# Discrete Nyquist-Shannon capacity

1. Sep 2, 2011

### xts

I've already posted this question some time ago, but got no answer. So let me try again:

Variation of Nyquist-Shannon's noisy channel.

We have some classical continuous real physical variable (e.g. voltage).
Alice is allowed to assign any value s she likes in a period 0<s<1.
The signal is then obscured by adding Gaussian noise of some sigma (possibly much bigger than 1).
Bob receives sum of original signal and the noise.

What is the maximum amount of information to be transferred if such experiments are repeated?

What is the best encoding scheme (I guess that for high noise 0/1 in some error correction sequence) ?