Please explain the Es/No definition

[davyzhu]

Es/No problem?

Hi all,

In Error correcting code, code_bits={data_bits+parity_bits}.

I am confused with Es/No definition:
Es/No = Eb/No*data_bits_per_frame/symbols_per_frame;

Why not:
Es/No = Eb/No*code_bits_per_frame/symbols_per_frame;

Best regards,
Davy

 

[rostam63]

Re: Es/No problem?

you should explain more and clarify what your question is to be able to answer.

 

[Sal]

Es/No problem?

Hi,

It is because the first definition doesnt include the error correcting codes, I mean you can apply different error correcting techniques and for each one the error rate may be different.

Sal

 

[davyzhu]

Es/No problem?

Cited from comp.dsp
Because Eb/No is (usually) defined in terms of the energy per *data* bit.

So for an (n,k) code:

Ebc = Ebi * (k / n)

where Ebi is the energy per data bit, and Ebc is the energy per code bit.
And for an M-ary symbol constellation:

Es = Ebc * log2(M)

Therefore:

Es = Ebi * (k / n) * log2(M)

But (k / n) * log2(M) is the number of data bits per symbol. Therefore:

Es/No = Ebi/No * data bits per symbol