Around the World

Making Music with McGill

By Philip Galanter

Convention attendees were brought by shuttle bus from the Javits Center to NYU's Cantor Film Center for the demo. A DVD-quality surround sound audio signal was streamed from a live performance by the McGill University Swing Band, playing in Montréal, to an audience in the 300-seat cinema. A live video feed of the swing band was simultaneously streamed and projected on the cinema's screen. In addition, Renata Celichowska, an NYU student in the School of Education Dance Education program, performed onstage to the music.

Dolby Digital Sound
The McGill Swing Band performance was streamed live in a 5.1 multi-channel surround-sound format. Microphones were placed to capture not only the direct sound, but also the ambience of the performance space. The Dolby Digital format reproduces this ambient space by using three speakers in the front (left, right and center) and two speakers in the rear (left and right), as well as a sub-woofer speaker system for low-frequency tones and effects.

The Dolby Digital format used for this demonstration is commonly used for major Hollywood motion pictures shown in commercial theaters. Dolby Digital immerses the audience in an aural environment where sounds can come from any direction.

The Internet2 Connection
Internet2 is the U.S. initiative to create next-generation Internet resources in support of robust digital media and other high bandwidth applications not currently possible on the original Internet. (For more information about Internet2, see Bill Russell's article.)

In order to connect NYU to McGill, this demonstration used not only Internet2 resources, but also the high-speed CA*net 3 network developed by CANARIE in Canada. The streaming sound and video data had to traverse a number of gateways, each of which had to be tested and debugged to prevent bottlenecks that might break up the real-time audio and video.

Software and Technical Details
The high-speed connectivity provided by Internet2 and CA*net 3 alone would accomplish very little. Virtually all new Internet2 services require the creation of application software to actually execute the intended function. The underlying software for these demos was developed at McGill University by a programming team led by McGill professor Jeremy Cooperstock.

At the performance space in Montréal, the six-channel mix of 48 kHz 16-bit surround sound was encoded using an off-the-shelf Dolby DP569 encoder unit at 640 kbps. The Dolby Digital stream was encapsulated within an industry standard AES/EBU stream at 1.5 Mbps. This stream was then input to a Silicon GraphicsIndy with a standard S/PDIF digital audio connector. The Indy, running software developed at McGill, encapsulated the AES/EBU stream within standard Internet UDP packets.

These packets then traversed the CA*net 3 and Internet2 networks to a similar SGI Indy at the NYU Cantor Center. McGill's software running on the NYU Indy then extracted the AES/EBU stream from the UDP packets and sent it out the S/PDIF digital audio connector for final decoding using a Dolby DP562 unit. Once the six-audio channels were reconstructed, they fed the Cantor Center theater-quality sound system.

Using a pair of independent Intel based Windows NT systems, video was transmitted in a distinct 1.5 Mbps stream using Cisco's IP/TV system with MPEG-1 compression.

As described by Professor Cooperstock, "When you combine multi-channel Internet audio transmission with high fidelity video, it creates an environment that replicates the same sorts of interactions that occur in the physical classroom."

The Demo Results
The event was not only a technology demonstration, but also an experiment in pushing the known limits of the network. The first of four demonstrations used a 23-second buffer against network congestion, while the other three used a far less conservative three-second buffer. No effort was made to limit unrelated network activity, other than disabling the incoming Usenet news feeds at both universities in the first three demonstrations.

During the final demo, the NYU news feed was switched back on, subjecting the program to intense competition for bandwidth. Nevertheless, at no time in any of the demonstrations was the Dolby Digital sound audibly interrupted. To further check for anomalies, digital recordings made at NYU of the streamed audio are being compared bit-for-bit with recordings made simultaneously at McGill.

Professor Robert Rowe, the Associate Director of the Music Technology Area in the School of Education, was NYU's lead faculty person in the collaboration. He noted, "The event was a compelling demonstration for us and the AES delegates that the time for high-quality surround sound over the Internet has arrived. As a consequence of this project, I have become vice-chair of the AES technical committee on Networked Audio, and we look forward to continuing this work at NYU."

Professor Wieslaw Woszczyk, Director of the McGill Graduate Program in Sound Recording and Chair of the AES Technical Council, remarked, "This technology opens the way for people in entertainment, business, education or research to collaborate live online. It will be much more appealing than the current tele-conferencing telephone model because it will offer an experience more like a movie theater. For collaborative musical performances and compositions over the Internet, it will be like a virtual classroom."

Posted October 9, 2000