It seems not that long ago the great advances in audio coding technology and lower cost satellite equipment made it feasible for many radio networks around the world to distribute their programming by satellite. Many of these radio networks uplinked directly from the radio station studios to feed AM / FM / MW / SW and cable systems directly in order to by-pass major shared hub operators. The economics were first driven APT-X and later by MPEG Layer II audio coding and compression technologies implemented on relatively low cost equipment.
Today, many of these distribution networks are still in operation with 10 to 15 year old equipment that has withstood the test of time. However, the manufacture and supply of this old generation equipment has generally stopped long ago and even repair of this old equipment is becoming more of a problem. The age and the lack of support for the equipment answers the Why should I replace the distribution system question for most station engineers. The more difficult questions in front of many station engineers to-day are How and when to replace their satellite distribution systems. and What should the new satellite distribution system architecture be to carry me through the next 10 years?
A survey of a number of radio engineers yielded a diverse wish list for the new network infrastructure.
- Allow for a flexible number of channels and bandwidth
- Decrease distribution costs per channel and improve quality
- Allow broadcaster to better implement localization
- Increase availability by eliminating sun and rain outages
- Permit advertisement and commercial insertion
- Integrate better with new terrestrial IP systems
- Easier integration with my broadcast automation system, the internet, DVB-T, DVB-H, DMB, web servers, telephone systems and HD radio
- Be capable of multichannel sound
- Re-use existing RF equipment as much as possible
This list highlights how the demands on a radio network have evolved over the past years. The list also suggests that radio will continue to evolve at a great pace. Thus any next generation system must be flexible and adaptable enough to deal with the unknown demands of the future.
Developments in three areas of technology are driving developments in the area of satellite radio network distribution. So lets begin by looking at the major technological advances in (1) Audio Codecs and Compression, (2) Transmission Architecture including Modulation and Forward Error Correction as well as (3) IP Networking to understand their impact on the implementation of a new generation radio distribution system with respect to the desires of network engineers as expressed above.
(1) Advances in Audio Codecs and Compression
Audio encoding technologies have evolved over the past 15 years from MPEG Layer 2, Layer 2.5 and Layer 3 (MP3) into a wide variety of customized codecs from Real, Microsoft, Dolby and others, to open source codecs like Vorbis, to a new generation of MPEG standards based codecs know as Advanced Audio Codecs. The MPEG-4 specification now includes MPEG-4 AAC, MPEG-4 High Efficiency AAC (HE-AAC), MPEG-4 AAC Low Delay and HE-AAC v2. New variants are being developed and implemented to deal with speech systems requiring low delay and loss-less codecs are being implemented for those driven by quality.
Modern satellite distribution systems implement MPEG4-LC, MPEG4-LD and MPEG-4 HE-AAC for radio network applications. The figure below shows the evolution of MPEG4 AAC codecs.
One of the major issues with early generation satellite radio distribution systems was the pops and squeaks introduced into the audio channel by errors in the data caused primarily by rain fades and sun outages. In the MPEG-4 AAC family, significant error concealment and error masking techniques have been introduced which virtually eliminates these issues.
The benefit of all these codec improvements are that a good stereo radio service can be transmitted in the range of 80 to 96 kb/s, significantly reducing the space segment cost to distribute a radio channel. Audio transparency quality is generally agreed to occur at 128 kb/s for a stereo pair; audio purists may chose to operate with speeds higher than 128 kb/s.
As such, operators with APT-X and MPEG-2 systems operating at 256 kb/s can achieve transparency with their current capacity, a quality improvement and a savings of 50 percent when they operate a stereo pair at 128 kb/s using MPEG4. Conversely, stations can transmit 2 services in the capacity of one previous service doubling their capacity. Lastly, many codecs can support the common configurations of 5.1 and 7.1 leading the way to the future.
The last word on codecs is that users should choose products which incorporate software decoders and codecs. This allows for the implementation of future improvements to existing codecs or the implementation of new generation codecs over the lifetime of the equipment.
