NUMERICAL OPERATIONS ON RANDOM FUNCTIONS.
Abstract
The work investigates the problem of obtaining a sampled-data representation of a random function, and operating on this representation to obtain approximate reconstruction of the random function or the result of performing a linear operation on it. The author considers cases in which the input is available to the presampling filter only for a finite time, and only a finite number of samples are available for reconstruction. He describes a set of computer programs in which the optimum filters are designed, and the associated mean squared errors evaluated, by time-domain techniques. These programs compute the effects of both sampling errors and quantization errors. He includes quantitative comparisons of PCM system performance with Shannon's rate-distortion function. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 01, 1966
- Accession Number
- AD0639300
Entities
People
- Willis C. Kellogg