High Fidelity: Brandenburg on audio compression
Professor Karlheinz Brandenburg talks about MP3, the audio coding standard that remains popular two decades he first introduced it to the world.
The enduring popularity of MPEG-1 Layer-3 (MP3) as an audio coding standard is closely tied to the development of digital media as a whole. Largely defined as an open standard, the widespread use of computers and increasing speed of Internet access further added to the ease of encoding and decoding MP3. Beyond its portable audio applications, MP3 has been invaluable to the growth of digital media in the areas of digital audio broadcasting, ISDN transmission for broadcast and distribution, archival storage for broadcasters, digital TV, Internet streaming and audio file storage/exchange.
Little has changed in the basic underlying technology since its introduction to the general public in the early 90s, says Professor Karlheinz Brandenburg, whose doctoral thesis on digital audio coding and perceptual measurement techniques formed the basis of the MP3 and most other modern audio compression schemes. “If you did some MP3 compression 15 years ago, all the decoders today are able to decode it. But audio compression technology has evolved and in fact there have been additions to MP3. Just one year after the MP3 compression format was introduced, lower sampling frequency was made available and subsequently, MP3 surround,” he tells Asia Image in a phone interview. “The underlying technology has not changed. In fact for AAC (advanced audio coding), the second innovation in audio coding standard, our team at the Fraunhofer Institute in Erlangen played a very important part in the development. Nowadays nearly every mobile phone can play MP3 and AAC. These formats are still state-of-the-art, giving listeners high audio quality and remain the top in the market for music on computers, streaming and all that.”
The MP3 format compresses digital audio data in way that uses the smallest file size possible to reconstruct (during decoding) audio to sound as close as possible to the original. The compressed file can be easily decoded by inexpensive software or hardware decoders (such as flash-memory based portable audio devices) and is flexible enough to be used in multiple applications. As a lossy compression technique, it has limitations when it comes to high quality audio coding. “MP3 at its time was the best audio coding system but still due to the standards process had little bits of compromises in there. For certain sounds, specifically percussion such as castanets, if you have very good headphones or a high-end stereo system you can detect the differences during playback, but these are minor. If you use high enough bitrates, MP3 is fine in terms of quality,” Brandenburg explains, adding, “While MP3 and AAC can give you very high quality, lossless formats are needed whenever you need to do some more processing like encoding and decoding.”
Even in capturing spatial audio it has limitations. “MP3 is a two-channel format, so it has the same limitations as the CD. If you have a good system and sit at the right spot it gives you very nice sound, but nowadays there are technologies which are able to capture spatial sound much better. I don't mean just surround sound. With MP3 surround you can do surround sound and AAC has multi-channel versions. I'm talking about systems built using the wave field synthesis, developed by Delft University in the Netherlands,” Brandenburg says. Wave field synthesis, he explains, uses a ring of loudspeakers around the room to give listeners much better spatial audio impression and immersion than any two-channel sound format such as CD or MP3.
Despite these limitations, MP3 is likely to remain one of the most popular audio compression formats for a while. “It depends on the available storage capacity and bitrates for streaming,” says Brandenburg. Twenty years from now things might change, but for now, as long as people see some advantage to have some compression around for this time there will be MP3 and I think it will be for a very long time,” he adds. “The age of digital media has just started.”
Professor Karlheinz Brandenburg is director of Fraunhofer Institute for Digital Media Technology. At SIGGRAPH Asia, held in Singapore from November 28 to December 1, he will talk about the story of the MP3 as well as new research aiming for the ultimate sound experience.
- Share this article
- Got more on this story? Email Asia Image