AIB2 Components - Voice MFCC

The MFCC component provides a flexible plug and play implementation of the Mel Spectrum set of triangular filters. These filters and the resulting coefficients are computed based on the incoming FFT output. In digital signal processing of speech, including speaker recognition and word recognition, the Mel Spectrum is commonly a critical component of filtering and spectral analysis, as it focuses largely on the range of frequencies produced by humans and recognized by humans. Using the component properties dialog box, the filters can be set, along with which ones will be exposed as outputs for consumption by other components. See the detail on inputs and outputs below for a description of how to utilize. The dialog box will visually present the use and application of these filters as well.