I agree with AntcuFaalb, you should capture YUY2 (4:2:2) - interlaced YV12 can be the source of all sorts of confusion.
Also if it's film-based material you are archiving then carrying out an IVTC and encoding as progressive YV12 would be more efficient.