We introduce a speech analysis system that performs the complete recording-playback interface, and allows some edition facilities. The system can display parameters such as energy, zero-crossing rate, fundamental frequency (pitch contour), spectrogram and LPC envelope. Another capabilities are filtering of the pre-recorded speech, labelling of the different acoustic segments and the possibility of laser screen hardcopies.
The system needs a PC and some specific hardware. The PC must be AT or later, with a VGA graphic card and a serial mouse. The specific hardware is a DSP based board which was completely developed in the Department. The software is divided in two blocks: the software running on the iX86 processor, which controls the system and displays the speech information at screen, and the program running on the DSP, that performs the signal processing, the recording and the playback.
Since the system is oriented to not technical people (phoneticians, etc.) we focused on the user interface to make it as easy and friendly as possible. The user can configure the windows in the screen in a wide manner using a keyboard or the mouse driven menu system.
This speech analysis device takes advantage of all the standard PC features (video monitor, hard disk, printer, memory, etc.), what results in the lower cost compared with any other analysis tool in the market.