Version 1.1
January, 2010
Table of Contents
Part I Introduction 4
1 What exactly
..............................................................................
is IP? 5
Part II 10 Great Reasons to Broadcast Audio over
IP 8
Part III Broadcast Applications 9
Part IV Types of IP Connections 11
Part V Selecting a Network 16
Part VI Important IP Network Considerations 21
1 Audio over
..............................................................................
IP Transport Protocols 21
2 Choosing
..............................................................................
an Algorithm 23
3 Concealing
..............................................................................
Packet Loss 25
4 Managing
..............................................................................
Jitter (Latency) 28
Part VII Dialing over IP Networks 30
1 NAT and..............................................................................
Port Forwarding 32
Part VIII Planning IP Network Installation 34
1 Regional..............................................................................
Factors Affecting IP Connectivity 34
2 IP Network
..............................................................................
Suitability and Reliability 35
3 Selecting
..............................................................................
a Data Plan 38
4 Redundancy
..............................................................................
Considerations 41
5 IP Interoperability
.............................................................................. 41
6 Checklist..............................................................................
for IP Connections 43
7 Testing a..............................................................................
Network 45
8 Assessing
..............................................................................
Hardware Requirements 47
Part IX Glossary of Terms 49
Part X Trademarks and Credit Notices 51
1 Introduction
Audio-over-IP has proved itself to be the broadcast network infrastructure for
today and into the future. As a consequence, increasing numbers of
broadcasters are migrating to low-cost wired and wireless IP networks from
more costly analog leased line, microwave and synchronous data
technologies like ISDN and X.21.
For many years Tieline Technology has recognised that the future of
broadcasting is in packet-switched networks supporting audio over IP, and
as a member of the Audio-via-IP Experts Group, Tieline has been at the
forefront of determining the direction of broadcasting audio over IP. Tieline
has assisted thousands of broadcasters to seamlessly transition audio
distribution, studio-to-transmitter link and remote broadcast infrastructure
into IP technologies.
The information in this guide is useful to users of all brands of audio codecs
and is supplemented by more detailed information in Tieline's IP and 3GIP
Streaming Reference Manual, which is available for download at www.tieline.
com/transports/Audio-over-IP. You can also contact Tieline support at
support@tieline.com to find out more if you have any further questions or
requests.
None of this is true, and after reading this guide broadcasters should
feel confident that they have sufficient knowledge to configure, run and
monitor broadcast audio connections over IP. The guide provides
information about audio over IP in a logical sequence and will provide:
1. An introduction to IP.
2. A description of the differences between IP networks and
traditional analog leased line and synchronous leased line data
networks.
3. An overview of how audio over IP can be used in different
applications and over different networks.
4. Detailed IP network information and considerations.
5. Recommendations of how to plan your IP network installation
and assess your IP network requirements.
What is IP?
IP stands for Internet Protocol, which is a protocol used to send data
across packet-switched networks. Packet-switching is used by
IP came along with the promise of more efficient use of bandwidth over
computer and wireless networks, but this came at a cost - well two
costs to be exact. The two key factors you need to understand to
manage network reliability are network 'jitter' and packet loss. Jitter
relates to the amount of time required for an audio codec to receive all
the data packets sent to it, then reorder them and play them out in
sequence and reliably stream audio without any audio interruption.
Packet loss relates to data packets sent from one codec to another that
are lost. Lost packets can potentially cause 'artifacts' or glitches in
quality when streaming audio, unless you have the right equipment to
manage it. We will discuss these factors in detail later, but the key
thing to remember is that software developments and improvements to
broadband network infrastructure have mitigated the effects of jitter and
packet loss to a large extent in most situations.
3 Broadcast Applications
IP audio codecs deliver a range of flexible solutions to broadcasters.
Remote Broadcasts
IP codecs are suitable for many different wired or wireless remote
broadcast applications.
Live sports.
Live news reports.
Live radio and television shows.
Live concerts.
