This attribute specifies whether the CCAP is to average spectral frequency domain sample power to remove spurious spectral peaks and troughs, and the number of samples to use to calculate the average power. The CCAP should not calculate the average of the upstream spectrum samples when the value of this attribute is zero. The CCAP should calculate the average power of upstream spectrum samples, over the number of samples specified by this attribute when the value of is nonzero. The CCAP should use quantities in the linear power domain when performing time averaging over multiple spectra. Time averaging provides for a smoother resulting spectrum. In time averaging, the spectrum is captured multiple times, and each FFT bin is averaged over the successive values in that bin to provide the final spectrum value for that bin. A leaky integrator may be used to perform the averaging. Let x = x_I + j*x_Q be the vector of time domain input samples to the FFT and y = y_I + j*y_Q be the vector of complex output frequency bin values of the FFT. Then p = |y|^2 = y_I^2 + y_Q^2 is the power or squared magnitude of the bin values and r = |y| = sqrt(y_I^2 + y_Q^2) is the magnitude of the bin values. The values may be converted to dB using s = 10*log10(p) = 20*log10(r). Only values of p are used for time averaging. Magnitude values need to be squared before averaging. dB values need to be converted into linear power using p = 10^(s/10) before averaging; after averaging, the smoothed bin values may be converted back to dB.