); Government of Catalonia, Universidad del Pais Vasco, (Spain); Swedish Radio International (Sweden); EBU, International Committee of the Red Cross, ITU (Switzerland); Arab States Broadcasting Union (Tunisia); BBC, Christian Vision, Digital One Ltd., Imagination Technologies Ltd., QinetiQ, RadioScape Ltd., VT Communications, WRN (U.K.); Broadcast Electronics, Inc., Dolby Laboratories Incorporated, Dolby LaboratoriesLicensing Corporation, DRS Broadcast Technology (formerly Continental Electronics), Harris Corporation, Broadcast Communications Division, IBB/VOA, Kintronic Laboratories, Inc., National Association of Short-waveBroadcasters, Sangean America, Inc.,TCI, a Dielectric Company, Via Licensing Corporation (U.S.A.); and Radio Vaticana (Vatican City).

DRM has 85 members from 27 countries– major international broadcasters– other broadcasters & network

operators– transmitter manufacturers– receiver manufacturers– research institutes

DRM ConsortiumMembers

5


1997 discussions lead to formation

1998 Consortium officially founded

2001 ETSI Specification published

2003 achieved triple recognition

IEC Standard 62272-1, ETSI Standard ES201980,as well as ITU-R Recommendation BS1514-1

and held ‘official launch’ alongside WRC Geneva

DRM very brief history

6


Starting a new service

BBC faced challenges in 1932 starting Empire Service at HF using AM – plus ça change!

“London calling”

Lord Reith“… neither very interesting

nor very good.”

7


Historically, these bands were used first

they use AM – ‘Ancient Modulation’!

AM listening experience doesn’t match modern expectations – ample scope for improvement

LF, MF, and HF broadcasting bands share the unique & valuable property that signals go a long way

as useful now as it was to the Empire Service

DRM for LF/MF/HF – why?

8


Ground wave

Long wave and medium wave offer good coverage by ground wave

transmitter– works best at lower

frequencies, and with good ground conductivity

ground wave hugs Earth surface

– supports verticalpolarisation

9


Short wave (& MW at night) travels further

refraction in the ionosphere ‘reflects’ waves

Sky wave

transmitter

NB: NOT TO SCALE!

Properties of the ionosphere vary cyclically, on a daily, seasonal and 11-year basisreflection also occurs

at the ground

10


NVIS

Near-Vertical Incidence Sky-Wave

used for national broadcasting

important for large, sparsely populated countries, especially in Tropics

signal sent upwards from transmitter

If frequency is too high, the wave isn’t ‘reflected’. For this reason, certain low-HF bands are reserved for NVIS Tropical Broadcasting

11


Bands below 30MHz have unique long-range coverage – vital for international broadcasting

only way to reach many ‘closed’ countriescannot reach by Internet or local relays

majority of BBC World Service listeners use MF/HF

Vital for some countries’ internal broadcasting

developing, large or tropical countries, using NVIS

Spectrum below 30MHz

12


Bands below 30MHz have unique long-range coverage – vital for international broadcasting

only way to reach many ‘closed’ countriescannot reach by Internet or local relays

majority of BBC World Service listeners use MF/HF

Vital for some countries’ internal broadcasting

developing, large or tropical countries, using NVIS

Spectrum below 30MHz

There’s just one snag – analogue

AM doesn’t meet modern

expectations for quality!

That’s the challenge for DRM…

13


The DRM challenge

DRM had to provide audio quality that could bear comparison with FM stations

while being a drop-in replacement in the spectrum!

frequency{

…

channel pitch, 9 or 10 kHz

AM

14


The DRM challenge

DRM had to provide audio quality that could bear comparison with FM stations

while being a drop-in replacement in the spectrum!

frequency{…

channel pitch, 9 or 10 kHz

AM

DRM

15


DRM: how?

We have to fit a digital signal in a restricted RF bandwidth (9 or 10 kHz), which can adequately describe the audio signal

reduce the bit rate needed to describe the audiousing source coding

squeeze as much data as we can in the limited RF bandwidth (and difficult propagation channel) using advanced modulation & error coding

16


DRM source coding

Two general types are specified, with options

waveform coding (for any programme, incl. music)

speech coding (for speech only, not music)

Waveform coding is AAC plus optional SBR

Speech coding, two options

HVXC (with option of SBR)

CELP

17


Waveform coding

AAC (Advanced Audio Coding)

latest development MPEG 1 layer 2 MP3 AAC

achieves lower bit rate at same qualitybut ideally needs reduced bandwidth (say 7 kHz) in order to keep coding artefacts to acceptable limits

SBR (Spectral Band Replication)

restores the top octave by synthesis in the receiver,transmitting a small amount of helper data

18


Spectral Band Replication

Sounds in the top octave are

high overtones of musical notes in first 7 kHz

noiselike (sibilants, cymbals, wire brush on snare)

Can synthesise this spectral content

‘helper’ flags noiselike/overtone, & spectral shape

receiver reconstructs a sufficient replica

19


Why have speech coders?