Audio Distribution
With the advent of digital radio broadcasting there has been exponential
growth in audio distribution using IP. Multichannel digital radio has
opened the door to new networking and narrowcast opportunities for
radio networks and IP audio distribution delivers a cost-effective and
flexible solution for:
4 Types of IP Connections
IP offers the ability to create much more flexible broadcast networks for a
much lower investment than traditional analog and synchronous digital
networks. Next we outline the three basic audio codec application concepts
important to understanding the capability of broadcasting audio over IP -
unicasting, multicasting and multiple unicasting.
What is Unicasting?
In computer networking a unicast transmission is defined as the
sending of data packets to a single end point or node. A similar
principal is employed in audio over IP broadcasting and a unicast
connection is a one-to-one connection between transmit and receive
audio codecs.
What is Multicasting
IP multicasting is used by broadcasters to deliver a single audio stream
to many recipients. In some ways it is a lot like traditional radio
broadcasting where you transmit a single signal over a wide area and
anyone with a radio can tune in. When multicasting, the audio stream
sent from the transmitting codec is distributed over the IP network to
other codecs and only a minimal amount of bandwidth is required to
Only small sections of the internet are multicast enabled and many
Internet Service Providers (ISPs) block multicast traffic over wide
area networks like the public internet. This restricts most multicast
broadcasts to private local area networks.
Multiple unicasts can be performed over either LANs or WANs and are
most suited to broadcasting over the internet when compared with
multicasting. Multiple unicasting is limited only by the number of
connections the codec is able to dial and the bandwidth available at the
transmitting codec.
They are a great way to send multiple feeds from any broadcast
5 Selecting a Network
IP networks come in various shapes and sizes and the network that is most
suitable for your requirements depends on your broadcast application (e.g.
remote broadcast, audio distribution or STL). In this section we explain the
different types of networks and suggest which ones can be used to perform
studio-to-transmitter links, audio distribution and remote broadcasts.
Wired LANs/WANs/MANs
Ethernet connections to LANs, WANs, MANs are used extensively for
wired IP connections over local, metropolitan and wide area networks.
Wired networks are capable of high data transfer rates and are more
reliable than wireless networks. Depending on data requirements, fiber-
optic cabling is used increasingly for high-bandwidth data networks,
particularly over local area networks. Depending on the network
infrastructure available over private LANs, higher data rates may provide
the opportunity to send uncompressed digital audio at very high bit-
rates.
Wireless 3G Networks
There are basically two different types of 3G networks; UMTS/HSDPA/
HSPA+ and EV-DO. Speeds vary from network to network and are also
affected by the hardware used (i.e. type of antenna) and environmental
factors. The data bandwidth provided by 3G wireless broadband
networks is often sufficient to send up to two channels of high quality
20kHz audio. Wireless 3G networks provide low-delay connections with
typical latency of around 100 to 200 milliseconds.
UMTS/HSDPA/HSPA+
W-CDMA is the technology behind the UMTS (Universal Mobile
Telecommunications System), HSDPA (High-Speed Downlink
Packet Access ) and HSPA+ (also known as HSPA Evolution,
Evolved HSPA, I-HSPA or Internet HSPA) standards for 3G.
These networks are the most suitable for streaming audio over IP
and are typically found in Europe, the Middle East, Africa and
Australia (AT&T in the USA).
EV-DO
EV-DO (Evolution Data Optimised) was evolved from CDMA2000
standards and EVDO Rev 0 can potentially deliver 400 - 1000Kbps
on the downlink and 50 - 100Kbps on the uplink. EVDO Rev A
delivers 600Kbps - 1,400Kbps downlink and 500Kbps-800Kbps
uplink. These networks are typically found in the USA (e.g.
Verizon, Sprint, Alltell).
WiMAX operates using the IEEE 802.16 wireless standard and it has
been developed primarily for medium to long-range outdoor transmission
hops. WiMAX is more efficient than Wi-Fi connections and it has higher
data rates and a greater range.
Satellite IP
Satellite IP connections are a dependable way to send broadcast audio
to the studio from very remote locations where other wireless network
infrastructure is unavailable. Using a BGAN satellite terminal it is
possible to send one or two channels of studio FM quality audio from a
remote location.
6 Important IP Network
Considerations
Packet switching optimizes the use of bandwidth over computer and
wireless networks by dividing data streams into packets with destination
addresses embedded within them. In this way packets are routed through
ISP routing tables to find the best route to their destinations.
