next up previous
Next: Application to VDSL2 Up: Performance of optimal codes Previous: C program implementation


Word Error Rate versus Information Rate

We illustrate in this section the word error rate of optimal spherical codes on a Gaussian channel at finite length n=100, 500, 1000, and 2000.
Ten different values of the information rate $ R$ are considered, $ 1/10 \le R \le 4$ . The word error rate $ P_e$ versus the signal-to-noise ratio per bit $ E_b/N_0$ is given in Figures 1-4.

Figure 1: Word error rate versus signal-to-noise ratio for length n=100.
\begin{figure} \epsfxsize =14cm \centerline{\epsfbox{rot_Qtheta_large_n_anyR_100.eps}}\end{figure}




Figure 2: Word error rate versus signal-to-noise ratio for length n=500.
\begin{figure} \epsfxsize =14cm \centerline{\epsfbox{rot_Qtheta_large_n_anyR_500.eps}}\end{figure}




Figure 3: Word error rate versus signal-to-noise ratio for length n=1000.
\begin{figure} \epsfxsize =14cm \centerline{\epsfbox{rot_Qtheta_large_n_anyR_1000.eps}}\end{figure}




Figure 4: Word error rate versus signal-to-noise ratio for length n=2000.
\begin{figure} \epsfxsize =14cm \centerline{\epsfbox{rot_Qtheta_large_n_anyR_2000.eps}}\end{figure}


Joseph Jean Boutros 2006-11-11