LTranscriber - a tool for segmentation, labeling and transcription of audio recordings
|Inserted by:||Ing. Kamil Ekštein, Ph.D.|
|Date last modified:||29.12.2013|
|Year of insertion||2009|
|Number of downloads:||1|
The LTranscriber tool was created within the framework of the NPV2 project Cot-Sewing 2C06009 and its purpose is to considerably ease the task of preparing training and testing data for automatic speech recognition systems. This tool presents a comfortable and efficient way of creating transcripts of the audio recordings that constitute training corpora of the speech recognition system. In order to make the recognition system properly trained it is necessary to have not only a sufficient quantity of the audio recordings: It is necessary to have a transcription of each such recording too. LTranscriber enables to segment, label and transcribe the recording into either phonetic or orthographic form. The design of the of the user interface helps the expert worker to carry out the transcription significantly faster and easier compared to using e.g. a standard text editor.