I’ve run the same procedure (multilooking + terrain correction + calibration) on the same scene in both v1 and v2 toolboxes. The output is significantly different. I’ve noticed before that the exact output values are very sensitive to, for example, the choice of the projection during terrain correction, so if there is even a minor difference in the way terrain correction is implemented in v2, I’m not surprised. Still, it would be interesting to know if the difference is expected.
The difference is in the resampling methods whether the value of the pixel is in the corner or in the middle of the pixel. We still need to make this consistent.
BTW, I found that after 3-pixel multilooking terrain correction computes pixel size for IW_GRDH to be 30.02 meters, not 30.03 meters as before. Using these two numbers with the same toolbox binary produces somewhat different results as well.