by Steve Church & Michael DoschAxia AudioCleveland, Ohio, USA
There is growing interest among broadcasters to deliver a surround listening experience to their audiences. Surround is clearly the hot topic at audio, consumer electronics, and computer shops. Visit any of these and you will see plenty of surround audio set-ups. Indeed, it would appear to a casual visitor that stereo has become nearly obsolete. The systems on display are home systems, mostly to be used for “home theater” listening to accompany surround DVD-Video disks, which are near-universally produced with a digital surround audio track. But many people have discovered surround as an impressive way to enhance general music listening as well. Two audio disk formats offer an audiophile grade carrier to provide surround audio to consumers: DVD-Audio and Super Audio CD (SACD). In the shops, you will find DVD players costing around $140 that can play all three formats. In all cases, surround is being delivered to consumers digitally in the so-called 5.1 format, providing six discrete digital channels: Left Front, Right Front, Left Surround, Right Surround, Center, and Subwoofer.
Our current modern systems offer a tremendous jump in quality over the early "quadraphonic" attempts to woo consumers with multi-channel sound. These were “matrix” systems that combined four tracks into two using phase-shifting techniques. They had the advantage that existing stereo vinyl records and FM transmission could be used to convey the audio to consumers. They were even compatible! One record, one broadcast could serve both quad and stereo listeners, so they said. This was all certainly convenient to broadcasters: just playing a surround-encoded record in the same way and with the same equipment that you would play a stereo one made you a surround broadcaster.
Alas, compatibility was achieved only in the imaginings of the record company PR departments. The reality was quite something else, as the systems were heard and their faults became apparent: the vast majority of people still listening in stereo received music that sounded quite different from what they were used to hearing, with distinctively strange placement, reverb, and cancellation effects. And the separation of the surround channels proved disappointing, dipping to as low as 3dB between some of the channels. And so the great quad broadcasting experiment came to an end.
The modern "5.1 channels" idea was the first surround method conceived in the digital era. Work on it was begun in 1987 by the Society of Motion Picture and Television Engineers when it looked to be possible to digitally encode audio for film releases. SMPTE decided that 5.1 channels were satisfactory to create the aural sensations film producers desired. The name was proposed by film sound innovator Tomlinson Holman to initial confusion but eventual acceptance. The “point one” channel is the subwoofer, with the decimal value suggesting the limited frequency response of the channel.
Prior to the establishment of 5.1 as a standard, a few surround films had been released on 70mm that had the capability for six audio channels in the longstanding analog optical format. Star Wars was the first, followed by Close Encounters, Superman, and Apocalypse Now. These were all big hits and all used essentially the 5.1 L-C-R-LS-RS-Sub arrangement, so there had been successful real-world experience with the format. There is limited physical space on film for the marks needed to encode digital values and at first it seemed there would not be enough space to hold six channels. But a new development came along just in time: audio “coding” or compression. With the possibility to reduce the bitrate by a factor of up to 10 over simple PCM, multichannel digital for film became a practical reality. Dolby Digital, DTS, and Sony SDDS were invented to exploit the opportunity and remain in widespread use today.
What about radio? By now you have probably noticed what’s missing from this picture: Radio. While it always been possible to transmit matrix surround over traditional FM, the introduction of iBiquity HD Radio in the USA offers the opportunity to give listeners digital surround with quality commensurate with their current and soon-to-come experiences with movie theaters, DVD film, TV broadcast, surround music disks, computer audio, and portable players. European DAB offers a similar opportunity for surround upgrade, and first steps are underway to enhance this radio service.
The iBiquity HD Radio service has a bitrate of 100kbps, of which 96kbps is used for audio. Only a couple of years ago, this would have been thought to be too little for high-fidelity stereo. Surround at this rate was but a dream. As with the introduction of digital surround to film, enabled by the just-in-time availability of audio compression technology, another bit of magic seems to have appeared just when needed for surround radio broadcasting: parametric surround coding technology. This amazing development allows a stereo signal to be expanded to surround with an additional 5-32kbps added to the basic stereo rate. Recent tests of the latest surround codec version have shown that 6kbps offers performance as good as our first demonstrations that used 16kbps.
A station using all 96kbps for a single high quality program would probably take a 90/6kbps split for the core stereo and surround streams. Stations wanting to "multicast" will divide the bits according to the needs of the various programs.
iBiquity has a proposal before the FCC to increase the HD bitrate to 150kbps, which would offer more options for additional programs and/or quality enhancement.
