The difference between the first raw bias and the second raw bias (in ADU) shifted by 10x10 pixels, is computed. This is the combination of fixed pattern noise and read out noise (scaled by sqrt(2)). The population standard deviation of this difference is computed and divided by the square root of 2. The read out noise contribution evaluated as QC OUTi RON is then quadratically subtracted from the total noise.
Details
can be found in the <a href="http://www.eso.org/observing/dfo/quality/FORS2/qc/bias_qc1.html">plot tutorial</a>.

</font><p><font size="2">
<i>DB Name: fors2_bias <br>
TPL.ID: img_cal_bias <br>
Fits header keyword: QC.BIAS.FPN <br>
QC1 DB parameter: FP_r</i>