1773 search results - page 269 / 355 » Music |
157
click to vote
ICASSP
2008
IEEE
15 years 12 months ago
2008
IEEE
We propose a new approach for singer melody extraction, based on blind source separation techniques. The short time Fourier transform (STFT) of the singer signal is modelled by a ...
106
click to vote
ICASSP
2008
IEEE
15 years 12 months ago
2008
IEEE
Recently the concept of ideal binary time-frequency masks has received attention and their optimality in terms of signalto-noise ratio has been presumed. However the optimality is...
145
click to vote
ICASSP
2008
IEEE
15 years 12 months ago
2008
IEEE
While state-of-the-art approaches obtain an estimate of the a priori SNR by adaptively smoothing its maximum likelihood estimate in the frequency domain, we selectively smooth the...
143
click to vote
ICASSP
2008
IEEE
15 years 12 months ago
2008
IEEE
In this paper, we propose an MMSE a priori SNR estimator for speech enhancement. This estimator has similar benefits to the well-known decision-directed approach, but does not req...
140
click to vote
ICDIM
2008
IEEE
15 years 12 months ago
2008
IEEE
This paper presents a simple mathematical analysis of some features of the Harmony Search algorithm (HS). HS is a recently developed derivative-free optimization algorithm, which ...