In architectural acoustics, there exist some general rules and guidelines to predict speech intelligibility for a given design (e.g., meeting rooms, theatres, public-address systems covering large areas, or in public transportation). Still, many people experience daily that the result is often far from satisfactory, even for the normal-hearing person.
Speech intelligibility can be predicted by means of calculation models. Usually we take advantage of listening with two ears and thus the calculation models may be enhanced by taking this binaural advantage into account.
In adverse listening conditions, Signal Enhancement procedures may be used. Some information and sound demos may be found at http://hearcom.eu/prof/Algorithmdevelopment.html
Demonstration of speech intelligibility under bad acoustic conditions are available at this page. Some further demos may be found at the Odeon page:
Link to Odeon, a room acoustic simulation software: http://www.odeon.dk