BER to Q
Q=(2^0.5)*erfcinv(2*BER)
Q_dB=20*log10(Q)1. Why do you want to turn like this:
I don't know yet, I'll make up when I know
2. About erfcinv:
y=erf(x) error function
y=erfc(x) complementary error function
y=erfinv(x) inverse error function (inverse of error function)
y=erfcinv(x) inverse complementary error function (inverse of complementary error function )
definition
3. Combining the two courses I have attended, "Communication Principles" and "Signal Detection and Estimation", let me talk about some understanding:
In the course "pre-estimation in signal detection", the Q value is defined as 
where I1 and σ1 are the mean and variance of the 1 Gaussian pulse output, and I0 and σ0 are the mean and variance of the 0 Gaussian pulse output, respectively.
The relationship between Q value and BER is: 
there is a coefficient difference between this expression and the code
Q (dB) is defined as: 
transfer from: qiuyiL: BER to Q