Virtual Sources in Front of the Array

In all cases considered so far, the virtual sources (primary sources) have been restricted to be behind the speaker array by some minimum distance for valid sampling. We can now extend as in WFS to allow sources between the array and the listener, but we still must maintain the same minimum distance, but now from the other side of the array. There are differences, however, to keep in mind relative to the behind-the-array case. For simplicity, consider a line or plane array as a starting point.

• Unlike the primary sources behind the array, those in front of the array have to first propagate the ``wrong way'' to the array to provided recorded signal components (spatial samples) that can be played back from the array to create a converging wavefront back to the virtual source and then on to the listener. This means there is an added delay between the source and the listener, as if the listener can only hear the first-order reflection of the source bouncing off the array, with no direct signal from the virtual source. The listener hears the reflection from the array after it passes through the point of convergence at the virtual source and then propagates to the listener.

• The listener receives a mirrored reflection of the rear of the virtual source. This is no problem for an isotropic source, like any monopole, but it can be an error for distributed sources trying to achieve a specific natural radiation pattern. As a result, primary sources should face the array instead of the listener, and be flipped as needed.

• In offline applications the extra delay for interior sources is easily removed in post-processing.