A Bit About Audio Processing

NEW YORK

Just as a new generation of video coders is finding its way into the market, audio is keeping up the pace. Audio coders handle both ends of the bit-rate spectrum (and in the middle as well). Low bit-rate coders for 5.1 surround sound are a good match for the higher efficiency video coders. At the other end, higher bit-rate audio coders handle the ever increasing number of channels required, including multiple-language 5.1 channels along with description channels.

LOW BIT-RATE CODERS

Increasing efficiency in the emission path is driving the core market for low bit-rate coders.

"HD video uses considerable bandwidth," even with advanced coding techniques such as H.246 and VC-1, said Ted Laverty, director of business development, broadcast for DTS, an Agoura Hills, Calif.-based developer of audio coding technology. "If a broadcaster can reduce the requirement for carriage of audio at the transport stream level without compromise in the quality and ease-of-use at the consumer end, then that leaves extra capacity. This can be used to either boost video quality, add extra audio services [extra languages, descriptive narrative etc.] or even introduce more channels on a given transponder due to cumulative data-rate savings."

aacPlus, a low bit-rate perceptual coder from Coding Technologies of Nuernberg, Germany, can produce 5.1 audio channels at rates as low as 128 kbps, and combines three technologies: advanced audio coding (aac), spectral band replication (SBR) and parametric stereo (PS). According to the company, these three technologies are part of the ISO/IEC standard for MPEG-4 and are also standardized by the European Telecommunications Standards Institute, for coding for DVB services over IP and for broadcast applications.

Applications include satellite transmission, cable systems, and DVB, systems that aren't bound by ATSC standards, according to Tim Carroll, founder and president of Linear Acoustic, a developer of audio processing technologies in Morris Plains, N.J. Linear Acoustic incorporates aacPlus coding in its Aeromax 5.1 audio and metadata encoder for transmission applications. "I didn't believe the quality of the coder could be this good at this low bit-rate, but it does sound excellent," Carroll said.

For Dolby, its enhanced AC-3 (E-AC-3 or Dolby Digital Plus), is an extension of Dolby Digital (AC-3) coding for multichannel, stereo or mono. Dolby Digital Plus expands the range of data rates down to 32 kbps and up to 6.144 Mbps, compared with AC-3 range of 56 to 640 kbps, with ATSC specifying either 384 or 448 kbps.

At the lower end of the bit range, "Dolby Digital provides for lower data rates for efficiencies when the bandwidth is strained, like for satellite and next generation IPTV," said Rocky Graham, director of broadcast products at Dolby. For this application, Dolby Digital Plus can code 5.1 channels in around 200 kbps. Enhanced AC-3 is part of the ATSC standards for the robust data channel, Graham said.

Like AC-3, E-AC-3 is a perceptual coder and according to Dolby, the improved coding efficiency is the result of improved filter bank and quantization, enhanced channel coupling, spectral extension, and transient prenoise processing.

As for applications, Graham said, "don't expect to see MPEG-4 in cable systems right away, but you will see it sooner in satellite systems and IPTV. Satellite will deliver many hundreds of channels and needs to deal with reducing data rates significantly."

THE CONSUMER SIDE

Using low bit-rate highly efficient coders does indeed save precious bandwidth at the transmission and distribution end, but what about the home consumer? A typical home theatre receiver (at least those available today) can't decode E-AC-3 or aacPlus.

So the output from the low bit-rate coders needs to be converted back into whatever format the A/V receiver supports. Two of these formats (there are others) are Dolby Digital, and DTS Coherent Acoustics.

Dolby has developed a transcoding scheme that converts E-AC-3 to AC-3 without going to baseband PCM. This is because "with Dolby Digital and Dolby Digital Plus, you share features and common building blocks," Graham said, and avoid needing a decode/re-encode cycle.

To convert aacPlus to something a consumer receiver recognizes, Coding Technologies partnered with DTS to create what the companies refer to as a transcoder to convert the aacPlus signal to a 1.5 Mbps DTS Coherent Acoustics signal in the consumer set-top box. The conversion involves decoding the aacPlus signal to baseband PCM then re-encoding it to DTS Coherent Acoustics.

"DTS has never had consumer applications allowing any third party codec to be fed into a DTS-enabled home theatre system until now," Laverty said. "We also have worked closely together to create proof-of-concepts, demos and give technical support to silicon implementors."

Whatever the process, the end product is the chips that will find their way into the next generation of set-top boxes. For Dolby, Graham said that Conexant, Broadcom and MIPS currently have chips available.

"STMicro [STMicroelectronics] is currently undergoing certification and [chips] will be available in the first quarter of 2007," Graham said. "There are additional ICs that will be available soon but have not yet been announced."

For aacPlus/Coherent Acoustics, chip makers Broadcom and STMicro have made public announcements on DTS. Others are in the pipeline, Laverty says, but haven't gone public as yet.

"We have considerable interest from some significant U.S. broadcasters," Laverty said. "We believe we have a technology application that meets the needs and aspirations of broadcasters and consumers with the widespread roll-out of HD video. Time will tell how it works out, but the commitment that each company is giving to this program is an indication on how big a deal we believe it to be."

At the higher end of the bit-rate spectrum, the Dolby Digital Plus "technology itself can go up to 6 Mbps and beyond, but HD DVD constrains the maximum to 3 Mbps, and Blu-ray allows a maximum of 1.7 Mbps," Graham said. With these bit rates, more channels can be encoded.

Dolby Digital Plus also supports stream mixing, where the audio channels on a DVD can be mixed with those from an Internet stream. "The disc player would be connected to a broadband connection and be able to receive additional updated content, for example, mixed in with the audio from the disc," Graham said. "It's not clear yet exactly how content creators will want to use this capability, but we support it."

Dolby Digital Plus debuted in the Dolby Media Producer for authoring packaged media formats and early tools were available to Hollywood studios for months before that, according to Graham.

Linear Acoustic is incorporating DTS Coherent Acoustics in its StreamStacker-HD series of products along with its own technology to create a new mezzanine format called "e2" (e-squared) which carries up to 16 channels and metadata in a 2 Mbps AES-compatible stream.

e-squared relies on the ADPCM (adaptive differential or delta pulse code modulation) nature of Coherent Acoustics that Carroll said prevents clicks, pops or noise bursts when there's a bit error. The e2 signal is locked to an audio reference rather than a video frame rate and is editable.

"There's about one NTSC frame of delay through the box, and it won't change with different video frame rates," Carroll said.

e-squared supports advanced audio metadata and full Dolby-E metadata, according to Carroll. Incoming metadata can be received via RS485 or through the VANC (vertical ancillary data) embedded in an HD signal.

"We need to send more channels than anyone ever expected," Carroll said. "As we began looking at the rest of the world, we found that countries like China and those in Europe need multilanguage. It's more important than in the United States. There it's a fact of life and mandated by law, so we looked at how to make the change."