HD-ACC Scalable Audio Codec

Data compression is probably the single most important factor in the meteoric success of digital audio, especially when it comes to online downloads and portable players like the iPod. Lossy compression formats such as MP3 discard as much as 90 percent of the original data—hence the term "lossy"—so that music tracks can be quickly downloaded. In addition, such files require very little memory, allowing thousands of songs to be stored in a device no bigger than a matchbook.

On the other hand, there are many who bemoan the sacrifice of sound quality in favor of convenience, going so far as to claim that lossy compression removes much of the music's emotional impact. Also, lossless compression, which provides a perfect bit-for-bit representation of the original audio file, is critical for archiving high-quality recordings. In that case, of course, the decrease in file size is not nearly as dramatic—losslessly compressed audio files are about half the size of the original.

Fraunhofer Institute for Integrated Circuits, the organization that brought us MP3, has developed a solution that could bridge the gap between the convenience of lossy compression and the quality of lossless compression. Dubbed HD-AAC, the new scalable audio format combines two codecs from the Moving Picture Experts Group (MPEG), an international organization that also standardized MP3, more technically known as MPEG-1 Audio Level 3.

At the core of HD-AAC is MPEG-4 AAC (Advanced Audio Coding), a type of lossy compression that is more efficient—and therefore higher quality at a given bit rate—than MP3. To achieve lossless compression, HD-AAC adds an extension layer called MPEG-4 SLS (Scalable Lossless Coding—I've found no one who can explain why this is abbreviated "SLS" and not "SLC"). Together, these codecs compress any audio file losslessly.

The brilliance of HD-AAC is its scalability—some or all of the extension data can be ignored depending on the bandwidth and storage requirements of the playback device as depicted in the diagram above. If all the extension data is ignored, you're left with standard AAC, which most portable players can easily handle. However, if you have, say, a media-server system with lots of storage and bandwidth, you can enjoy the same file with higher quality.

Another advantage is backward compatibility. Older devices simply ignore the extension layer and play the AAC core, while newer players with an HD-AAC decoder can access more of the data for higher-quality playback.

HD-AAC can encode two-channel audio up to 192kHz/24-bit, but 96kHz/24-bit is more common and can also be used for multichannel surround mixes. Losslessly compressed, two-channel 192/24 audio requires a bit rate of about 3.5Mbps, while the AAC core is normally encoded at 128Kbps, illustrating the wide range of bit rates this codec can accommodate within a single file. Even better, an HD-AAC file can also include album art, liner notes, lyrics, and other metadata normally associated with physical media like CDs.

There are many possible applications for HD-AAC, from lossless archiving and audiophile listening to high-quality broadcast transmission at a near-lossless bit rate to playback on portable devices at lower bit rates. In all cases, the source is a single file with no need to re-encode for different applications. For example, you could rip your CD collection losslessly and listen to it on your home-entertainment system, then play the same files in lossy versions on your iPod. This is a very exciting development that could become increasingly important in our digital world.