| 摘要 |
<p>Improved analog to digital and digital to analog signal processors are disclosed wherein a quick approximation of the input signal to the signal processor is attained and a more accurate approximation is later attained for the input signal. For the analog to digital conversion, a standard analog to digital converter having a finite resolution or a predetermined quantization error is used to create part of the digital representation. The remainder of digital representation is created by processing an error signal due to the finite resolution of a standard analog to digital converter and the infinite resolution of the analog input signal to a delta modulator. In a digital to analog signal processing circuit, the first part of the digital representation is converted by a standard digital to analog converter and the remaining portion is integrated into a frequency limited analog signal and then summed to reconstruct the analog signal. To avoid overshoot, the integrators of the signal processors are preset based upon the increase or the decrease in the first part. Further, the rate of integration is controlled so that the rate slows as the digital representation of the analog signal becomes increasingly more accurate representation of the input signal.</p> |
