With all the conflicting information out there about annealing, I decided to experiment a bit.
I'd heard I needed a pan of water; that no, a air cooling was fine; to make the neck glow and to keep the neck from glowing and the flame from changing color, as either indicated a cooking-off of zinc from the copper (zinc melts at about 790°).
Structural changes begin in brass at 650°, so I used a TempilStik crayon that turns to liquid at that temperature.
I marked the case in three places: 1/4" below the shoulder, halfway down the case body, and at the head.
The TempilStik melted below the shoulder at 8 seconds. An additional 2 seconds did not melt the crayon halfway down the body, so it looks like for this particular setup, which is a 9/16" deep wall socket stuck in an electric screwdriver, 10 seconds per case, air cooled, gives adequate annealing without burning out the zinc.