The exact form of a packet is determined by the protocol (see Audio over IP
Transport Protocols) a network is using and this affects the actual size of
the packet. Packets are generally split into three parts which include:
Create IP packets
Provide statistics and feedback about IP streams
Establish connections.
UDP (User Datagram Protocol) is the protocol used most commonly for
sending internet audio and video streams and the European
Broadcasting Union (EBU) standard for audio over IP recommends using
RTP over UDP rather than TCP. The UDP protocol is different to the
TCP protocol in that it sends datagram packets. These packets include
information which allows them to travel independently of previous or
future packets in a data stream. In general, UDP is a much faster and
more efficient method of sending audio over IP and RTP over UDP
sometimes has a higher priority than TCP in internet and network
routers. Tieline has written special Forward Error Correction software
(FEC) for UDP data streams, which significantly increases the stability
of a connection over IP.
other devices over the internet and carries SDP messages. It is used to
find call participants and devices even when they move from place-to-
place and is the method used by most broadcast codecs to connect to
competing brands of codec for interoperability. SIP and SDP combine to
negotiate the type of audio coding that can be used over a connection.
Other Protocols
Other important IP protocols are listen in Appendix 1 of this document.
Most audio codecs allow you to select your preferred compression algorithm
using software menus. The algorithm you select will depend on how much
bandwidth you have available and it will affect not only the quality of the
broadcast, but also contribute to the amount of latency or delay introduced.
For example, if MPEG Layer 2 algorithms are used, program delays will be
much longer than when using Tieline Music, MusicPLUS, aptX or AAC
algorithms. This is due to the additional inherent encoding delays involved
when using MP2 algorithms. This can be a major consideration for live
applications where you need bidirectional communications.
The algorithm you choose to connect with will also depend upon:
It is a good idea to listen to the quality of your program signal using each
algorithm and to see how it sounds when it is sent at different connection
bit-rates (as well as different FEC and jitter-buffer millisecond settings). This
will assist you to determine what the best algorithm is for the connection
Packet Loss
Packet loss in IP networks can be caused by:
Concealment
Network protocols like TCP provide for reliable delivery of packets by
asking for retransmission of lost packets. This can be inefficient and
lead to the connection bit-rate being higher than expected if many
packets are lost. Packet loss concealment can also be used to mask
The amount of FEC that you require will depend on how many data
packets are being lost over the network connection and it can only be
used over networks where bandwidth congestion is not an issue. Well
designed codecs let you to manually adjust the FEC setting using
software.
A high quality broadcast codec should provide statistics that allow you
to view how many packets are being lost over the network. This let's you
gauge the amount of FEC that you require to maximise connection
quality and stability. For example, if you are losing one packet in every
five that is sent, and you have a FEC setting of 20%, the lost packets
will be replaced by FEC to maintain the quality of the connection. If you
are losing more packets than this, say one in three, it will be necessary
to increase the FEC setting to 33% to compensate.
You should also consider the remote end too. What is the remote
codec's maximum upload speed? Is the connection shared at either
end? Your bit-rates, FEC settings and buffer rates must be pre-
configured at both ends before you connect, so it's always better to set
your connection speed and balance your FEC according to the available
uplink bandwidth at each end for best performance.
Packet jitter occurs when data packets sent over a network do not arrive
in regular intervals. This occurs because packets can travel over any
route to their destination despite being sent in regular time intervals.
The random delays that occur, and the severity and frequency of these
delays, will be different for every connection. The combination of factors
contributing to the total latency over a network mean that a temporary
buffer is required to ensure reliable play-out of audio streams when
broadcasting.
If a jitter buffer delay setting is not high enough then it is likely that
interruptions to streams will occur as a result of late packets. If the time
value or depth of the jitter buffer is set at a point larger than the longest
experienced jitter delay, then all packets received by a device will be
delivered to the decoder and the best possible audio quality is
recreated.
1. There is no way to predict for sure what the longest jitter delay
will be, and
2. The larger a jitter buffer is (to increase the chance of catching all
late packets) the longer the end-to-end and round trip delay of
data becomes. (In extreme circumstances this can become
unacceptable for bidirectional audio applications that need low
delay)
Certain IP address ranges have been allocated for private use and these
private addresses help to create secure private networks. Private
addresses can be used by anyone on a private LAN but computers or
devices using these numbers are unable to connect directly over the
internet without using Network Address Translation (NAT) and a public
IP address.
