PNW Chair Dan Mortensen opened the meeting with Section news, then had all attendees briefly introduce themselves.

Dana Troxel is a Washington state native who received his degree in Electrical and Computer Engineering from Brigham Young University in 1990. He worked for 5 years at Active Noise and Vibrations Technologies, 3 years at Telex/EV, and for the last 6 years at Rane Corp.

Dana told us another name for acoustic feedback - the Larsen effect, named for Danish physicist Soren Larsen (1871-1957). He noted that there is artistic use of acoustic feedback, although the sound reinforcement operator is interested in reducing it. His PowerPoint slides reviewed basic terms such as the sinusoid, frequency, magnitude, phase, transfer function, open-loop and closed loop response.

Next, he showed the transfer function of a delay block and its effect when viewed as a part of an acoustic feedback loop, showing magnitude and phase of various delays, and the magnitude peaks in the open & closed loop response that that arise at each instance of zero degrees phase when gain is raised. More feedback frequencies occur and are closer when delay is increased (2mS delay has 500Hz spacing of feedback peaks; 10mS-100Hz etc.) Different delays also change the rate of increase or decrease of the feedback component (in dB/sec). He generated an audio demo, changing delay times.

He noted temperature has effects on feedback. With higher temperatures, the speed of sound increases, delay is reduced, and feedback frequencies shift higher. The higher frequency feedback peaks shift more than lower frequencies due to the delay phase shift affecting the shorter wavelengths more. This is important if you later use auto notching filters to suppress feedback, as one will need to account for this shifting.

Real rooms also have reflections which add a delayed signal, and create comb filtering effects. Another speaker's signal, or a second mic can also produce a similar effect. When you total the random multiple reflections plus room modes plus temperature and other factors you get...chaos (or KAOS if you were a "Get Smart" fan, Dana noted).

Controlling feedback electronically was covered next. The major methods were categorized as follows:

1. Adaptive filter modeling - It is widely used in teleconferencing applications to get rid of acoustic echo. Dana categorizes this as OK for speech.
2. all pass filters - Such filters introduce 180 degrees of phase shift to null the feedback points. However, there is phase shift at other that the desired peak, possibly causing a problem elsewhere (the whack-a-mole effect). Dana thought this of limited utility, and gave an audio demo.
3. frequency shifting - constantly adding phase causes a frequency shift - which can be audible. 5-7 Hz of shift is commonly used. There can be side effects such as chirps, and not much gain increase can be had. He gave a humorous vocal audio demo with varying shifts.

• Various Rane T-shirts were won by Peter Toms, Ken Lorenzen, Dave Thiel, Paul Bostic, Dave Tosti-Lane, and Chris Roberts.
• Rane "Dragnet" 32MB USB memory watches were won by Aaron Lipinski, Kevin Whitaker, Josh, Baulch and Rod Johnson.
• A copy of the Rane Pro Audio Reference book by Dennis Bohn was won by Steven Douglas.

There are two significant elements of autonotching:

1. frequency detection - by using an FFT or an adaptive notch filter
2. feedback frequency discrimination - by using the presence of harmonics and/or growth/decay rate of a signal to discriminate feedback from desired sound.

1. transducers can generate harmonics.
2. some natural sounds lack harmonics.
3. It is not impossible to have the natural system transfer function cause 2 feedback frequencies that are harmonically related.

Therefore, Dana felt that utilizing signal strength and growth/decay and checking with a notch is more workable than relying on harmonic content.

All manufacturers provide user controls of notch quantity, bandwidth and depth, as well as whether a filter is fixed, or floating and thus, "recyclable."

Dana left us with some final tips:

• Place notches last (after all system EQ and delay changes).
• Don't use one box to auto place notches and then copy settings to another box that has different delay or processing in it.
• Avoid deep narrow notches, use shallow wide instead (remember you only need to bring the open loop gain below unity gain to keep things stable).
• Remember the temperature shift effect. Higher frequencies are more susceptible to phase differences caused by temperature and delay changes.
• Only keep static filters that you have convinced yourself are necessary.
• Only leave a couple of floating filters active and restrict their depth.