Recognition of human speech remains a very difficult task for machines
, although one that humans effortlessly achieve every day. Nevertheles
s, the underlying technology has matured enough for many advanced spee
ch-based applications to become commercially viable.