An educational software tool is developed for teaching speech signal coding theory and algorithms with the National Instruments LabVIEW™ package.  We choose to use LabVIEW because of its intuitive visual interface, ability to process real-time signals and capability to interface with DSP hardware.    The tool is based on the National Instruments LabVIEW™ environment. The framework of this tool was built using existing C and MATLAB code as a library along with LabVIEW’s native functionalities. Standardized linear predictive coding algorithms have been implemented; these are used to demonstrate in our DSP classes how digital filters and signal modeling is utilized in cellular and military communications. Experiments covered include the introduction of speech synthesis models, parameterization of speech in terms of filter and excitation parameters, and robustness of speech parameters to additive and channel noise.   The tool provides capabilities for both quantitative and subjective assessment of the synthesized speech signal.       The choice of LabVIEW as a visual programming environment was motivated by the real time signal acquisition capabilities of this environment and the presence of several native graphical functions that enable users to visualize different aspects and parameters of a speech coding algorithm.

The graphical user interface of the LabVIEW speech coding tool is shown in Figure 1. The framework mimics the speech coding standard by defining blocks separately for analysis and synthesis, with additional features for real-time input handling, playback and appropriate graphical plots. The software can access either an audio (‘.wav’) file or real-time speech input.  The user also has options to change certain speech parameters to analyze the performance and behavior of the algorithm under different conditions. The preprocessed input speech is displayed and processed on a frame-by-frame basis. Frame-by-frame display is also used  to view the excitation, the filter parameters, the spectrum of the preprocessed and decoded speech, the quantized LPC spectral envelopes, the pitch estimates, pole-zero plots of the synthesis filter, formants, speech synthesis waveforms, SNRs, etc. The software has options to save and read the input data and the coded data.  The user can also analyze the subjective quality of these algorithms by listening to the synthesized speech with the aid of the playback feature.