The Howling Phenomenon
Acoustic feedback (also known as howling) is inevitable whenever sound reinforcement is necessary. For instance
in large meeting rooms, court rooms, classrooms, auditoriums, theaters, etc. In such large rooms, although audience
(or participants) are seated in the same room, they may be seated far from the presenter or from each other and
may not hear one another clearly.
In sound reinforcement systems, microphones are used to pick up the talker voice, the microphones signals are amplified,
and the amplified signal is played to loudspeakers placed in the same room, so that audience can hear clearly.
The problem, however, is that the microphones not only pick up the presenter's voice. They also pick up the loudspeaker
sound after traveling through the room back to the microphones, causing a closed electroacoustic loop.
If the loop gain at any frequency becomes larger than 1, the loop becomes
unstable at that frequency, and the system starts oscillation, producing the well known howling phenomenon.
To avoid howling, an electronic device (Acoustic Feedback Canceller or Howling Suppressor) must be used to monitor
the closed loop and ensure that the loop stays stable at all times. The sample audio files below demonstrate the Acoustic Feedback Canceller at work.
The audio signal at the top has been recorded in a room where sound reinforcement is used at moderate
amplifier gain with the Acoustic Feedback Canceller switched OFF. In this case the coupling between the loudspeaker and microphone
is already too high and oscillation at around 5.8 kHz and many other frequencies are easily excited. The 5.8 KHz tone and other oscillations increase in
amplitude rapidly resulting in saturating the amplifier and other equipment.
The sample at the bottom is recorded in the same room with the amplifier gain increased by 3 dB after the Acoustic Feedback Cancelelr has been switched ON. This sample
audio file shows that the AFC suppressed the 5.8 KHz oscillation even before it happens and other oscillations almost
immediately suppressed as they appear until the loop has been stabilized completely.
- Efficient block frequency domain processing resulting in low resource consumption and high quality loudspeaker signal.
- Minimum distortion compared to solution relying on frequency shifting or phase modulation.
- Up to 14 dB additional gain in most (well designed) sound reinforcement installations.
- Supports multiple microphones.
- Multiple loudspeakers can be connected to the Acoustic Feedback Canceller and placed at different locations in the room with no degradation in performance.
- Fast and robust convergence even with multiple microphones and multiple loudspeakers in one room.
- Low processing delay (typical 5.3ms at 48 kHz sampling rate).
- Processing delay depends on user adjustable block length.
- Longer processing delay results in lower CPU cycles but higher memory consumption.
And of course the other way around, shorter processing delay results in higher CPU load but lower memory consumption.
EVALUATING THE ACOUSTIC FEEDBACK CANCELLER
A real-time demonstrator is directly available on the
This demonstrator also includes the CANEC speech enhancement, to request this demo,
please fill the CANEC demonstration request form.
For more information on the Acoustic Feedback Canceller algorithm we appreciate taking the time to contact us.