MPEG Surround System via HD Radio end-to-end
Remember that the quad broadcasting experiments came to an end primarily because it was not possible to achieve acceptable compatibility in stereo. Fortunately, this is guaranteed not to be a problem with the modern MPEG approach because the stereo signal is taken directly from the original source and sent on to the listener without modification. Unlike with the matrix systems, there is no requirement for the original 5.1 source to be downmixed to create the stereo broadcast. Instead the system takes both the stereo and the 5.1 signals from the source, such as a DVD-A or SACD disk, and uses the 5.1 to create the 6kbps surround-coded stream. This stream contains the information necessary to expand the stereo to surround at the decoder; a format that its developers, Fraunhofer Labs, call "2 + 5.1".
So we have a fortunate match of the coder’s characteristics to the application. If for some reason we have only the 5.1 source, it would be possible to downmix it automatically to create the stereo signal, but if we have a handcrafted stereo mix available, as we usually do, we are able to use it. Advances in digital transmission and codec technology let us get that so often sought and so rarely achieved “something for nothing.” Station owners bought an old-fashioned analog stereo license and now find themselves with the potential to offer a state-of-the-art digital surround service just when they need it to compete. Not bad!
Building a modern surround radio studio facility To broadcast in MPEG surround, a station has to upgrade it's studio facilities to surround. Specifically, we need to store, network, and mix in the 2 + 5.1 format. This is the main objection that proponents of matrix systems proffer – that their systems can be used with existing stereo facilities.
So, since we need to examine the cost associated with a surround upgrade, let’s walk through how one would build a modern 2 + 5.1 plant, with a careful eye to expenses. We'll start with the routing and distribution infrastructure, then the PC-based delivery system, move to the mixing console, then on to the surround encoding, dynamics processing, STL, and transmitter. Then we’ll discuss monitoring and the production studio’s needs. All will be in the context of using computers and computer networking to provide the functions we used to get from the old proprietary radio station machinery.
Surround Radio Station: Functional Perspective
The Routing and Distribution Infrastructure Here we are talking about the glue that binds the studios together. In smaller stations, this may only be a couple of distribution amps that provide a couple of rooms with network feeds. But larger facilities, including the common consolidated ones in the USA, usually need to have flexible audio routing so that sources originating at any point within the plant may be consumed anywhere else. There is also the need to distribute a multitude of network feeds to a number of studios, switch various studios to transmitters and other outgoing lines, etc. Thus in the past years, we have seen the increasing application of facility-wide audio routers such as have been common in TV facilities for some time. These are proprietary boxes filled with cards that communicate via a backplane and offer various kinds of input/output. They look very much like the telephone PBXs that have been in use over the past decades and share many characteristics. These are manufactured in low volume for our very small industry, and are consequently expensive. Each input or output requires a port on a card which needs physical space, conversion chips, etc. An 8-channel input such as we need for surround would require 8 individual XLRs for analog or 4 for AES3 connections. Same for a surround output.
We propose a system that uses an Ethernet switch as an audio router. When analog or AES3 inputs and outputs are needed, these are converted in “nodes” to Ethernet. But this system requires many fewer of these because most devices communicate directly via a single Ethernet RJ-45.
An Ethernet 100BaseT link has 100Mbps capacity, enough to transport 25 uncompressed stereo signals or 3 8-channel surround signals. And these are bi-directional. One RJ-45 thus substitutes for as many as 100 XLRs!
Delivery System Most stations are using PC-based delivery systems to play music, promos, commercials, etc. With today’s low-cost, high-capacity hard drives, there is no significant barrier to storing the required 8 channels. A 300 Gigabyte drive costs less than $200 and can store 1200 surround songs with no compression.
With an Ethernet infrastructure, there is no need for soundcards and their associated connectors. There is also no need for router or console inputs and/or outputs at the other end. Driver software passes the audio to and from the audio playback application and the Ethernet. Physical connection is via a single RJ-45. Modern Ethernet switches support “Quality of Service” prioritization, so that general data may share the same link as audio. That means that you can use the same network for both audio playback and for other applications like file downloads from a server.
We propose to store audio in eight 24-bit integer packed PCM (uncompressed) channels in standard Windows interleaved wav format, organized as shown in the table at left.
