My Lch and Rch are sample2sample coherent, the feature works correctly with "Default input" only(if you need, I can make a video), actually, it's ok to me, however, I spent 20 minutes to find out that way. I thought the REW should be multichannel to handle the Virtual balanced input feature...