Phototransistors would be fast enough for large slow bullets. I'm leaning to photodiodes which are faster and cost about the same.
If there is a single light source it would have to be set pretty far back or the detectors would need to extend beyond the window to have full coverage. The system will need to fit in one shooting lane (about 3-4') and I won't be able to leave it set up, so it needs to be semi-portable.
Single source would simplify the system. I've been assuming I would need multiple sources and switch them in sequence reading all the detectors in each state. I dismissed a single source early in my thought process, but I think I'll go back and look again.
It would be tremendously simpler if I could reliably trigger the the system and latch the detectors, which leads back to my original question about US sensors or other methods of detecting a the passage of a bullet.
Thanks for the ideas. If nothing else it makes me question some of my assumptions.