General radio programme material has music, street ambience etc as well as speech

AAC waveform coding will normally be used

Some niche applications justify speech codecs

speech codecs require less bit rate, transmit one spoken service very ruggedly?

use for multilingual news within one 9/10kHz slot?

additional spoken service alongside main audio?

20


Modulation & error coding

AAC audio typically in range 15 ~ 25 kbit/s

Need modulation/FEC which fits in 9/10 kHz

has to cope with multipath, Doppler and fading …

Use Coded Orthogonal Frequency-Division Multiplex (COFDM) with Multi-Level Coding

provide ‘knobs’ to adapt to broadcasters’ requirements (trade ruggedness v. capacity)

21


Why COFDM?

… …n+1 n+2 n+3 n+4nn–1n–2n–3n–4

… …n+1nn–1n–2n–3n–4n–5n–6n–7

… …n–2n–3n–4n–5n–6n–7n–8n–9{delayed paths

Multipath causes Inter-Symbol Interference (ISI) between successive data symbols

in principle can handle with an equaliser, but this is very difficult if multipath is long w.r.t. symbol1st path

t

22


Why COFDM?

… …n+1 n+2 n+3 n+4nn–1n–2n–3n–4

… …n+1nn–1n–2n–3n–4n–5n–6n–7

… …n–2n–3n–4n–5n–6n–7n–8n–9{delayed paths

Multipath causes Inter-Symbol Interference (ISI) between successive data symbols

in principle can handle with an equaliser, but this is very difficult if multipath is long w.r.t. symbol

reception of nthsymbol disturbed by fragments of many

others

1st patht

23


COFDM

We reduce the ISI by making symbols longbut reduces data rate

and occupied bandwidth

Get data rate back by using multiple carriersby making them orthogonal we avoidInter-Carrier Interference (ICI)

and adding a Guard Interval TG eliminates ISI

24


COFDM





∫τ+TU

τ

ψm(t)ψ∗

n(t)dt = 0, m �= n,

= TU , m = n

Orthogonality

Pick

so carrier spacing

ψm(t) = ej m ωU t

ωU/2π

= 1active symbol period TU

Just use

FFT!

25


COFDM





∫τ+TU

τ

ψm(t)ψ∗

n(t)dt = 0, m �= n,

= TU , m = n

Orthogonality

Pick

so carrier spacing

ψm(t) = ej m ωU t

ωU/2π

= 1active symbol period TU

Just use

FFT!

n+1n–1 symbol n

Use of guard interval

symbol n n+1n–1

symbol n n+1n–1

symbol n n+1n–1

symbol n n+1n–1 too early!

earliest

medium

latest

too late!

FFT window TU

total symbol lengthTS

TG

26


COFDM





Finally add error-correction coding (the “C”)

too early!

earliest

medium

latest

too late!

27


DRM is for all broadcasters, large & small

bandwidth options (4.5, 5, 9, 10, 18, 20kHz)

ruggedness against capacity trade-off

different constellations (16-QAM or 64-QAM)different code rates

options to match propagation requirements

degrees of multipath (& single-frequency networks)degrees of Dopplershort or long interleaver

DRM options

28


DRM is for all broadcasters, large & small

bandwidth options (4.5, 5, 9, 10, 18, 20kHz)

ruggedness against capacity trade-off

different constellations (16-QAM or 64-QAM)different code rates

options to match propagation requirements

degrees of multipath (& single-frequency networks)degrees of Dopplershort or long interleaver

mode use TU ,ms TG,ms TS ,ms fU ,Hz

AmostlyLF/MF

24 2 2

326 2

341 2

3

B‘workaday’

HF21 1

35 1

326 2

346 7

8

C ‘robust’ HF 14 2

35 1

320 68 2

11

D‘extreme’ HF

(NVIS)9 1

37 1

316 2

3107 1

7

DRM options

29


Let’s hear a real comparison

Broadcast from UK

medium-wave transmitter at Orfordness

Received at Erlangen in SE Germany

night-time, so skywave with fading, and co-channel interference

Alternate brief AM and DRM transmissions

So does it work?

