Digital data encoding apparatus and method thereof

A digital data encoding apparatus for encoding first channel digital data divided into blocks each containing a predetermined number of samples and second channel digital data divided into blocks each containing said predetermined number of samples, comprising:

• coefficient data generating means for transforming said first channel digital data into data on a frequency axis on a block-by-block basis, to thereby generate first channel coefficient data on a frequency-by-frequency basis while transforming said second channel digital data into data on the frequency axis on a block-by-block basis to thereby generate second channel coefficient data on a frequency-by-frequency basis;
• channel sharing means for determining mean values of said first channel coefficient data and said second channel coefficient data to thereby generate shared-channel data;
• channel sharing/non-sharing decision means for determining a degree of correlation between said first channel coefficient data and said second channel coefficient data to thereby make decision as to which of said shared-channel data and said first and second channel coefficient data is to be selected in dependence on said determined degree of correlation;
• switch means for selecting either said shared-channel data inputted from said channel sharing means or said first and second channel coefficient data inputted from said coefficient data generating means in dependence on a result of said decision made by said channel sharing/non-sharing decision means, to thereby output as output data either said shared-channel data or said first and second channel coefficient data as selected;
• and
• floating-point transform means inputted with said output data outputted from said switch means, for dividing said inputted output data into a plurality of sub-bands each containing a single or a plurality of said inputted output data, while allowing said divided output data to undergo a floating-point transform on a subband basis for thereby generating one subband-common exponent data, which is common to said divided output data contained in each of said sub-bands, and mantissa data the number of which is equal to the number of said divided output data contained in each of said sub-bands.

A digital data encoding apparatus for encoding first channel digital data divided into blocks each containing a predetermined number of samples and second channel digital data divided into blocks each containing said predetermined number of samples, comprising:

• coefficient data generating means for transforming said first channel digital data into data on a frequency axis on a block-by-block basis to thereby generate first channel coefficient data on a frequency-by-frequency basis while transforming said second channel digital data into data on the frequency axis on a block-by-block basis to thereby generate second channel coefficient data on a frequency-by-frequency basis;
• sign inverting means for detecting on a frequency-by-frequency basis whether or not sign of said first channel coefficient data coincides with sign of said second channel coefficient data, wherein unless the sign of said first channel coefficient data coincides with the sign of said second channel coefficient data, said sign inverting means inverting either the sign of said first channel coefficient data or the sign of said second channel coefficient data, to thereby make the sign of said first channel coefficient data coincide with the sign of said second channel coefficient data;
• channel sharing means for determining mean values of said first channel coefficient data and said second channel coefficient data both of which have the respective signs made to coincide with each other by said sign inverting means, to thereby generate shared-channel data;
• channel sharing/non-sharing decision means for determining a degree of correlation between said first channel coefficient data and said second channel coefficient data to thereby make decision as to which of said shared-channel data and said first and second channel coefficient data is to be selected in dependence on said determined degree of correlation;
• switch means for selecting either said shared-channel data inputted from said channel sharing means or said first and second channel coefficient data inputted from said coefficient data generating means in dependence on a result of said decision made by said channel sharing/non-sharing decision means, to thereby output as output data either said shared-channel data or said first and second channel coefficient data as selected;
• and
• floating-point transform means inputted with said output data from said switch means, for dividing said inputted output data into a plurality of sub-bands each containing a single or a plurality of said inputted output data, while allowing said divided output data to undergo a floating-point transform on a subband basis for thereby generating one subband-common exponent data, which is common to said divided output data contained in each of said sub-bands, and mantissa data, the number of which is equal to the number of said divided output data contained in each of said sub-bands.

A digital data encoding method for encoding first channel digital data divided into blocks each containing a predetermined number of samples and second channel digital data divided into blocks each containing said predetermined number of samples, comprising:

