Exactly my point.
Any longer than the maximum slew rate is only needed to make sure that any starting transitions don't peak high before things settle. My own driver rises on an exponential slope at startup, and settles within about 10 microseconds, and the slope to full current rises to settle about 20 microseconds later. This is faster than 30ms by a thousandfold, but it is good enough. I can repeatedly jam splayed wire ends on a 12V battery drawing rough sparks if I want, the transitions never hit the laser diode. So I agree, any longer than 30ms is luxury.
Swami, I think the ambient light ought to be enough to solve the first point, if it doesn't the projector is likely too bright for the conditions. Though I accept that we can't easily scale this and not at all fluently, a diode is what it is... We can scale other lights though, set a mimimum so we don't perceive the annoyance in context.
The timing thing is more interesting, I'll accept this if there is no way to avoid it. Thing is, DPSS does have nastiness, direct diodes not so much, I thought. What you're saying seems to imply that they have much slower responses than I thought. Hundreds of microseconds, where I thought they'd be a thousand times faster. I've not seen this because I have never modulated more than one colour. If the degree below threshold can influence the timing of fast modulation accuracy for colour mixing at low levels, then it's worth using to do just that. But again, can this not be done purely by setting that sub-threshold value and accepting that if very low, there may be slight impact on diode life? I still think it's worth accepting the slight low level light because in proportion to full it will be very weak, and as with audio the best way to hide low level aberrations is in a 'noise floor' that our brains will tune out.