Impulse response calculation from frequency domain complex data

Hi,

I have some channel measurement data obtained in the frequency domain using vector network analyzer and I need to obtain the time domain impulse response. The simplest thing that I tried is to take IFFT from my data (X), also 1/sqrt(N) normalization is considered. But the result plot of abs(x) exceeds 0 dB which menas my channel introduces gain! and it is not correct. Am I simplifying the procedure too much? or do I need to use more accurate estimation algorithm?

Thanks

[Nagarjuna]

how can u say that channel introduces gain when ifft is done ?

[seyran khademi]

I plot the abs of time domain values obtained by taking ifft from my data ; plot(abs(ifft(X))), and the peak exceeds the zero dB value. Isn't that a correct interpretation?