One major complaint to surface since the introduction of surround sound to television is the audibility of the announcer, especially during surround-rich programming like sports. It seems to be a recurring problem that frequently happens when the crowd swells and the announcers are overpowered. The problem is made worse by certain "hyper surround" modes included with many new television sets--and in many cases these effects are active by default.
These modes work by breaking two-channel audio into sum and difference signals, and processing the difference signal to make it stand out. The sum, or mono signal contains all of the audio while the difference signal contains only the information unique to each channel. Since the difference signal then contains all of the "stereo" information (with dialog being almost exclusively in the sum channel), the processing results in a wider or "more stereo sounding" signal when recombined to left and right.
Unfortunately, surround programming also relies on this difference signal to carry audio intended for the surround speakers--this includes crowd noise. So, when a mix aggressively pushes crowd audio into the surround channels, the stereo difference channel grows in intensity. The train wreck occurs when the "hyper surround" process further accentuates this difference signal and the dialog gets buried.
The solution? Turn off the hyper surround mode in all of the television sets, or apply pre-processing at the point of transmission to minimize these problems.
