Skip to Main Content
The study of coding for constant-data-rate systems begun in Part I is continued. A simple pulse model of such a system is introduced which is easier to visualize than the correlation detection scheme of Part I. The Wagner code is extended to two multipleerror-correcting schemes. The efficiency of the Wagner code is evaluated by comparing it with Shannon's ideal coding. It is shown that the Wagner code is just the minimum probability of error detector for parity-checked sequences of independent binary digits.