Speech synthesizer, audio watermarking information detection apparatus, speech synthesizing method, audio watermarking information detection method, and computer program product
According to an embodiment, a speech synthesizer includes a source generator, a phase modulator, and a vocal tract filter unit. The source generator generates a source signal by using a fundamental frequency sequence and a pulse signal. The phase modulator modulates, with respect to the source signal generated by the source generator, a phase of the pulse signal at each pitch mark based on audio watermarking information. The vocal tract filter unit generates a speech signal by using a spectrum parameter sequence with respect to the source signal in which the phase of the pulse signal is modulated by the phase modulator.
1. An audio watermarking information detection apparatus comprising:
a memory; and
one or more processors configured to function as a pitch mark estimator, a phase extractor, a representative phase calculator and a determination unit, wherein
the pitch mark estimator estimates a pitch mark of a synthesized speech in which audio watermarking information is embedded and extracts a speech at each estimated pitch mark;
the phase extractor extracts a phase of the speech extracted by the pitch mark estimator;
the representative phase calculator calculates a representative phase to be a representative of a plurality of frequency bins from the phase extracted by the phase extractor; and
the determination unit determines, based on the representative phase, whether the audio watermarking information exists in the synthesized speech.
2. The audio watermarking information detection apparatus according to claim 1 , wherein the determination unit
calculates, in each frame which is a predetermined period, an inclination indicating a variation of the representative phase in elapse of time, and
determines, based on a frequency of the inclination, whether there is the audio watermarking information.
3. The audio watermarking information detection apparatus according to claim 1 , wherein the determination unit
calculates, in each frame which is a predetermined period, a correlation coefficient between the representative phase and a reference straight line which is assumed as an ideal value of a variation of the representative phase in elapse of time, and
determines that there is the audio watermarking information when the correlation coefficient exceeds a predetermined threshold.
4. An audio watermarking information detection method employed for an audio watermarking information detection apparatus including a memory and one or more processors configured to function as a pitch mark estimator, a phase extractor, a representative phase calculator and a determination unit, comprising:
estimating, by the itch mark estimator, a pitch mark of a synthesized speech in which audio watermarking information is embedded and extracting a speech at each estimated pitch mark;
extracting, by the phase extractor, a phase of the extracted speech;
calculating, by representative phase calculator, from the extracted phase, a representative phase to be a representative of a plurality of frequency bins; and
determining, by the determination unit, based on the representative phase, whether the audio watermarking information exists in the synthesized speech.
5. A computer program product comprising a non-transitory computer-readable medium that includes an audio watermarking information detection program to cause a computer to execute:
estimating a pitch mark of a synthesized speech in which audio watermarking information is embedded and extracting a speech at each estimated pitch mark,
extracting a phase of the extracted speech,
calculating, from the extracted phase, a representative phase to be a representative of a plurality of frequency bins, and
determining, based on the representative phase, whether the audio watermarking information exists in the synthesized speech.