If we are talking about transformers for SPDIF.........
Their coupling is too low. They, along with the idiots at Crystal, seem to have this misguided notion that you need low coupling to prevent noise from getting coupled in.
Too bad that also kills the HF content of the signal, and that is essential to get a good, clean edge.
The other problem........low coupling means low capacitance. Which means high leakage inductance. It is hard to get rid of the effect that has on the signal. Someone like me can design networks to help compensate for it, but you guys don't have all the fancy test equipment to allow you to try.
Bottom line..........go with one that has real tight coupling. The old Schotts...........the cheap imitation one by Pulse Engineering, or the ones Newava makes. Newava may be the same thing, made by ex-employees at some new, nearby location after Schott went belly-up. They sure look like it.
BTW......Crystal are the same idiots that design receiver chips with Schmitt trigger inputs. Another rant for another time. Bottom line is: they do all the worst things that one can do to a SPDIF signal, and have no clue how wrong they are.
Pat