Here’s a comparison with your RED footage on the left, a render matching the lighting conditions in the middle, and your render on the right. These are all at the high exposure of 5, viewed through the ACES 1.1 RRT. As you mentioned your render is at 15 on the red spheres and around 3-5 on the ground.
I believe the reason the other two look so different is that on both the bulbs are over 100 and the floor is around 0.2 meaning the difference in brightness between the bulb and floor is much greater.
Here are the images at exposure 1. Note how the floor is nearly black on the left and middle and middle grey on the right
Here are the images at exposure 1, through the OpenDRT (v0.75)
and through OpenDRT at exposure 5
FWIW, the red light in my render is at 100% saturation (1,0,0) in ACEScg primaries.
Finally, here are a comparison of OpenDRT 75, 80 and 81b. Note the red bulbs go to magenta. The difference between 80 and 81 appears to be due to the lack of gamut compression, that is, when I add a gamut compression node above the v81b DRT in Nuke it looks virtually identical to v80.
Here’s a comparison of spheres with ACEScg primaries at high exposure.
All have a path to white, which is important, but the v81b (and v80) retains its colors more in comparison to v75. I think this is a plus and this was the reason I was bringing up the issue of saturation in CG colors earlier. Most things in the real world are not at 100% saturation, but in CG it’s easy to pick that color in the color picker creating an unreal/surreal image. There may be reason to discourage artists from doing that, but I’d propose that the DRT is not the place to do that, and that instead the DRT should not limit what is possible. That’s why I’m thinking the direction of v80 and 81 is an improvement.
FWIW, we have a matrix added to the OCIO color picker role that keeps all colors (in ACEScg primaries) under 0.98 saturation, which artists can easily override in Maya if desired.