If you want to dial a codec with a private IP address you will require
Network Address Translation (NAT). NAT allows a single device, such
as a broadband router, to act as an agent between the public internet
and a local private LAN. Usually this will be set up at the studio end so
you can dial into the studio from the remote codec.
In TCP and UDP IP networks the codec port is the endpoint of your
connection. Software network ports are in a sense doorways for systems to
communicate with each other. For example, several codecs in your studio
may use the same public static IP address. Therefore it is necessary to
allocate port numbers to these codecs so that when an incoming call
comes in, the network knows which codec to send the call to.
Picture a house and imagine the front door is the entry point represented by
an IP address. You want to get to several codecs in different rooms of the
same house and the doors to each of those rooms are represented by
different port numbers. In principle this is how port addressing works. When
a studio with a designated public IP address receives data from several
Whenever possible use wired IP connections that are not being shared with
other devices.
IP Network Alternatives
There are a range of common wired IP networks available for
broadcasting audio over IP
IP Description Recommen-
Network dation
Interface
DSL/ADSL Common and transmits bi-directional Point to Point
(Digital digital data over the public internet using STL/Audio
Subscriber a POTS/PSTN line. Typically uses most Distribution
Line) of the data channel bandwidth to Point-to-Point
download data to a subscriber and will Remote
only transmit data as fast as the DSL/ Broadcasts
ADSL data uplink will provide. The Multicasting
2. You will get the best quality connection if both the local (studio) and
remote codecs use the same Internet Service Provider. This can
substantially increase reliability, audio bandwidth and reduce audio
delay. Using the same service provider nationally can give better
results than using different local service providers. This is especially
true if one of the service providers is a cheap, low-end domestic
service provider, which buys its bandwidth from other ISPs. Second
and third tier providers sub-lease bandwidth from first tier providers
and can result in connection reliability issues due to multiple switch
hops. We also highly recommend using Tier 1 ISPs if connecting two
codecs in different countries.
4. Ensure that the speed of the connection for both codecs is adequate
for the job. The minimum upload speed recommended is 256 Kbps for
a studio codec and 64 Kbps for a field unit connection.
Once you have calculated the total connection bit-rate (64Kbps) and
how high the ISP connection bit-rate needs to be (128Kbps = twice the
connection bit-rate), you can shop around for the most suitable and
The methods employed depend on the hardware being used and the
connections both supported by the codec, and available at the studio and
transmitter sites.
8.5 IP Interoperability
In the past, audio codec manufacturers have designed codecs that have
largely been incompatible with each other in many different situations due to
the use of:
All Tieline codecs are EBU N/ACIP Tech 3326 compatible over IP
and the company is committed to developing new IP and 3GIP
applications that take advantage of emerging network
infrastructures around the globe.
There are two very distinct parts to a call when dialing over IP. The initial
stage is the call setup stage and this is what SIP is used for. The
second stage is when data transference occurs and this is left to the
other protocols used by a codec (i.e. using UDP to send audio data).
SIP can also be used for other elements of a call but it is important to
remember that SIP only defines the way in which a communication
session between devices should be managed. It does not define the
type of communication session that is established.
SIP leverages on the use of web architectures like DNS, and SIP
addresses are similar in appearance to email addresses. A device using
SIP can dial another devices SIP address to find its location. This task
is performed by SIP servers, which communicate between registered
SIP-compliant devices to set up a call.
Check Result
1 Connecting using a reputable Tier1 ISP thats part of
Internet backbone.
2 The same ISP is being used for both codec
connections.
3 The ISP data plan is a Business Plan or equivalent.
4 The ISP connection speed is adequate (e.g. higher
than audio bit-rate plus packet overheads).
5 Equipment is high quality and suitable for media
streaming.
6 The ISP connection speed has been tested.
7 The ISP connection is not shared with PCs or other
devices.
8 UDP is being used as the audio transport protocol.
9 Only 50% of ISP connection uplink bandwidth is being
used.
