摘要 |
A voice-excited vocoder analyzer divides the audio input signals into a baseband and upper band by deriving Fourier transform coefficients and transmitting the coefficients in a time-multiplex frame. The coefficients (an and bn) are derived in two paths, each path multiplying the input signal by a stepped-frequency local source, (one sine, one cosine), the product feeding a recirculating shift register. The vocoder synthesizer effectively performs the same operations in reverse. |