Applied Technology: MPEG Surround for HD Radio
Nov 1, 2005 12:00 PM, By Shawn Hopwood
Flash back to the early 1960s. All radio is mono and sounds like it is coming to you through string and tin cans. Suddenly a new technology arrives on the scene � FM, and it blows your mind. Not only does the music leap right out of the speakers, but it sounds as awesome as your hi-fi, thanks to stereo broadcasting. It took awhile to catch on, but today, 40 years later, no one would even think of broadcasting FM music in mono. Stereo is the obvious choice for broadcasters and for consumers.
Today, a new shift is just beginning. Stations are experimenting with a variety of ways to broadcast surround sound over the air. Although the shift from stereo to surround sound will take a few years to transpire, in much less than 40 years, the idea of stereo broadcasting for music will be just as foreign as mono is today. And just as FM was the driver for stereo, HD Radio will be the driver for surround sound.
Radio broadcasts are a perfect candidate for an upgrade to surround sound. With most radio listening happening in cars, and most of today's cars carrying at least six speakers, the rendering infrastructure is nearly in place. Additionally, the ongoing series of evolutionary changes in radio during the past couple of years, from analog FM to satellite radio and IBOC, has made radio aficionados receptive to another upgrade to their listening experience.
The bandwidth requirements of the various surround transmission systems. Click here to enlarge this image.
There are three main techniques that can be used to broadcast surround sound over the HD Radio system: discrete, matrix/watermark and MPEG Surround. A discrete system discretely encodes 5.1-channel programming on the IBOC signal, possibly using a 5.1 channel version of the already deployed HDC codec. This technique, similar to the Dolby Digital audio distributed on DVD soundtracks, is not feasible, however, because of the bandwidth limitations in HD Radio. Even with minimum quality, discrete 5.1 encoding would use up all the available bandwidth on the IBOC channel, eliminating the possibility of multiple programs or data services. At a minimum, discrete coding would require computationally expensive downmixing inside the radio for stereo-only environments, complicating implementation.
Matrix systems, such as SRS Circle Surround and Dolby Pro Logic, use matrix encoding to encapsulate a surround signal within a stereo signal. The L-R energy of the stereo signal is used to carry the surround sound information over standard stereo infrastructure. This allows matrix-encoded surround sound programming to be transmitted on the analog and digital portions of an HD Radio signal. When such a signal arrives at a radio with a matrix decoder, the stereo signal is upmixed for 5.1. Also, because the stereo signal is already the basis of transmission, the program can play back in a stereo-only environment with no modifications. The downside of matrix techniques is accuracy. Due to the inherent limit on the amount of information that can be transmitted using matrix techniques, the quality of the surround mix cannot match the original 5.1 channel mix. Also, by encoding the surround sound as part of the stereo waveform, it does not take full advantage of the digital information path offered by HD Radio.
MPEG Surround is a compromise between the discrete and the matrix methods. It provides the accuracy of discrete encoding while preserving the stereo compatibility and bandwidth efficiency of the matrix encoding. Borne out of a collaboration between Agere, Coding Technologies, Fraunhofer and Philips, MPEG Surround is a compression technique in the process of being standardized by the Moving Picture Experts Group. The goal of MPEG Surround is to provide a scalable, digital system to transmit high-quality surround sound in a way that is backward compatible to stereo. At the highest level, MPEG Surround adds digital information to a stereo mix, enabling reconstruction of an original 5.1 surround audio with little bandwidth overhead. The nature of the MPEG Surround technology allows it to scale in quality proportional to the amount of bandwidth allocated to the surround signal. Quality in this case is measured by the ability of the encoded signal to accurately reproduce the original 5.1 channel audio signal sent into the encoder. While the nominal operating bit rate for MPEG Surround on HD Radio would be around 6kb/s, MPEG Surround can scale all the way up to full transparency.
A multi-program HD Radio station would likely allocate 64kb/s out of the total 96kb/s to the main stereo program, then add 6kb/s MPEG Surround to achieve a full 5.1 surround sound broadcast. By tuning the encoder even further the stereo bit-rate could even be pulled back a bit thereby reducing the surround sound overhead to be almost negligible. MPEG Surround provides surround sound to a stereo broadcast with room for extra audio programs or other digital broadcast services.
But making the move to surround will not come easy. A major concern for broadcasters is studio and broadcast infrastructure. Regardless of the technique chosen � matrix, discrete or MPEG Surround � modifications will be required to enable the studio to take full advantage of surround sound. While the matrix method can be transmitted and manipulated using a stereo infrastructure, libraries, play-out systems, consoles and studio links may require upgrades.
For new surround facilities, there are two main techniques being considered to handle the transition. The first is shifting to a full IP audio infrastructure provided by a system like that from Axia. In an IP audio system, audio is transported over twisted-pair or fiber networks, and controllers can be configured and ganged on the fly. The second is a dedicated system that creates a parallel path for the MPEG Surround information by adding boxes that interface to the various stereo AES buses. Both methods can easily handle stereo or surround sound.
The Moving Picture Experts Group is in the process of defining and standardizing MPEG Surround with a final specification scheduled for mid-2006. Hardware implementations will be available shortly thereafter.
Hopwood is senior director of licensing for Coding Technologies.