Hi, I have an assignment and I need some help.

The question follows: "Construct a funcion [quantized_signal,var_lin,var_sat,SNqR]= My_quantizer(unquantized_signal,V_p,N), that uses an 2^N quantization levels, equally spaced on [−Vp, Vp], to quantize a signal. The function should return a vector containing the quantized levels of the input signal, empirical estimations of the variance of the linear as well as of the saturated errors"

Does anyone have any suggestions or solutions on how I should approach this question or does anyone have any idea for a code for it?

[Geoff Hayes]

Simon - since this is homework, we can only give hints. What have you tried so far? From your class notes, how are the "empirical estimations of the variance of the linear" and the "saturated errors" defined? For creating the equally spaced quantization levels, see linspace.

[Simon Ekström]

I can handle the empirical estimations, but I am mainly wondering how I should define the quantization levels? Should I use something similar to: quantized_signal = linspace(-V_p, V_p, 2^N);?

[Geoff Hayes]

yes, that seems like a good start.

[Simon Ekström]

lets say for example, we have a sampled signal s = [-4 1 5] that should be quantized with 2 bits and with Vp = 4 so that the the quantization levels should be [−3, −1, 1, 3], how should I modify my code so this becomes the result?

[Guillaume]

Have a look at the discretize function which will give you the indices into your quantization vector.