To provide a computerized method which aligns text segments of a text file with voice segments of a voice file.

The computerized method aligns the text segments 310 of the text file with the voice segments 330 of the voice file. The text file includes written words and the voice includes spoken word. A vocabulary and a language model are generated from the text segments 310. The vocabulary and language model are used to recognize a word list 331 from the voice segments 330. The word list 331 is aligned with the text segments 310 and corresponding anchors are selected with the word list 331 and text segments 310. The anchors 200 are used to section the text segments 310 and voice segments 330 into unaligned and aligned segments. Those steps are repeated for the unaligned segments until end conditions are met.