Orthogonal transform apparatus, orthogonal transform method, orthogonal transform computer program, and audio decoding apparatus

1. An audio decoding apparatus for decoding a multichannel audio signal from a data stream containing a main signal code into which a main signal representing a main component of each channel obtained by downmixing each channel signal of the multichannel audio signal is encoded, a residual signal code into which coefficients obtained by performing a modified discrete cosine transform on a residual signal orthogonal to the main signal are encoded, and a spatial information code into which spatial information representing the degree of interchannel similarity and interchannel intensity difference is encoded, the audio decoding apparatus comprising:

a processor configured to:

demultiplex the main signal code, the residual signal code, and the spatial information code from the data stream;

reconstruct the main signal in a time domain by decoding the main signal code;

transform the main signal in the time domain into quadrature mirror filter coefficients in a time-frequency domain by applying quadrature mirror filtering to the main signal;

reconstruct the spatial information by decoding the spatial information code;

reconstruct the modified discrete cosine transform coefficients of the residual signal by decoding the residual signal code;

transform, for each of a plurality of prescribed intervals generated so as to overlap each other by one half by dividing an entire frequency band, the modified discrete cosine transform coefficients of the residual signal contained in the prescribed interval into quadrature mirror filter coefficients in the time-frequency domain;

compute quadrature mirror filter coefficients for each channel of the audio signal by upmixing the quadrature mirror filter coefficients of the main signal and the quadrature mirror filter coefficients of the residual signal by using the spatial information; and

reconstruct each channel signal of the audio signal by applying inverse quadrature mirror filtering to the quadrature mirror filter coefficients of each channel, and wherein

the transforming the modified discrete cosine transform coefficients of the residual signal contained in the prescribed interval into quadrature mirror filter coefficients in the time-frequency domain comprises:

interchanging the modified discrete cosine transform coefficients contained in a first half of the prescribed interval with the modified discrete cosine transform coefficients contained in a second half of the prescribed interval;

inverting the sign of the modified discrete cosine transform coefficients contained in the second half of the prescribed interval after the interchange;

computing real components of the quadrature mirror filter coefficients of the residual signal by performing an inverse modified discrete cosine transform using a fast Fourier transform to both of the modified discrete cosine transform coefficients contained in the first half of the prescribed interval after the interchange and the sign-inverted modified discrete cosine transform coefficients contained in the second half of the prescribed interval after the interchange;

computing imaginary components of the quadrature mirror filter coefficients of the residual signal by performing an inverse modified discrete sine transform using a fast Fourier transform to both of the modified discrete cosine transform coefficients contained in the first half of the prescribed interval after the interchange and the sign-inverted modified discrete cosine transform coefficients contained in the second half of the prescribed interval after the interchange; and

obtaining the quadrature mirror filter coefficients of the residual signal by combining the real components of the quadrature mirror filter coefficients of the residual signal with the imaginary components of the quadrature mirror filter coefficients of the residual signal.

2. The audio decoding apparatus according to claim 1 , wherein computing the real components of the quadrature mirror filter coefficients comprises:

computing first coefficients by performing the inverse modified discrete cosine transform using the fast Fourier transform to the modified discrete cosine transform coefficients contained in the first half of the prescribed interval after the interchange;

computing second coefficients by performing the inverse modified discrete cosine transform using the fast Fourier transform to the sign-inverted modified discrete cosine transform coefficients contained in the second half of the prescribed interval after the interchange; and

obtaining the real components of the quadrature mirror filter coefficients by combining the first and second coefficients.

3. The audio decoding apparatus according to claim 1 , wherein computing the imaginary components of the quadrature mirror filter coefficients comprises:

computing first coefficients by performing the inverse modified discrete sine transform using the fast Fourier transform to the modified discrete cosine transform coefficients contained in the first half of the prescribed interval after the interchange;

computing second coefficients by performing the inverse modified discrete sine transform using the fast Fourier transform to the sign-inverted modified discrete cosine transform coefficients contained in the second half of the prescribed interval after the interchange; and

obtaining the imaginary components of the quadrature mirror filter coefficients by combining the first and second coefficients.