Measurement of Urdu Speaker´s Speaking Speed in Ordinary Conversation

Speaker identification traditionally employs technologies of MRCC and LPC, and most of the proven existing technologies are strictly text-dependent. New technologies are text independent. Whatever technologies are used for speaker identification their performance degrades if the list of the speakers stored for identification grows. Therefore, if other pertinent parameters are added to the extracted data from the speaker´s speech, speaker identification will get enhanced. This research aims to enhance speaker identification by adding a new parameter namely speaker´s speaking speed to the extracted data to enhance the accuracy of the speaker identification. This paper has manually determined the speaking speed from many recorded conversations, which were saved in computer by Bluetooth. A reputed software Praat was used to watch sound graphics and to make calculations. The Praat´s calculation confirm that our study can be used to make an algorithm to automatically determine correct speaker speed.