This paper addresses the problem of single channel speech enhancement at ve
ry low signal-to-noise ratios (SNR's) (<10 dB), The proposed approach is ba
sed on the introduction of an auditory model in a subtractive-type enhancem
ent process. Single channel subtractive-type algorithms are characterized b
y a tradeoff between the amount of noise reduction, the speech distortion,
and the level of musical residual noise, which can be modified by varying t
he subtraction parameters, Classical algorithms are usually limited to the
use of fixed optimized parameters, which are difficult to choose for all sp
eech and noise conditions. A new computationally efficient algorithm is dev
eloped here based on masking properties of the human auditory system. It al
lows for an automatic adaptation in time and frequency of the parametric en
hancement system, and finds the best tradeoff based on a criterion correlat
ed with perception. This leads to a significant reduction of the unnatural
structure of the residual noise. Objective and subjective evaluation of the
proposed system is performed with several noise types form the Noisex-92 d
atabase, having different time-frequency distributions. The application of
objective measures, the study of the speech spectrograms, as well as subjec
tive listening tests, confirm that the enhanced speech is more pleasant to
a human listener, Finally, the proposed enhancement algorithm is tested as
a front-end processor for speech recognition in noise, resulting in improve
d results over classical subtractive-type algorithms.