This analysis certainly seems to show that MPEG-4 AAC codecs perform a vital role in addressing the requirements of operators to decrease the cost per channel and improve audio quality while paving the way to the future.
As a recommendation, engineers should understand the benefits of the various new codecs and select a codec that meets their needs. The newest isnt always the best choice.
Make certain that the satellite receiver allows the choice of several codecs and that the receiver can be updated/changed over the satellite to allow the next generation codec to be implemented if necessary. As such, a software implementation of MPEG-4 HE-AAC probably fits most peoples needs.
(2) Advances in Transmission Architecture including Modulation and Forward Error Correction
In early generation SCPC radio distribution systems, proprietary BPSK and QPSK modulation and energy dispersal was used with convolutional 1/2 rate forward error correction encoding. These systems were great at the time and the low cost implementations contributed significantly to the success of early SCPC radio distribution systems.
However, DVB for satellite was standardized in 1994 and has subsequently been adopted for use in many satellite radio network distribution systems. The adoption of first generation DVB for satellite (DVB-S) contributed significantly to the improvement in performance of radio distribution systems as well as the virtual elimination of proprietary SCPC transmission systems for radio distribution.
To this day, DVB-S is used almost exclusively for radio distribution systems. The leverage provided by large DBS systems has created a wide variety of components and equipment that has been steadily being improved in performance and reduced in price to the benefit of all users.
Most companies have focused their development and product efforts around DVB-S to make it a great transmission system but now 13 years after it was standardized, even DVB-S is been overtaken by new technologies.
Over this 13 year period, significant advances have been made in the capabilities of analog and digital integrated circuit technologies which permit the implementation of circuits that could not previously be done cost effectively. These advances have lead to the release of the new DVB-S2 standard in 2005.
At the transmission layer, the DVB-S commonly used QPSK modulation has been supplanted in DVB-S2 with QPSK, 8PSK and 16QAM modulation.
Additionally, the traditional concatenated Reed-Solomon and convolutional FEC coding has been replaced by the more effective combination of BCH (Bose-Chaudhuri-Hocquengham) and LDPC (Low Density Parity Check). Furthermore, DVB-S2 supports better filtering with permits a narrower carrier as can be seen in the more aggressive carrier shaping factor which has been reduced to 0.2 from 0.35. The occupied bandwidth for a carrier of a given symbol rate can be calculated from the symbol rate. In DVB-S2, a 30 ms/s carrier occupies 36 MHz.
The combination of better FEC with less overhead and narrower carrier implementation has yielded an improvement in the data throughput of approximately 40 percent. The improvement in the coding gain by about 1.5dB allows for larger rain fade margins, or the use of smaller antennas or the use of 8PSK modulation into existing antennas.
DVB systems are inherently RF bandwidth agile and can contain a variable number of audio channels referred to as PIDs. Typically, the maximum RF symbol rate is 30ms/s which will fill up a 36 MHz transponder and the maximum number of PIDs is 8192, which allows for considerable expansion of the audio channel capacity.
How a new system benefits from these innovations in modulation, carrier shaping and FEC needs to be part of system design to determine the optimal choice for the system. This table shows the most common choices of modulation and the FEC rates available for DVB-S and DVB-S2. Be careful here, not all FEC rates are available in all products on the market.
The figure above shows two interesting items.
Firstly, the reduction in Eb/No between DVB-S and DVB-S2 implementations with QPSK at 1/2 rate FEC shows between 1.4 and 1.8 dB. Secondly, the figure shows the intersection of DVB-S at QPSK rate 3/4 with DVB-S2 at 8PSK at rate 2/3 implying that users could switch to 8PSK and realize a higher throughput
It can be concluded that DVB-S2 is a better transmission solution generally that offers many advantages over DVB-S and traditional SCPC systems. DVB-S however may be a better choice for some however given its maturity and the existing preponderance of existing DVB-S systems.
