Ok looked at the Royer, it seems to me it gets out of control too easily wrt amplitude and frequency as it is resonance based, which is a problem for transformer saturation (runaway current -> mosfet kaboom) and output voltage stability (gate>Vgsmax -> mosfet kaboom) for the intended forward converter use
Maybe just a drive winding driven by a NPN transistor from a digital PWM output and transformer reset winding with a diode to ground...?
Bootstrap supply yes looked at it, but it kind of worries me it needs to be charged first, while I shouldn't have 100% duty cycle typically due to winding current control there may be times I need to for some period of time, then the caps would have to be big, then they would charge slower... complicates things. Thats why I wanted real gate voltage rails, to not have to worry about it.. maybe I'm just paranoid sorry

As for the tiny DCDC modules, yes I have been lazy and used something like this (but 15V) in the prototype, but they werent cheap like this