30



Broadcast from UK





So does it work?

Analogue (AM):Low audio bandwidth, fading and co-channel interference

31



Broadcast from UK





So does it work?

Digital (DRM):Would you rather listen

to this or AM?

32


Receivers were needed for both system development and field trials

Present receivers of 2 main types

analogue front end, feeding a computer

with software demodulation and decoding

free-standing hardware

Real consumer receivers coming soon!

DRM receivers

33


Software receivers

conventional superhet radio

down-converter

IF, say 455kHz

soundcard

low IF, say 12kHz

f

computer, with demod/decoding

software

34


Fraunhofer produced PC software used by many DRM members, often with AOR 7030 receiver

Cut-down version made available cheaply to radio enthusiasts via website

FhG software receiver

35


Fraunhofer produced PC software used by many DRM members, often with AOR 7030 receiver

Cut-down version made available cheaply to radio enthusiasts via website

FhG software receiver

There’s also an open-source software-

receiver project called DREAM.It’s free!

36


Inspired by a paper at the IEE HF 2000 Conference, we made a ‘fully-digital’ design

BBC R&D ‘pro’ receiver

37


Inspired by a paper at the IEE HF 2000 Conference, we made a ‘fully-digital’ designA

D

27MHz LPF

digitaldown-

converter

DSP, with demod/decoding software

65Msample/s

BBC R&D ‘pro’ receiver

38


Coding Technologies brought partnerstogether to make small DRM receiver

BBC R&D: DSP card (based on ‘pro’ main board)

AFG: analogue front end

First consumer-size rx

BBC R&D DSP card

39


Coding Technologies brought partnerstogether to make small DRM receiver

BBC R&D: DSP card (based on ‘pro’ main board)

AFG: analogue front end

First consumer-size rx

BBC R&D DSP card

Secondgenerationis smaller

40


Future receivers for consumers must be cheap and low-power

technically possible

huge market ( 2.5 billion AM sets to replace)

Needs DRM ICs

several collaborations are working on this

User interface must be more user friendly

Future receivers

41


User friendliness

DRM has features built in to help this

conveys station name, text message, lists of other frequencies where same service is available, schedule of programmes and frequency changes…

listeners shouldn’t have to know or type frequency

signal structure supports alternative-frequency switching (AFS) or even full frequency diversity

42


User friendliness





Basis of AFSSignal segmented in time,regular periods containing no audio data

frequency f1 audio data, etc audio data, etc t

receiver uses f1 initially… … then reverts to f1

… checks f2 in the ‘gap’… … if f2 better, switchesin next ‘gap’…

frequency f2 audio data, etc audio data, etcReceiver can use periods of non-audio to check second

frequency — note transmissions must be synchronised

43


User friendliness





can use some capacity for other features, e.g. spoken news on demand, mini ‘web’ site

44


Added features

Here is a demonstration we gave this year at the International Broadcasting Convention in Amsterdam

Two low data-rate services were sent (from Orfordness) alongside the main audio

on-demand spoken news headlines

mini broadcast web site

Locally generated tonight!

45


We need broadcast infrastructuremany broadcasters have 2 or more transmittersin a network

single or multi-frequency (SFN or MFN)

DRM-DI Distribution Interface defined to distribute to transmitters and synchronise them

First synchronous MFN set up in June 2003by BBC R&D/VT Merlin

R&D receiver re-programmed as modulator

Broadcast chain

46







Broadcast chainstudio centre

content(audio/data)

content server (makes DRM-DI)

GPSreceiver

GPS satellite network

DRM-DI

uplink station

distribution satellite network

DVB-Scarrying DRM-DI

DRM distribution for wide-spread transmitter network

transmitter site 1

GPSreceiver

satreceiver

modulator (acceptsDRM-DI)

transmitter

transmitter site 2

GPSreceiver

satreceiver


transmitter

coveragearea

DRM RF signal

DRM RF signal

47








content(audio/data)


GPSreceiver


DRM-DI

uplink station




transmitter site 1

GPSreceiver

satreceiver


transmitter

transmitter site 2

GPSreceiver

satreceiver


transmitter

coveragearea

DRM RF signal

DRM RF signal

studio centre

content(audio/data)

content serverGPS

receiver

t & fref

GPS

DRM-DI

48








content(audio/data)


GPSreceiver


DRM-DI

uplink station




transmitter site 1

GPSreceiver

satreceiver


transmitter

transmitter site 2

GPSreceiver

satreceiver


transmitter

coveragearea

DRM RF signal

DRM RF signal

studio centre

content(audio/data)

content serverGPS

receiver

t & fref

GPS

DRM-DI

transmitter site

GPS GPSreceiver

satreceiver

satellitedistribution

modulator

t & f reference

DRM-DI

transmitterDRM

DRM RFsignal

49


Field Trials

BBC engineers took their field-trial brief seriously!

