Shannon-Fano

From section 5.3, we have \(l_i^*=-\log_D p_i\), but it may not be integer, and we should choose \(l_i\) close to \(l_i^*\). So round it up using the ......