It seems clear however that DVB-S & DVB-S2 both address the desire for a flexible number of channels and flexible audio bandwidth in addition to contributing to a greater rain fade margin while re-using existing RF uplink and downlink equipment.
Advances in IP Networking
No discussion about future networks would be possible without including a discussion on the internet and computers to look at how IP networking and computer technology will impact the operation of the next generation radio distribution platform. We all know that the internet has developed over the past decade to the point that IP based audio systems play a big roll in radio today. Whether it is an IP codec that is used for contribution or remotes, linear IP networked audio inside studios, terrestrial or satellite IP distribution to transmitters or interfacing with our web based audio servers, 3G phone systems or DVB-H, IP is here to stay in the radio industry and offers many advantages.
Consequently, a new generation radio distribution system must be IP friendly and interconnect transparently with the existing IP infrastructure of a radio station at several levels. IP connectivity can be used for many cost saving applications.
- IP connection of remotes to studio
- IP networking inside the studio of linear audio
- IP networking of multicast audio over the satellite
- IP monitoring of remotes performance
- IP networking to provide backup in case of satellite link loss
- for rain/sun fades
- For damaged antennas & LNBs
- IP connectivity to web servers
The diagram below shows an all All Digital, IP Enabled, DVB-S/S2 radio distribution system similar to several being installed currently.
In this system, linear digital audio is delivered from the playout system to the Audio Encoder over IP over Ethernet. This eliminates the need to run balanced audio coaxial cable which can degrade the audio. This also eliminates the need to do an A/D conversion in the audio encoder which serves to protect the quality of the audio. This project uses MPEG-4 AAC-HE audio compression which is multicast over a terrestrial network to the satellite uplink where the IP packets are encapsulated inside DVB transport packets for transmission.
br /> The Network Management System (NMS) authorizes the receivers and forwards ad/commercial/program files to the HDDs of the receivers. The NMS system also schedules or triggers the ad/commercial/program insertions.
The DVB transmission in this case is transmitting the audio IP packets inside DVB transport packets at 128 ks/s. The carrier can be expanded up to a full transponder should the need arise to add more audio, video or data services.
On the downlink, Ku-band antennas equipped with LNBs are used to receive the signal which is cabled to the satellite receiver at the affiliate. The satellite receiver removed the DVB transport layer and reconstructs the IP multicast packets.
The IP packets are usually decoded to linear PCM audio and output over IP on Ethernet or converted to analog and output on balanced connectors. Additionally, the compressed audio can be output directly for use by systems requiring compressed audio for subsequent distribution as is the case with web servers, IPTV, 3G/3.5G phone, and DVB-H systems.
Internal to the satellite receiver is a large HDD which is used to store advertisements or programs which could be alternate language or even complete shows. New to this IP implementation is the ability for the satellite receiver to connect via the internet or other IP capability back to the hub and take the program through the terrestrial network as might be the case during a rain fade, sun outage or otherwise when the satellite link may not be available because of a mis-pointed antenna or defective LNB.
Remote sites are monitored via the terrestrial network using conventional SNMP/IP monitor and alarm system.
As such, the implementation of a flexible satellite architecture with the ability to implement multiple program services while reducing costs and improving quality has been shown to be completely capable of transmitting monaural, stereo and multichannel sound. The addition of the IP network layer and satellite receivers with internal HDDs can add the desired interconnectivity with internal station LANs as well as the internet. As such, local program, ad and commercial insertion is made possible as is connectivity to secondary IP enabled distribution systems such as web servers, DVB-H, IPTV and even 3/3.5G telephone systems.
It is evident that the implementation of modern MPEG-4 audio codecs & compression combined with the benefits of DVB-S and DVB-S2 go a long way to addressing the future requirements of station engineers for these new networks as identified earlier on in this article. Add into the mix, IP networking, and the resulting network solution will address all of the aforementioned requirements of station engineers.
About the author
Gary Carter is Vice President and Chief Technology Officer of International Datacasting Corporation and has over 25 years of experience in the satellite communications industry directly addressing the needs of broadcasters.