Peter Chu has spent many years behind the scenes at Polycom working to create and fine tune our innovative, signature audio technology in our Media Lab. Learn how this audio guru got his start, and get Peter’s take on the industry, including what’s hot now and what we can expect to see—or hear!—in the future.
What was your career path to Polycom?
My path to Polycom began when I started working on echo cancellation for video conferencing audio at PictureTel in 1989. At that time, usable echo cancellation didn’t exist—audio echo and cut-out speech were big problems in video conferencing. The PictureTel echo canceller was the first commercial implementation that performed echo cancellation individually on many separate spectral bands as opposed to all frequencies at once. This sub-band approach gave very good performance (trademarked as “IDEC” – Integrated Dynamic Echo Cancellation) and helped PictureTel gain market share from competitors. The echo canceller demonstrated that a high quality speakerphone could be built.
In 1990, PictureTel Co-Founder Brian Hinman co-founded Polycom with Jeff Rodman in order to build speakerphones. I eventually joined Polycom in 1999. Polycom funded a lab to facilitate my team’s development efforts. Over the years, many innovations have arisen from this lab (stereo audio echo cancellation, Eagle Eye Director, G.719, Siren LPR, Live Music Mode, ceiling microphone array, ultrasonic pairing, noise blocker, voice-activated AGC, etc.) and never-ending refinements to improve audio quality.
What are some interesting examples of your audio/sound experiences? The work requires that I do lots of experiments with loud audio in my office. I can’t use headphones because I’m testing loudspeaker audio. Neighboring colleagues would go crazy, hearing the phrase “Testing 1, 2, 3” over and over again. In consideration of neighbors, my current office is acoustically isolated.
What is a hot/trending topic in audio/sound today, and how do you see Polycom playing a role to help facilitate change? If you can’t see the visual details of a person, video conferencing has no benefit over a phone call. In a group video conference, presenting visual details of the talker to the far-end is a problem. For example, the motivation behind the design of huge telepresence systems with 80-inch screens is to show visual details of the speaker. A much cheaper alternative is to use sound source localization and video face detection to automatically zoom the camera in on the speaker. Polycom’s Eagle Eye Director is the first commercial system to successfully accomplish this. As Polycom continues to develop more automated camera control features and lower costs, automated camera control will become a necessary requirement, not optional. This is a future trend where Polycom is the industry leader.
How do you see audio/sound technology working in the future? (E.g. in the year 2020)
By the year 2020, I hope that the “CD” audio quality we experience in Polycom video conferencing makes its way to our everyday cell phones. The poor quality of cell phone audio is very irritating.
What’s the most unique situation where you had to use Polycom technology? Over several months I worked with Christianne Orto of Manhattan School of Music to develop “Live Music Mode” for teaching music at a distance. Music Mode is a Polycom feature that more faithfully reproduces live music picked up by microphones so remote listeners can hear the true sound of instruments and voices produced by remote participants
It was quite an experience to get to know musicians who care more passionately about audio quality than me.
What do you like to do in your free time?
I like endurance sports due to the mental and physical benefits. I regularly do long distance running, rowing, and ride the “Elliptigo” (outdoor elliptical bicycle).
This is 1 of 2 blog entries we're publishing timed to the National Day of Listening (Nov. 29), which encourages people to sit down and record a meaningful conversation. To read the other 'Meet the Expert' blog entry with Jeff Rodman, please click here.