New audio coding system solves audio latency problem

May 27, 2011

A new codec (data coding system) called Enhanced Low Delay Advanced Audio Coding (AAC-ELD) developed by researchers at the Fraunhofer Institute for Integrated Circuits IIS has solved a major problem with Skype and other videoconference calls: latency (annoying sound delay between participants).

Their new Enhanced Low Delay Advanced Audio Coding (AAC-ELD) allows for latency only about 15 milliseconds. In the process, they also managed to reduce the audio data to less than one-thirtieth of its original size without major loss of sound quality, thus reducing the required bandwidth.

The researchers developed an algorithm that requires a certain amount of time to encode data and then decode it again at the other end of the line. The process requires data that is still in the future, as it must wait for the data to arrive. They optimized the algorithm to shorten the delay (and not impair the quality at the same time) by minimizing the area that is forward-looking and only process current data until there is an optimum balance between quality and delay.

To test the new codec, they developed an app to play games across the borders of cities or countries. “Thanks to the optimized image and sound quality, there is the impression that game partners who are far apart from each other are not in front of screens, but actually sitting across from one another,” said Manfred Lutzky.

Fraunhofer IIS is known as the main inventor of mp3, the audio codec that made it possible to greatly reduce the size of music or other audio files without impairing the sound. AAC Low Delay, the forerunner of AAC-ELD, is the codec used for many video-conferencing systems and in some radio broadcasting links.