Classification

Home > Technology > Classification

Active Noise Control (ANC) Technology Classification I

Single-Channel ANC vs. Multi-Channel ANC
  • 01 Single-Channel ANC
    • Uses no reference microphone or one reference microphone
    • Uses one control speaker
    • Uses one error microphone
    • 1(0) reference mic x 1 control speaker x 1 error mic
Single-Channel ANC
Multi-Channel ANC
  • 02 Multi-Channel ANC
    • Uses one or more reference microphones
    • Uses one or more control speaks
    • Uses one or more error microphones
    • J reference mic x K control speaker x M error mic (e.g. 12 x 8 x 8)
Why is Multi-Channel ANC Necessary?
  • To overcome a narrow Quiet Zone, Multi-Channel ANC overlaps and expands the Quiet Zone.

Active Noise Control Technology Classification II

  • 01 Feedforward ANC
    • Existence of a reference microphone
    • Uses a reference microphone signal as prior information to remove random noise
    Feedforward
  • 02 Block Processing ANC
    • Reduces computation through FFT-based Fast Convolution, useful in multi-channel ANC with high computational demands
    • Increases output latency and adaptive filter update latency due to block buffering
    • Operates on a DMA basis; low interrupt overhead
    • Frequency-domain algorithm: Frequency-domain Block FxLMS(FBFxLMS)
    • Utilizes the temporal statistical properties of signals obtained through block processing
    • Representative algorithms
      • Self-Orthogonalized FBFxLMS: Easy calculation of the inverse of Diagonalized Autocorrelation Matrix
      • Subband FxLMS: Filter-bank based algorithm; computation reduction through Decimation
  • 03 Hybrid Processing ANC
    • For the output of ANC controller, sample-processed in Time-domain
    • For the adaptive filter update of the controller, block-processed in Frequency-domain
    • No output latency, but there is adaptive filter update latency
    • Disadvantageous when the noise characteristics change rapidly over time
    • Utilizes the temporal statistical properties of signals obtained through block processing
    • Representative algorithms
      • Time-Frequency FxLMS (Delayless Frequency-domain FxLMS)
      • Delayless Subband FxLMS