VOICE SYNTHESIZER, VOICE SYNTHESIS METHOD AND PROGRAM

PROBLEM TO BE SOLVED: To shorten response time from input of a synthesizing target text to output start of voice.SOLUTION: A voice synthesizer 100 stores first synthesized voice data corresponding to a character string of a text in advance in a storage part 120, and a head fraction text extraction part 111 acquires an input text. Then, the head fraction voice selection part 112 extracts a character string coinciding with a head character string of the input text while referring to the storage part 120, outputs the first synthesized voice data corresponding to the coinciding character string to a voice reproduction part 115 and the voice reproduction part 115 reproduces it immediately. During this reproduction processing, a synthesized voice configuration information generation part 113 and a voice synthesis part 114 with a configuration information designation start generation of second synthesized voice data corresponding to a character string following the coinciding character string, and output it to the voice reproduction part 115. As a result of this, the voice reproduction part 115 reproduces the second synthesized voice data consecutively after the first synthesized voice data.