|
Features
In ANC (Active Noise Control), feed-forward control is used in many cases. You input a reference signal near the noise source, convolve the signal with an inverse filter to make an anti-phase control signal, and emit it out of a control speaker. Control algorithm adjusts the inverse filters to minimize the power of the controlled point signal, which is called an error signal. In other cases, feed-back control is used due to the following reasons.
(1) You cannot input a reference signal because you cannot locate the noise sources. In feed-forward control, an anti-phase control signal is calculated and emitted to reduce noise level while the noise signal is propagating in the air or other mediums. So, the time-lag from input to output in the control system needs to be short, but the time-lag always exists because of the limitation of hardware. You usually keep enough distance between a reference input and a control speaker to let the control signal overcome the noise propagation. In feed-back control, you set the reference input and the error input on the same place, so you cannot stall for time to keep up with the noise signal propagation. From the point of view of feed-forward control, it might be impossible or unstable to control by feed-back method because the control signal never catches up with the noise signal, especially in a not-minimum-phase system like in the air, except for the following cases.
(1) The noise is periodic. Redec Feed-back ANC "DQancFB4" operates at very high sampling frequency ( max.48KHz ) to reduce the system time delay, and has as high-performance as to convolve long-tap inverse filters (max. 2K taps). Moreover, it has howling cancellers to operate with stability.
|
Last update : 2010/06/19
|