Switch uses a speech recognition engine to generate a rough approximation of the words spoken in an audio file. To use this feature, just set the 'Output Format' to 'txt' and click 'Convert', this feature is avaliable on windows vista and later version.