The amount of the compiled resources on Child Directed Speech has rapidly increased since the first attempt in 1967. Most of the records are digitized and are sitting in the server of the CHILDES project. The transcripts of the audio and the video were used in many studies. Although the automatic processing of the orthographic transcripts can be done in vast amounts, thanks to the natural language processing tools, this is not true for the audio. The studies that analyzed the wave had to go through a tedious hand tagging procedure. 

The main purpose of this document is to suggest a new method to automatically analyze the wave by using speech recognition tool box. This document gives a detailed sketch starting from a CHILDES corpus to getting detailed phonemic transcription of the audio, by using HTK and doing forced alignment. Although the sketch is English specific, the extension to some other languages are described in other sources.