update phi problem

[zhi93]

        phi_noise(t, f) = trace(corre * R_noise_inv) / num_channels;
        phi_noisy(t, f) = trace(corre * R_noisy_inv) / num_channels;
        
        % update lambda
        k_noise = obs' * (R_noise_inv / phi_noise(t, f)) * obs / 2;          
        k_noisy = obs' * (R_noisy_inv / phi_noisy(t, f)) * obs / 2;

update phi , k_noise and k_noisey allways be 2

[funcwj]

Yes, the computation is duplicated, k_{n,s} will always equal to number of the channels