A speech signal filterbank, using melscale frequency and fr a mebanking.- 1. For speech signal can be represented as a discrete sequence of fr a mes (or feature vectors) that can be used as the input to a speech recogniser.　Important ideas and techniques that are used in the assignment are the design of a (Mel frequency scale) audio filterbank, , windowing of a continuous audio signal, spectrum analysis of the signal, filtering as multiplication in the frequency domain, the visual representation of a speech signal as a spectrogram, appreciation of the acoustic variability in real speech utterances.



2. To gain a deeper knowledge of the application of MATLAB programming to audio signal processing.



3. To gain practice in the art of writing a formal report: structure, content, style, use of diagrams, presentation etc. etc.