50


Field trials performed throughout project

both channel-sounding and system-testing

short- and long-term, all sorts of paths

long-term tests automatically gathered data from widespread receivers on members’ sites

some work EC-funded in the Radiate project

Ecuador NVIS tests led to extra modes

DRM Field Trials

51


Demonstrations at DRM Launch June 2003

included synchronous 2-frequency transmissions

had to drive kit there anyway, so…

Grasped chance to measure frequency-diversity reception

mobile measurement on journey

static measurement in Geneva

Recent BBC R&D tests (1)

52


Demonstrations at DRM Launch June 2003

included synchronous 2-frequency transmissions

had to drive kit there anyway, so…

Grasped chance to measure frequency-diversity reception

mobile measurement on journey

static measurement in GenevaA pause in France on the way to Geneva


Three BBC R&D receivers logging MFN and frequency-diversity reception

53


Frequency-diversityFrequency 1

Frequency 2

Single-frequencyreception suffers some drop-outs

(poor SNR)

But solid reception using

diversity!

54


Sorting the synchronisation of separated transmitters enabled us to try a big SFN

long-distance HF

2 transmitters

Rampisham in UK (VT Merlin Communications)Bonaire in Netherlands Antilles (R Nederlands W)

Reception in Portugal

It worked!


55



long-distance HF

2 transmitters



It worked!

Recent BBC R&D tests (2)Rampishamtransmitter

Bonairetransmitter

Portugalreceiver

56



long-distance HF

2 transmitters



It worked!

Recent BBC R&D tests (2)Rampishamtransmitter

Bonairetransmitter

Portugalreceiver

Impulse response as received in

Portugal

2ms

note logarithmic vertical scale Bonaire

Rampisham

57



Possible benefits of polarisation diversitywere rediscovered when calibrating antennas

So investigated here and in staff homes

58



Possible benefits of polarisation diversitywere rediscovered when calibrating antennas

So investigated here and in staff homes

59


Polarisation diversity

Each component of the received skywave has dominant polarisation that often rotates

Antennas normally respond to one polarisation only

vertical whip responds to vertical polarisation

thus get polarisation fading

Use crossed loops to get polarisation diversity

60


Polarisation diversity

Each component of the received skywave has dominant polarisation that often rotates

Antennas normally respond to one polarisation only

vertical whip responds to vertical polarisation

thus get polarisation fading

Use crossed loops to get polarisation diversity

Polarisation fading on crossed loops

Voice of Russia received at KW

61


Every silver lining…

Interference

… has a cloud!

other HF radio users

62


Interference remains a worry

especially from systems using the mains wiring

PLT/PLC – Power Line (Tele) Communications

some concern about VDSL on phone wiring

Strong political pressure for PLT

EC and UK in favour (‘Broadband for all’)

happy to turn EMC regulation on its head!

even the IEE position was toned down…

Interference

63


Future of DRM

Broadcasters looking to launch serious services into W Europe in a year’s time

several broadcasters (including BBC World Service) co-ordinating efforts

target France, Germany & Benelux countries

aim to have several services and receivers allavailable at the same time

… then The World!

64


DRM is now ready for introduction

so at last we can have good quality audio andthe long-distance reach of the ‘AM’ bands

DRM uses audio coding & COFDM to squeeze a digital signal into space of one AMtransmission

BBC R&D has played a significant rôle

but there’s plenty more to do!

Conclusions

65


The End

Thanks for listening!

and thanks also to:IEE for inviting me to give presentation,BBC colleagues for help with preparation,Catering staff for refreshments

66


Contact detailsJonathan Stott

BBC Research and Development,Kingswood Warren, TADWORTH,Surrey KT20 6NP, United Kingdom

[email protected]

More papers: www.bbc.co.uk/rdwww.drm.orgIEE Electronics & Communication,February 2002

67

R&D White Paper - BBCdownloads.bbc.co.uk/rd/pubs/whp/whp-pdf-files/WHP109.pdf · Research &...

Documents

Transcript of R&D White Paper - BBCdownloads.bbc.co.uk/rd/pubs/whp/whp-pdf-files/WHP109.pdf · Research &...