How does a speech recognition work?

When sound is inputted, sound card in device converts sound wave into digital data, software takes digital value and breaks them up into phonemes, phonemes compared with words in dictionary, word spoken will then be suggested and output