A voice retrieval apparatus executes processes of: obtaining, from a time length memory, a continuous time length for each phoneme contained in a phoneme string of a retrieval string; obtaining user-specified information on an utterance rate; changing the continuous time length for each obtained phoneme in accordance with the obtained information; deriving, based on the changed continuous time length, an utterance time length of voices corresponding to the retrieval string; specifying a plurality of likelihood obtainment segments of the derived utterance time length in a time length of a retrieval sound signal; obtaining a likelihood showing a plausibility that the specified likelihood obtainment segment is a segment where the voices are uttered; and identifying, based on the obtained likelihood, an estimation segment where, within the retrieval sound signal, utterance of the voices is estimated, the estimation segment being identified for each specified likelihood obtainment segment.