This layout is standardized within the ITU and SMPTE for interchange of program content accompanying a picture and is widely used with TV digital tape recorders. The Music Producer’s Guild of America has also endorsed it. For Windows PCs, this will be stored in the RIFF/WAVE audio file format, which is a variation of the longstanding .wav format. The “fmt” (format description) chunk is a WAVEFORMATEXTENSIBLE structure that allows description of multichannel formats as well as any other PCM and non-PCM audio formats.
We choose 24-bit because DVD-Audio has this resolution and SACD has dynamic range that could take advantage of this bit depth. The Axia Livewire network also has 24-bit resolution, so we have a match between the source, the storage, and the network. Compact disks have 16 bits and this has been the norm in broadcasting, but 24-bits are the future. 16-bit systems have theoretically 94dB dynamic range, and 24-bit systems 141dB. Both are plenty enough for radio broadcasting, but having more bits means that distortion at low audio levels is reduced, which may be audible – even (or particularly?) after aggressive processing. Were big cheap hard drives not available, we’d probably want to stay with 16 bits – but with drives so cheap, why not splurge?
Audio that is stored on other formats: compressed, fewer bits, only mono, stereo, or surround should be uncompressed and/or up or down-mixed as needed to convert to the network’s standard format. The file header tells the application about the format so that it knows what to do. We could consider compressing the surround channels to extend capacity. Since they are only used as inputs to the surround position encoder and are not actually transmitted, there would be no degradation of the on-air quality at all. On the other hand, swapping to a bigger drive or adding another one is so cheap, so perhaps there is no compelling reason to bother with it. A software “driver” installed in the PC makes the network look like a standard Windows Driver Model (WDM) soundcard, so any audio application that works with usual soundcards should work without modification to send and receive audio from the network.
Mixing Console A modern mixing console can be built with two ingredients: A control surface and a mixing and processing Engine with PC motherboard, CPU, and Ethernet connection. While the control surface has to be manufactured in the small volumes our industry dictates, the Engine can take advantage of powerful, high-volume, low-cost components from the computer world. A commodity 2.4 Gigahertz Pentium 4 CPU has plenty of horse-power to support mixing, equalization, panning, dynamics control, etc. for a 24-fader surround broadcast console.
Next-generation consoles will have a surround metering option that show both stereo and 5.1 channel levels (as demonstrated by Axia's Element modular control surface, above). Because this is only a software change for a PC display, there is no additional cost compared to stereo.
The Engine has only two connectors: power and Gigabit Ethernet. All audio and control pass via the single RJ-45 Ethernet jack. With no hard drive (software is stored in a Compact Flash card), embedded Linux as the operating system, and all parts mounted on one PCB, reliability is probably higher than a traditional digital mixing engine with its many plug-in DSP, CPU, input/output cards, etc.
Cost to provide surround mixing in this PC Engine-based console is the same as for stereo. There is no incremental increase in cost going to surround from stereo because the P4 platform has so much headroom that surround mixing software can be added without changing any hardware. The Gigabit Ethernet connection has enough capacity as well to support the additional surround signals. Contrast this with a surround upgrade to a traditional console. You would have four times the dozens of audio in/out connectors already needed for stereo and many more plug-in cards, leading to probably having to increase the size of the frame. Your cost increment would be tens of thousands of dollars.
Via the Ethernet switch, the console has access to any audio source in the system. Its various outputs may consumed anywhere within the facility.
There will surely be a lot of experimentation with microphone ideas for surround. In most cases, mono mics will be panned to a position within the surround stage. Perhaps reverb with multiple outputs, time-delay, pitch-shifting, or comb-filtering processing will be used to create a sense of immersive spaciousness. Surround panning will be part of the console and so microphones without additional processing will cost no more to support in surround than in stereo. Other local inputs and outputs would require corresponding ports in the audio-to-Ethernet nodes. Stereo CD players would need only the usual two input ports and would be panned to surround within the console. Only surround SACD and DVD players would require surround inputs. They would normally connect 5.1 channels with the downmixing to stereo happening within the console.
Surround Encoding The surround encoder can another Ethernet-connected box. One RJ-45 serves all required inputs and outputs. The 2 + 5.1 channels from the console program output are the inputs and the output is a 6kbps coded surround stream that gets sent to the transmitter. Alternatively, the surround encoding could be done within the dynamics processor or HD encoder.
