At Gerber Acoustic Systems we have been using downward expanders for many many years as inserts on the speech subgroups. We do this to minimize the amplifying of the inevitable acoustical reflections from the room returning to the open microphones in the room, and thus lowering the gain before feedback. A downward expander will suppress sounds below the threshold volume that the engineer sets. We have used the Symetrix 528 channel strip for years, but have just now(Dec 20 2014) purchased the Neve 5045 Primary Source processor, and are awaiting its arrival. This purchase will cost our customers no additional rental fee when renting our complete sound system packages.
The practice of using downward expanders kind of began years ago, before the advent of digital audio consoles, with awards shows broadcasts, where very clever broadcast audio engineers were able to use the decode card device of a certain multi band Dolby noise reduction system to push down all the background room reverb, and other acoustic noise floor sounds, to deliver a dry, articulate speech signal to the broadcast audio signal. It worked pretty well, and these guys were definitely very cool in doing this.
Recently Cedar Audio, a long time audio de-noising company, has come out with a few rather expensive digital de noisers and downward expanders that , I am sure, are very effective. They are still somewhat rare in the rental inventories of most rental house, and due to their significant purchase price,are an extra cost item to the producer.
The Neve Primary Source Processor, which can add as much as 20 db of gain before feedback, is a 2 chnl analog unit, and because it is fairly affordable, can be included in the sound system package for no additional cost.