Abstract: There is provided a method for enlarging a location with optimal three-dimensional audio perception. Optimal three-dimensional audio perception may relate to a fully spatial sound effect. The method includes deriving three-dimensional encoded localization cues from an audio input signal having a first channel signal and a second channel signal; decoding the first channel signal and the second channel signal into a plurality of decoded channel signals, the plurality of decoded channel signals being equal to a number of speaker units; performing crosstalk cancellation on the plurality of decoded channel signals to eliminate crosstalk between the plurality of decoded channel signals; and outputting the plurality of decoded channel signals which have been subjected to crosstalk cancellation to each of the number of speaker units. It is advantageous that the crosstalk cancellation includes further processing to generate a smoothed frequency envelope.

Abstract: There is provided a method for accessing at least one audio file from a collection comprising more than one audio file stored within or accessible with an electronic device. The method includes generating one index comprising of information entries obtained from each of the more than one audio file in the collection, with each audio file in the collection information being linked to at least one information entry; receiving an input during an input reception mode; converting the input into a digital signal using an analog-digital converter; analyzing the digital signal using either frequency spectrum analysis or melody contour analysis into discrete portions; and comparing the discrete portions with the entries in the index. It is advantageous that the audio file is accessed when the discrete portions substantially coincide with at least one of the information entries in the index. The discrete portions may be musical notes, melody contours or waveforms.

Abstract: There is provided a method for accessing at least one audio file from a collection comprising more than one audio file stored within or accessible with an electronic device. The method includes generating one index comprising of information entries obtained from each of the more than one audio file in the collection, with each audio file in the collection information being linked to at least one information entry; receiving an input during an input reception mode; converting the input into a digital signal using an analog-digital converter; analysing the digital signal using either frequency spectrum analysis or melody contour analysis into discrete portions; and comparing the discrete portions with the entries in the index. It is advantageous that the audio file is accessed when the discrete portions substantially coincide with at least one of the information entries in the index. The discrete portions may be musical notes, melody contours or waveforms.

Abstract: There is provided a method for enlarging a location with optimal three-dimensional audio perception. Optimal three-dimensional audio perception may relate to a fully spatial sound effect. The method includes deriving three-dimensional encoded localization cues from an audio input signal having a first channel signal and a second channel signal; decoding the first channel signal and the second channel signal into a plurality of decoded channel signals, the plurality of decoded channel signals being equal to a number of speaker units; performing crosstalk cancellation on the plurality of decoded channel signals to eliminate crosstalk between the plurality of decoded channel signals; and outputting the plurality of decoded channel signals which have been subjected to crosstalk cancellation to each of the number of speaker units. It is advantageous that the crosstalk cancellation includes further processing to generate a smoothed frequency envelope.

Abstract: In a first aspect, there is provided a method for locating a position of at least one sound generating object using at least one audio pulse, with the at least one audio pulse being detected by a plurality of stationary microphones located at a first position being spaced apart by a pre-determined distance. In a second aspect, there is provided a method for locating a position of a microphone using audio pulses emitted from a plurality of sound generating objects. The at least one audio pulse may preferably be in a form of a logarithmic swept sine (LSS) signal, as the LSS signal is detectable at both low volumes and amidst background noises.

Abstract: A method of deriving a compressed acoustic model for speech recognition is disclosed herein. In a described embodiment, the method comprises transforming an acoustic model into an eigenspace at step 20, determining eigenvectors of the eigenspace and their eigenvalues, and selectively encoding dimensions of the eigenvectors based on values of the eigenspace at step 30 to obtain a compressed acoustic model at steps 40 and 50.

Abstract: A method and apparatus for accessing at least one audio file from a collection comprising more than one audio file stored within or accessible with an electronic device. The method includes generating one index comprising information entries obtained from each of the more than one audio file in the collection, with each audio file in the collection information being linked to at least one information entry; receiving a vocal input during a voice reception mode; converting the vocal input into a digital signal using a digital-analog converter; analysing the digital signal using frequency spectrum analysis into discrete portions; and comparing the discrete portions with the entries in the index. It is advantageous that the audio file is accessed when the discrete portions substantially match at least one of the information entries in the index. It is preferable that the discrete portions are either musical notes or waveforms.