I think I might have figured out why he's calling it FM.
When I was working with HV at my last job, I pulled apart an electronic NST. I ripped out the output transformer. It had a primary wound with magnet wire, a continuous ceramic core, and a potted secondary. It was a resonant transformer, which I thought was odd. When driven at 30kHz it would massively step up the voltage on the secondary to 10kVpk-pk. If you shifted down to 25kHz or up to 35kHz, it would drop to <1kVpk-pk.
If you were to take an audio input and use it to modulate the frequency of a sine wave so that it fluctuated between 25 and 30kHz, you could achieve an amplitude modulation of the spark voltage on the secondary. The linearity would depend on the linearity of the transfer function of output voltage as a function of frequency, but there might be ways to avoid or compensate for that.
Does that sound like a possibility AG? I'm tempted to go whip one up. I've been meaning to do a plasma speaker for a while, and this I could definitely whip something like this up. I've got a Tektronix rep swinging by next week to show me their new DPO3000 scope and I could ask for a quote on one of their HV probes.