Help Scalar Vs. Vector Quantizations

[electronicman]

Hi,I want to know the exact diffrences between scalar and vector quantization,and how can I implement Scalar and vector quantization by using Matlab.I hope any one give me this with elaboration because I read Proakis's book ,Digital communications,But I have not understood anything at all.

 

[bagi]

in scalar quantization what is done is a single number is approximated generally using series of step functions....for explaining vector quantization i'll assume that the discrete signal is having various frequencies. so what we can do is instead of approximating the signal discreetly at each frequency we will club them into low, middle and high frequency groups. so we have 3 vectors, namely one each for low, middle and high frequency. now we use some quantization principle to approximate the vectors. i don't know how well i've explained the funda here, but i would recomment you read section 4.14 from the book 'data compression' by david salamon. that should clear all your doubts or else u can get bcak to me...in case i was helpful pls press the helped button

 

[mimomod]

In scalar quantization, you quantize each sample of a signal independently to other samples. In vector quantization, you collect several samples (for example N samples) and then assign a certain code to represent each combination of N samples. You may assign a short code to a combination which occurs often so you will compress the data significantly. You may assign a tolerance such that N samples data which is not too different to other N samples data will be assigned to one code, instead of two different codes.

best