I tried microwave oven transformer first as a transformer, but it went up only to 1 kHz (probably because of core material selected -- it's normally used at 50...60 Hz), which would be quite a low frequency for efficient corona treatment. That's why I had to make my own with ferrite core and double, layered windings.
CRT TV Flyback transformers would go to high frequencies, but they have the rectifying diode incorporated that cannot be removed. What I need is AC. All my DC experiments, up to 30 kV, failed and the commercial unit we measured gives AC.
I think that uneveness is easily addressed by doing many passes "just in case", which is not possible or easy in fast, continuous web machinery. Ozone indeed does smell, but for small scale treatment probably does not matter.