aspcol.soundfieldcontrol.vast

aspcol.soundfieldcontrol.vast(Rb, Rd, rb, mu, rank, reg=0)

Frequency domain variable span trade-off filter for sound zone control

Parameters:

  • Rb (ndarray of shape (num_freq, num_ls, num_ls)) – spatial covariance associated with the bright zone

  • Rd (ndarray of shape (num_freq, num_ls, num_ls)) – spatial covariance associated with the dark zones

  • rb (ndarray of shape (num_freq, num_ls, num_virt_src)) – cross correlation between paths to bright zone and virtual source paths

  • mu (int) – non-negative, weights effort between error in dark and bright zones

  • rank (int) – between 1 and num_ls. Applies low-rank approximation via GEVD. Lower rank gives higher acoustic contrast, higher rank gives lower distortion

  • reg (float) – non-negative, applies l2 regularization to the loudspeaker effort, by adding reg * np.eye() to the dark zone spatial covariance

Returns:

w – the control filter used to generate loudspeaker signals. loudspeaker signals can be generated from the virtual source signals v : (num_freq, num_virt_src, 1) as w @ v.

Return type:

ndarray of shape (num_freq, num_ls, num_virt_src)

References

[leeFast2020] T. Lee, L. Shi, J. K. Nielsen, and M. G. Christensen, “Fast generation of sound zones using variable span trade-off filters in the DFT-domain,” IEEE/ACM Transactions on Audio, Speech, and Language Processing, vol. 29, pp. 363–378, Dec. 2020, doi: 10.1109/TASLP.2020.3042701.