Analysis-by-Synthesis of Speech
Step1: Choose any of the pre-recorded speech signals from the 'Select an audio file' dropdown menu.
Step2: Play the audio file and observe the waves.
Step3: Click on the 'Generate Spectrum' button, 'LP order', 'Graphy Type' dropdown options, Excitation, Log Energy and Pitch Countour will be displayed.
Step4: Study the characteristics of generated 'LP residual' signal. This residual is generated using a 'LP order' of 11 and will be used to synthesize the speech signal in further stages of experiment.
Step5: Observe the excitation signal that would be used to synthesize the speech signal. Also choose the order of LP coefficients to be used to synthesize the speech.
Step6: Study the synthesized speech signal according to the parameters chosen. Study the characteristics of the re synthesised speech by using different excitation types, namely, 'LP Residual' and 'Impulse' excitation excitation.
Step7: Study the effect of using the vocal tract system derived using different LP orders for resynthesis. Zoom into the waveform to study different regions of the synthesized waveform.
Step8: Repeat the experiment for sentences recorded with your own voice.