Programs to extract extra channels have been around for a long time. Yamaha pioneered the idea starting with the DSP-1 probably 15 years ago, which extracted 4 additional channels (a pair of front high and wide, and a pair of rear high and wide) from a stereo source.
Oh it goes further back than 15 years. Fosgate was selling 7.1 pre-pros in 1986; Lexicon followed a couple years with their 7.1 pre-pro in 1988. Keep in mind that 5.1 material wouldn't show up until 8 years later. So this is all extraction from stereo sources.
In fact, the first 25 years of consumer surround sound (starting in 1970) was all based on extraction. It's not until the mid 1990s that we consumers finally got discrete multi-channel and people started predicting the end of matrix processing. As if! Since then we got: EX, ES, PLII, Neo:6, PLIIx, DSX, PLIIz, and now Neo:X (just got a listen to DTS' new 11.1 processing at CES).
One thing to keep in mind, the Dolby and DTS modes extract ambience (using only sounds from the recording itself) while Yamaha DSP and Audyssey DSX generate ambience (add reverb and early reflections that weren't originally in the recording).