This looks like technology that could go mainstream very soon. The Sensear system employs multiple mikes in order to calculate even the direction of ambient noise and then apply noise reduction to both in-ear sound and also your outgoing voice. Via IntoMobile

Read on in the full article (http://www.allaboutsymbian.com/news/item/5262_Sci-fi_level_noise_reduction_o.php).

Active noise reduction sucks. What you need is passive noise reduction via a good set of IEM's.