Hello I just want to know if the computations I do to know the relation Energy_per_Bit/No are correct in the next situation: I add white noise after performing the FFT in a OFDM receiver as shown next: Signal_Power=var(Noiseless_Signal_in_Frequency); VarianceNoise=SNR/Signal_Power; Noisy_Signal=Noiseless_Signal_in_Frequency+sqrt()*(randn()+1i*randn())...; Is the Energy_per_Bit/No relation computed as: Energy_bit=(Signal_Power/Number_Subcarriers)*1/(OFDM_Symbol_Duration*Number_Subcarriers)*1/log2(4) (QPSK) No=NoiseVariance/Signal_Bandwidth ?

# Eb/N0 computation when adding noise in frequency domain

Started by ●April 7, 2009

Reply by ●April 10, 20092009-04-10

On Apr 7, 6:33 pm, "Crisanquito" <crisanc...@hotmail.com> wrote:> Hello > > I just want to know if the computations I do to know the relation > Energy_per_Bit/No are correct in the next situation: I add white noise > after performing the FFT in a OFDM receiver as shown next: > > Signal_Power=var(Noiseless_Signal_in_Frequency); > VarianceNoise=SNR/Signal_Power; > Noisy_Signal=Noiseless_Signal_in_Frequency+sqrt()*(randn()+1i*randn())...; > > Is the Energy_per_Bit/No relation computed as: > > Energy_bit=(Signal_Power/Number_Subcarriers)*1/(OFDM_Symbol_Duration*Number_Subcarriers)*1/log2(4) > (QPSK) > > No=NoiseVariance/Signal_Bandwidth > > ?It looks like Signal_Power is per OFDM carrier and per OFDM symbol. In that case, Energy_per_bit = (Signal_Power*Number_Subcarriers)*1/ (OFDM_Symbol_Duration*Number_Subcarriers*2) You have a typo above in the Noisy_Signal formula, but I think you mean: Noisy_Signal=Noiseless_Signal_in_Frequency+sqrt(NoiseVariance/2)*(randn ()+j*randn()) where NoiseVariance=N0*Signal_Bandwidth Otherwise, your formulae are correct. Thanks, Dilip.