10 There are no wireless connections being used.
There is a large range of codecs available that are suitable for different
broadcast situations. A sample of these products follows and they can all
connect over 3G wireless broadband networks, wired and wireless LANs,
WANs, the internet, satellite IP, WiMAX and Wi-Fi..
headphone controls/outputs
On-board PA output control with a built-in
telephone hybrid
Wired and wireless IP and POTS codecs
with wireless 3G/3.5GIP, ISDN, X.21,
GSM and Satellite Codec capability
On-board relays and RS-232 with full
studio remote control
9 Glossary of Terms
AES/EBU Digital audio standard used to carry digital audio signals
between devices.
AES3 Official term for the audio standard referred to often as AES/
EBU.
DNS The Domain Name System (DNS) is used to assign domain
names to IP addresses over the World-Wide Web.
Failover Method of switching to an alternative audio stream if the
primary connection is lost.
GUI Acronym for Graphic User Interface
ISP Internet Service Providers (ISPs) are companies that offer
customers access to the internet
IP Internet Protocol; used for sending data across packet-
switched networks.
Latency Delay associated with IP networks and caused by
algorithmic, transport and buffering delays.
Multicast Efficient one to many streaming of IP audio using multicast
IP addressing.
Narrowcast Transmitting a signal or data to a specific recipient or
recipients.
Network A system for forwarding data packets to different private IP
Address network addresses that reside behind a single public IP
Translation address.
(NAT)
Packet A formatted unit of data carried over packet-switched
networks.
Port Address Related to NAT; a feature of a network device that allows IP
Translation packets to be routed to specific ports of devices
(PAT) communicating between public and private IP networks.
Disclaimer
Whilst every effort has been made to ensure the reliability and accuracy
of the information contained in this guide, Tieline is not responsible for
any errors or omissions within it, and the guide should not be relied
upon solely when designing, purchasing and installing broadcast IP
networks. Always consult a qualified and experienced IP broadcast
network professional for advice or to undertake appropriate training prior
to purchasing and installing equipment for use over IP networks.
11 Appendix 1: IP Protocols
Additional IP transport protocols that can affect sending audio over IP
include the following:
Algorithms 23 Jitter 28
Audio distribution 9 Jitter Buffering 28
-C- Latency 28
MANs/WANs/LANs 16
Credit notices 51 NAT 32
-D- Network Address Translation 32
Network considerations 21
Data Costs 38 Network types 16
Data Plans 38 Planning Installation 34
Data Requirements 38 Port Forwarding 32
Disclaimer 51 Private and public networks 30
-E- Redundancy 41
Regional factors 34
EBU N/ACIP Tech 3329 41 Testing connections 45
Error Concealment 25 Transport protocols 21
-F- What is IP 5
Wireless 3G and WiMAX 16
FEC 25
IP Addresses 30
Forward Error Correction
About 25 IP Codecs 47
Conserving bandwidth 25 IP Hardware 47
IP LANs 16
-G- IP MANs 16
Glossary 49 IP Protocols 21
-I- Appendix
IP WANs 16
51
Internet Broadcasting 35
Interoperability 41
-J-
Introduction 4 Jitter 28
IP Jitter Buffering 28
-L- -S-
Latency 28 SAP 51
-M- Satellite IP
SDP 51
16
Multicasting SIP 21
About 11
SIP, how it works 41
Applications 11
SNMP 51
Multiple Unicasts
STLs 9
About 11
Studio to transmitter links 9
Applications 11
-N- -T-
TCP and UDP 21
NAT 32
Testing IP Networks 45
Network Address Translation 32
Trademarks 51
Network Types 35
Networks -U-
Considerations 21
Unicasting
-P- About 11
Applications 11
Packet Loss 25
Planning Installation 34 -W-
Port Forwarding 32 WiMAX 16
Private IP Networks 30 Wireless
Public IP Networks 30 3G 16
-Q- EV-DO
Satellite
16
16
Quality of Service (QoS) 35 UMTS/HSDPA/HSPA+ 16
-R- WiMAX 16
Redundancy 41
Reliability Checks 43
Remote broadcasts 9
RTCP 51
RTP 51
RTSP 51