I don't think there's a generic answer to this question, regardless of whether you use a transformer or a coupling cap. In both cases, you have to design your DC isolation carefully considering the exact circuit that you're attaching it to, also with a mind of what you're trying to test and what signals you expect to use.
This week, I'm using a home-built isolator box to test biased electrolytics for impedance and distortion. I drive the cap from my generator through a 100Ω resistor, so I need low impedance coupling. I decided to use 48x 10µF metallized polypropylenes since they are very linear and they can handle 200V of either polarity. The DC bias I'm applying to the cap under test gets filtered from the PSU to the cap, but that filter also has to include the 480µF polypropylene cap, since it too gets charged up by the bias voltage. The constraint on this DC bias voltage filter is thus the settling time of the entire system: the cap under test, the DC coupling cap, and the DC bias voltage RC filter.
The whole system works well, but it's a very thoroughly engineered complete system - there's little generic about it, and probably little other than the basic concept that would apply to other applications.
For simplicity, I'd suggest a simple coupling cap, preferably film if you can get one large enough for your use, and do some calculations to see how large it needs to be to get the required frequency response in the circuit you want to test. Thanks to solar power inverters, you can get some enormous polypropylenes now, but you probably don't need hundreds of µF (or... maybe you do?) The circuit details matter here: a U87 has some extremely high impedance nodes and some very low impedance nodes - it all depends on what you're testing, where you're injecting your test signal, and the series impedance you're using to inject the signal.