Dynamics Processing Initial testing indicates that existing stereo processing is satisfactory for the MPEG surround system. Any dynamics processing that is applied to the stereo channel affects the received surround channels as well. Because today’s processing is not enabled for direct Ethernet connection, a node is used adapt the network audio to the processor’s input and to apply the output back to the network.
Future processors may incorporate the surround encoder and offer more sophisticated individual processing control over the stereo and surround channels. For example, it may be interesting to have a way to “deprocess” the surround channels somewhat, while maintaining a more aggressive sound on the stereo program. More than a few listeners exposed to surround have said that the envelopment effect causes a perception of "high energy" similar to what programmers and engineers try to achieve with dynamics compression.
FM and HD require different processing styles. FM needs special attention to pre-emphasis, usually quite a lot of left/right clipping, and perhaps even some composite clipping. The HD encoder has to work harder on a clipped signal and will not have as good a result as from a non-clipped signal since the additional harmonics look like audio that needs to be encoded and therefore attempt to receive bit allocation. So a processor optimized for the HD channel will generally use a look-ahead limiter rather than a clipper. Stations probably will decide to process the HD program less than the FM in order to offer a more “purist” signal to listeners. (For now, anyway. When HD Radio gets popular, all bets are off.) The most popular processors use a common front-end AGC section and follow that with independent limiter sections for FM and HD. Thus, both outputs need to be connected ultimately to their respective transmitters.
HD Radio Encoding In iBiquity’s second-generation HD Radio system, the encoder is located at the studio. This has the advantage that the processing may be co-located at the studio and the STL only has to convey the encoded HD signal, tremendously reducing its bandwidth requirements. It also gives the benefit that any additional data that needs to be muxed-in can be applied at the studio. This data could be Program Associated Data (PAD) like song titles, or indeed our 6kbps coded surround stream. The input for this data is via Ethernet, so connecting it to the network easily enables a path from the surround encoder.
Studio to Transmitter Link In most cases, we need to get our FM program audio to the transmitter in either composite stereo or PCM form. And we need to send the 96kbps HD radio signal. We could decide to do this with two independent links, or we could use one STL radio to handle both.
A digital STL such as the Mosley Starlink can be used in this set-up. The FM audio goes via the usual input and the HD radio signal via the ancillary data channel. These radios don’t have much capacity because they operate in the traditional 950 MHz band, where not much bandwidth is available. Because their operating frequencies are protected by license and because the frequencies they use are (relatively) low, they are quite reliable.
Another way would be to use the new Ethernet radios like the BE Big Pipe. These operate with bitrates up to 45Mbps, so there is a lot of capacity for multiple audio channels as well as data, VoIP phone, etc. Since we already have all our facility’s audio on the Ethernet, no format conversion is required – just connect the radio’s Ethernet jack to a port on the Ethernet switch. These operate in the unlicensed ISM band at 5.2 and 5.7 GHz, so there is some risk. However, the few current users report good performance and overall satisfaction.
At the Transmitter With the HD encoding at the studio, there is not much to be done at the transmitter site. The HD exciter simply accepts the already encoded and multiplexed bitstream from the STL and modulates it for transmission. The FM audio is applied to the FM exciter and transmitted as usual.
Monitoring It’s going to be necessary to listen to your internal audio and your station’s on-air program in surround. This means 5 small speakers and one subwoofer in each serious monitoring position.
The old quad arrangement, with the speakers in each corner of the room, is not the right way to position your monitoring set-up. Human ears are not front-to-back symmetrical and that set-up not only sounds unnatural, but may indeed provoke stress as your deep genetic wiring causes your brain to tell you that “there is danger behind.”
The right way is defined in ITU standard 775 (illustrated at left). This specifies the left and right front speakers to be placed at 30° from the listening position. The surrounds go at 110° ±10° - just a bit back of straight out to the sides. The center goes in the center and the sub goes wherever it sounds the best or is out of the way. You should not be able to detect the position of the subwoofer. According to well-researched psycho-acoustics, humans are not able to localize frequencies below 80Hz. Our heads are too small and our ears too close together at these long wavelengths to detect any left-right difference. If you are able to locate the sub, it probably means that it is radiating audio at a frequency high enough to be localized. One cause of this is distortion-caused harmonics outside of the sub’s proper operating range.
