Beyond MP3 -- Multi-Channel Audio Home Theatre vs. Cinema


James D. Johnston


Chief Scientist, DTS, Inc.

Abstract:

MP3, MPEG-AAC, and other perceptual audio coders exploit the nature of the human auditory system in order to reduce the data rate necessary to transmit an audio signal. In an analogous fashion, a soundfield present in a real or imagined venue can be examined to see what parts of the enormous information (thousands of channels worth of data in 1 meter diameter sphere) present in a sound field are actually perceptible to the human listener.  Using this knowledge, it is possible to “overwrite” the sensation of being in a small room with that of being in a larger room (for instance, simulating an aircraft hangar inside of a home theatre space). It is not however generally possible to create the sensation of a smaller space inside of a larger space. In addition, the differential time delays and angular positions of loudspeakers in a larger (cinema) space vs. a smaller (home theatre) space create entirely different requirements for capture, production, and synthesis of a desired sensation.   This talk will discuss, at a conceptual level, what kinds of acoustic cues are present in a space, what kinds of cues are captured by the auditory system, and how one can use this knowledge to “transmit” a desired sensation of space and direction via multi-channel reproduction in the home theatre. Cinema production techniques will be briefly discussed, and the reason for their necessarily limited use of multi-channel audio effects explained.

Bio:

Jim Johnston (best known as JJ) is widely considered to be the "father" of perceptual audio coding, which is the basis for MP3.

 JJ was a Member of Technical Staff at AT&T Bell Laboratories from 1976 to 1996 and AT&T Research from 1996 to 2002. He is presently the Chief Scientist for DTS, Inc, working from
Kirkland, Washington.  DTS provides the high-quality audio system for BlueRay disc, and is an industry leader in audio coding, code preprocessing, multi-channel audio, loudness control, and other advanced signal processing algorithms for audio.

JJ is the primary inventor and architect for a variety of signal processing algorithms related to room correction, loudness processing, perceptual modeling of audio, audio coding, audio sound field perception and presentation, and standards and ancillary mathematics and science related to audio issues.  His current interests include loud speaker pattern analysis and control, loudness modeling, room simulation, stereo image control and analysis, filter design, speech coding, audio and speech testing methodology and execution, and implementation concerns in audio processing.


JJ is a Fellow of the Audio Engineering Society (1997) and a Fellow of the IEEE (2002).  In 2006, he was awarded the J. L. Flanagan Signal Processing Field Award from the IEEE Signal Processing Society for his work on creation and standardization of perceptual audio coding. He has also received an AT&T Technology Medal and an AT&T Standards Award (1998). In 2001, he received a New Jersey Inventor of the Year award for his contributions to MP3 and audio coding in general.


Monday, October 19 1:30-2:30 pm Room 1003 EECS

(Sponsor is Dave Neuhoff)