Pronunciation conversion apparatus, pitch mark timing extraction apparatus, methods and programs for the same

A pitch mark timing extraction apparatus comprising:

• a speech segment detection section which detects a speech segment of a speech signal;
• an LPC analysis section which determines an LPC residual signal by using an LPC coefficient a obtained from the speech signal of the speech segment through LPC analysis by using the speech signal and the speech segment, and the first speech signal of the speech segment;
• a pitch mark analysis section which calculates an impulse train corresponding to an EGG glottal closure time by using the speech signal;
• a cross-correlation estimation section which estimates a delay of the LPC residual signal relative to the EGG glottal closure time by estimating a cross-correlation between the LPC residual signal and the impulse train, and
• a pitch mark time estimation section which estimates a pitch mark time based on the delay, the impulse train, and the LPC residual signal.

A pitch mark timing extraction method comprising:

• a speech segment detection step of causing a speech segment detection section to detect a speech segment of a speech signal;
• an LPC analysis step of causing an LPC analysis section to determine an LPC residual signal by using an LPC coefficient a obtained from the speech signal of the speech segment through LPC analysis by using the speech signal and the speech segment, and the first speech signal of the speech segment;
• a pitch mark analysis step of causing a pitch mark analysis section to calculate an impulse train corresponding to an EGG glottal closure time by using the speech signal;
• a cross-correlation estimation step of causing a cross-correlation estimation section to estimate a delay of the LPC residual signal relative to the EGG glottal closure time by estimating a cross-correlation between the LPC residual signal and the impulse train, and
• a pitch mark time estimation step of causing a pitch mark time estimation section to estimate a pitch mark time based on the delay, the impulse train, and the LPC residual signal.