Another psychoacoustic phenomenon to be aware of is the human ear’s change in frequency response from different positions due to the Head Related Transfer Response (HRTF). Sound entering the ear from the side speakers will be perceived as bright compared to sound panned to the front speakers. The effect is significant – a broad curve starting at 1.6kHz, reaching an 8dB peak at 4kHz, and extending to 7kHz. Music producers have probably already compensated for this in their mixes, so it’s not an issue for normal listening. But if you are checking your set-up with white noise, you will likely notice this.
While the usual set-up calls for a one-to-one correspondence between channels and speakers, when you have small main speakers, you will probably need “bass management” to filter the low frequencies from the small speakers and re-direct them to the subwoofer. This means that the subwoofer will be responsible for the sum of the “.1” bass channel and the filtered lows from each of the other channels. This is how the “theater in a box” systems so popular with consumers do it.
As another compromise, you could leave off the center speaker and add the center signal to both the left and right front speakers. (I actually recommend this for your home music listening system. While the center speaker is helpful to stabilize the dialog that accompanies video, it has been my experience that music is better without it. You have the practical consideration that the center speaker is probably not nearly as good as your two front mains – and it can’t be if you have a TV in front of you since the screen and the speaker can’t share the same space. Movie theaters solve this by putting the speakers behind a screen with holes in it. Punching holes in your TV’s CRT is not very likely to be a satisfying operation…)
Production Studio Most PC-based audio editors such as Adobe Audition, ProTools, etc. support mixing for surround, a procedure not much more complicated than stereo mixing. For a production studio now equipped with a PC editor (are there many that aren’t these days?), a soundcard upgrade and a surround monitoring loudspeaker set-up may be all that are required to start producing in surround.
Adobe Audition's surround mix-down function.
For dubbing surround music from disks to the delivery system, the production studio will need a DVDAudio and SACD player, which may be one universal device. These players will not output stereo and 5.1 simultaneously, so the tracks need to be recorded separately and synchronized in an audio editor.
Data If the audio network is engineered with sufficient capacity and it correctly supports modern priority mechanisms, it could also be used for the station’s data needs. Email, web browsing, client-server downloads, etc. may traverse the common network. The Ethernet switch isolates the data traffic from the audio streams. When audio and data need to share the same switch port and link, the audio is assured to have first call on the bandwidth because it has higher priority than the data and the switch knows to hold any data packets until the audio is sent. The TCP (Transmission Control Protocol) part of TCP/IP in the network interfaces of computers automatically regulates the data transmission rate to fill the link capacity not occupied by audio.
A more conservative approach is to have two networks. Computers that need to have access to both could have two Ethernet cards with a connection from one to the audio and the other to the data network. Or an IP router could be used to safely pass data from one network to the other.
Cost Comparison - Conventional vs. Networked Studios Since our focus has been on the real-world practicality of upgrading a facility to surround, let’s explore the cost to build discrete 5.1 surround-capable studios with different approaches. For this discussion, we will have a basic studio set-up with a 12 fader console, 3 mic inputs, 4 automation source inputs, 2 inputs for codecs and 2 more for phone hybrids, and 1 input for an SACD player. We will compare the cost of building a stereo studio with that of a surround studio.
Total Cost: $23,495
Compare the cost of networked studios using Axia to the cost of a conventional studio build:
Stereo $40k - $60k
The Axia networked console approach is much less expensive than the conventional (digital router-based console) systems for stereo studios. The small increment in cost for the networked system when expanding from stereo to surround-capable is due to the inherent characteristics of the networked approach. The networked radio studio has only a single cable for each source, whether stereo or surround. The DSP mixing engine is the same as for stereo. Soundcards are replaced by a software driver and the router is replaced by a low-cost Ethernet switch, which handles stereo or surround equally well. With conventional systems you must increase the quantity of PC soundcard ports, console input cards, output cards, cables, DSP cards, frame size, etc. by a factor of four. These costs can vary significantly between vendors but as a rule of thumb, a networked stereo console should cost about half that of its conventional counterpart. And the upgrade to surround will be only a small incremental cost for speakers, extra hard drive space and some additional I/O.
Axia's Element modular control surface (left) and Studio Engine mixing engine (above) can be deployed to mix stereo audio and software-upgraded for discrete surround audio mixing duty.
