Geometric calculation of absolute phases for parametric stereo decoding

An audio decoding method for decoding a bitstream and reproducing two audio signals, the bitstream including (i) first coded data indicating a downmix signal obtained by downmixing the two audio signals, (ii) second coded data indicating a gain ratio D between the two audio signals, and (iii) third coded data indicating a phase difference θ between the two audio signals,

• wherein the third coded data is a value of cos θ associated with the phase difference θ,
• wherein said audio decoding method comprises:
• decoding the first coded data into the downmix signal;
• transforming the downmix signal into a frequency domain signal, the downmix signal being generated by said decoding;
• determining, using the value of cos θ that is the third coded data, two phase rotators, a first phase rotator of the two phase rotators forming a phase rotation angle α and a second phase rotator of the two phase rotators forming a phase rotation angle β, the phase rotation angle α and the phase rotation angle β being obtained by diagonally dividing a contained angle formed by two adjacent sides of a parallelogram where a length ratio between the two adjacent sides is equal to the gain ratio D indicated by the second coded data, and the contained angle being equal to the phase difference θ indicated by the third coded data;
• separating the frequency domain signal into two separation signals using the two phase rotators and the gain ratio D indicated by the second coded data, a first separation signal of the two separation signals for indicating an angle α as a phase difference between the first separation signal and the downmix signal, and a second separation signal of the two separation signals for indicating an angle β as a phase difference between the second separation signal and the downmix signal;
• and
• inverse transforming the two separation signals, respectively, into time domain signals so as to reproduce the two audio signals,
• wherein said determining determines, as the two phase rotators, either two complex numbers e−jα and ejβ or two conjugate complex numbers ejα and e−jβ of the two complex numbers e−jα and ejβ,
• wherein the bitstream further includes fourth coded data representing phase polarity information S indicating which of the two audio signals has a phase that is ahead of a phase of another of the two audio signals,
• wherein said separating generates the two separation signals by respectively multiplying, with the frequency domain signal transformed from the downmix signal by said transforming, either the two complex numbers or the two conjugate complex numbers associated with the phase polarity information S indicated as the fourth coded data, and
• wherein said inverse transforming accurately reproduces an advancement or a delay of the phase of each of the two audio signals by inversely transforming the two separation signals generated by multiplying, with the frequency domain signal transformed from the downmix signal by said transforming, either the two complex numbers or the two conjugate complex numbers associated with the phase polarity information S indicated as the fourth coded data.

A non-transitory computer-readable recording medium having a program recorded thereon, the program for performing audio decoding processing of decoding a bitstream and reproducing two audio signals, the bitstream including (i) first coded data indicating a downmix signal obtained by downmixing the two audio signals, (ii) second coded data indicating a gain ratio D between the two audio signals, and (iii) third coded data indicating a phase difference θ between the two audio signals,

• wherein the third coded data is a value of cos θ associated with the phase difference θ,
• wherein said program causes a computer to execute a method comprising:
• decoding the first coded data into the downmix signal;
• transforming the downmix signal into a frequency domain signal, the downmix signal being generated by said decoding;
• determining, using the value of cos θ that is the third coded data, two phase rotators, a first phase rotator of the two phase rotators forming a phase rotation angle α, and a second phase rotator of the two phase rotators forming a phase rotation angle β, the phase rotation angle α and the phase rotation angle β being obtained by diagonally dividing a contained angle formed by two adjacent sides of a parallelogram where a length ratio between the two adjacent sides is equal to the gain ratio D indicated by the second coded data, and the contained angle being equal to the phase difference θ indicated by the third coded data;
• separating the frequency domain signal into two separation signals using the two phase rotators and the gain ratio D indicated by the second coded data, a first separation signal of the two separation signals for indicating an angle α as a phase difference between the first separation signal and the downmix signal, and a second separation signal of the two separation signals for indicating an angle β as a phase difference between the second separation signal and the downmix signal;
• and
• inversely transforming the two separation signals, respectively, into time domain signals so as to reproduce the two audio signals,
• wherein said determining determines, as the two phase rotators, either two complex numbers e−jα and ejβ or two conjugate complex numbers ejα and e−jβ of the two complex numbers e−jα and ejβ,
• wherein the bitstream further includes fourth coded data representing phase polarity information S indicating which of the two audio signals has a phase that is ahead of a phase of another of the two audio signals,
• wherein said separating generates the two separation signals by respectively multiplying, with the frequency domain signal transformed from the downmix signal by said transforming, either the two complex numbers or the two conjugate complex numbers associated with the phase polarity information S indicated as the fourth coded data, and
• wherein said inverse transforming accurately reproduces an advancement or a delay of the phase of each of the two audio signals by inversely transforming the two separation signals generated by multiplying, with the frequency domain signal transformed from the downmix signal by said transforming, either the two complex numbers or the two conjugate complex numbers associated with the phase polarity information S indicated as the fourth coded data.