• a first coefficient data generating step of transforming said first channel digital data into data on a frequency axis on a block-by-block basis to thereby generate first channel coefficient data on a frequency-by-frequency basis;
• a second coefficient data generating step of transforming said second channel digital data into data on the frequency axis on a block-by-block basis to thereby generate second channel coefficient data on a frequency-by-frequency basis;
• a sign inverting step of detecting on a frequency-by-frequency basis whether or not sign of said first channel coefficient data coincides with sign of said second channel coefficient data, wherein unless the sign of said first channel coefficient data coincides with the sign of said second channel coefficient data, either the sign of said first channel coefficient data or the sign of said second channel coefficient data is inverted, to thereby make the sign of said first channel coefficient data coincide with the sign of said second channel coefficient data;
• a channel sharing step of determining mean values of said first channel coefficient data and said second channel coefficient data both of which have the respective signs made to coincide with each other, to thereby generate shared-channel data;
• and
• a floating-point transform step of dividing said shared-channel data into a plurality of sub-bands each containing a single or a plurality of said inputted shared-channel data, to allow said divided shared-channel data to undergo a floating-point transform on a subband basis for thereby generating one subband-common exponent data, which is common to said divided shared-channel data contained in each of said sub-bands, and mantissa data the number of which is equal to the number of said divided shared-channel data contained in each of said sub-bands.

A digital data encoding method for encoding first channel digital data divided into blocks each containing a predetermined number of samples and second channel digital data divided into blocks each containing said predetermined number of samples, comprising:

• a first coefficient data generating step of transforming said first channel digital data into data on a frequency axis on a block-by-block basis to thereby generate first channel coefficient data on a frequency-by-frequency basis;
• a second coefficient data generating step of transforming said second channel digital data into data on the frequency axis on a block-by-block basis to thereby generate second channel coefficient data on a frequency-by-frequency basis;
• a channel sharing step of determining mean values of said first channel coefficient data and said second channel coefficient, to thereby generate shared-channel data;
• a channel sharing/non-sharing decision step of determining a degree of correlation between said first channel coefficient data and said second channel coefficient data to thereby make decision as to which of said shared-channel data and said first and second channel coefficient data is to be selected in dependence on said determined degree of correlation;
• a switch step of selecting either said shared-channel data or said first and second channel coefficient data in dependence on a result of said decision made by said channel sharing/non-sharing decision step, to thereby output as output data either said shared-channel data or said first and second channel coefficient data as selected;
• and
• a floating-point transform step of dividing said output data into a plurality of sub-bands each containing a single or a plurality of said output data, to allow said divided output data to undergo a floating-point transform on a subband basis for thereby generating one subband-common exponent data, which is common to said divided output data contained in each of said sub-bands, and mantissa data the number of which is equal to the number of said divided output data contained in each of said sub-bands.

A digital data encoding method for encoding first channel digital data divided into blocks each containing a predetermined number of samples and second channel digital data divided into blocks each containing said predetermined number of samples, comprising:

• a first coefficient data generating step of transforming said first channel digital data into data on a frequency axis on a block-by-block basis to thereby generate first channel coefficient data on a frequency-by-frequency basis;
• a second coefficient data generating step of transforming said second channel digital data into data on the frequency axis on a block-by-block basis to thereby generate second channel coefficient data on a frequency-by-frequency basis;
• a sign inverting step of detecting on a frequency-by-frequency basis whether or not sign of said first channel coefficient data coincides with sign of said second channel coefficient data, wherein unless the sign of said first channel coefficient data coincides with the sign of said second channel coefficient data, either the sign of said first channel coefficient data or the sign of said second channel coefficient data is inverted, to thereby make the sign of said first channel coefficient data coincide with the sign of said second channel coefficient data;
• a channel sharing step of determining mean values of said first channel coefficient data and said second channel coefficient data both of which have the respective signs made to coincide with each other, to thereby generate shared-channel data;
• a channel sharing/non-sharing decision step of determining a degree of correlation between said first channel coefficient data and said second channel coefficient data to thereby decide which of said shared-channel data and said first and second channel coefficient data is to be selected in dependence on said determined degree of correlation;
• a switching step of selecting either said shared-channel data or said first and second channel coefficient data in dependence on a result of said decision made in said channel sharing/non-sharing decision step, to thereby output as output data either said shared-channel data or said first and second channel coefficient data as selected;
• and
• a floating-point transform step of inputting said outputted data and dividing said output data into a plurality of sub-bands each containing a single or a plurality of said output data, to allow said divided output data to undergo a floating-point transform on a subband basis for thereby generating one subband-common exponent data, which is common to said divided output data contained in each of said sub-bands, and mantissa data the number of which is equal to the number of said divided output data contained in each of said sub-bands.