Cost Comparison - Matrixed vs. Discrete Encoding Proponents of matrix have been hammering on the point that a studio upgrade to support surround is expensive and therefore impractical. But consider... to go the matrix route and stay with your stereo facility, for a typical station you would need:
This is 2-3 encoders, 3 decoders, and a mixer of some kind. Matrix audio encoders cost north of $5k, so you'd have $25-30k in these boxes, and another, say, $4k for some kind of surround mixer/panner. (Does anyone know of a product that does this without going to a full-blown console?)
On the other hand, a complete Axia surround-ready studio eliminates the need for this expensive extra gear. You have integrated surround panning, mixing, metering, routing, etc., so you don't need an awkward sub-mixer. You can pop a surround SACD or DVD directly on the air without pre-encoding it to matrix. (You could do this with the matrix studio approach, but you'd need yet another encoder.) You can monitor surround or stereo anywhere you can plug-in - no decoder needed. Sure, you will need a bigger hard drive in your delivery system to store your audio in discrete form, but 300 Gigabytes would suffice for most commercial stations, providing something like a 1200-music-piece capacity for less than $200. And you would need one MPEG surround encoder as part of the transmission chain. A stand-alone encoder might cost, say, $5k. But one included in the HD generator or dynamics processor would likely cost less.
And if you need to upgrade your studio anyway, an Axia networked solution gives you surround mixing and routing at no increase in cost over stereo. This becomes a lot *cheaper* than the matrix approach because you don't need all the encoders and decoders everywhere.
To summarize cost (not including the surround loudspeakers, SACD/DVD players, and maybe an upgrade to your production audio editor that you'd need either way):
The big cost benefit claimed by the matrix guys is just not there. Indeed, there is a significant *cost penalty* for those who are starting fresh.
By the way, the two surround approaches are not mutually exclusive. You could (and probably should) go with a discrete studio even if you plan to broadcast in matrix surround. You'd have only one matrix encoder, just after the studio output and before the on-air dynamics processor. You'd get convenient panning, metering, and monitoring with no outboard lash-ups - and save a lot on encoders and decoders.
By the same token, you could (but probably shouldn't) use a matrix system in-house to feed an MPEG on-air system. This could offer those who are adamant they want to stick with their stereo storage and consoles a way to do so without compromising broadcast quality for the rest of us.
Time for a Radio Revolution? While the world around swirls with change and opportunity, not much has happened to the technology side of radio since the addition of stereo to FM in the early 60s. Other established media (film, TV, music disks) have exploded with innovation, and completely new media (Internet, iPod, satellite broadcasting) have burst onto the scene. All of the preceding have gained capability and appeal from having transitioned to digital. We finally have HD to take our industry into the digital era, but in stereo it offers a very small improvement to the FM listener’s experience. When we were growing up, FM was cool because it was at the pinnacle of audio delivery technology. Just the letters FM connoted a general sense of quality. With competitive media having surpassed radio, this connotation has faded. Without action, we will surely be sidelined.
Digital surround may well be an answer. It’s a way to please older music fans by making the classics fresh and to excite younger listeners with aural fireworks – all the better in cars with huge subwoofers. As film-sound innovator Tomlinson Holman says: “Perceptually, we know that everyone equipped Any PC on the network can listen to surround audio streams. with normal hearing can hear the difference between mono and stereo, and it is a large difference. And…virtually everyone can hear the difference between 2-channel stereo and 5.1 channel sound as a significant improvement.”
There is an argument that the HD channel capacity should be split and used to "multicast" two or more programs. If we divide the current 96kbps channel into two 48kbps channels, we could certainly offer two good-fidelity talk services. (One of them would not benefit from HD Radio’s “revert to analog” feature in the case of digital failure, though.) There is no need for surround in this scenario. But 48kbps is probably not good enough for music. So we could imagine that a station operator could decide to use his available bandwidth for either two channels of talk or one channel of music. Should iBiquity’s proposal to increase HD’s bandwidth succeed, a division of 90/6 + 48 could nicely serve one high-quality surround music service and one talk service. Other splits would be possible, such as 64/6 + 64/6 for two surround services.
Either way, compared to a mere shift to stereo digital, there will be a clear motivation for a consumer to buy an HD-enabled receiver. And a state-of-the-art networked studio facility supporting the creation of on-air product for these services presents an opportunity for both cost savings and operational flexibility.