I notice something that I think is unique, but don't quite understand. Follow me for a second. Heres the scenerio. Lets say the track you're working on is peaking at -0.5dB or something. The Input Scale knob on the Adaptive Limiter is at 0.0dB, and Output Ceiling knob is at 0.0dB. The middle knob on the adaptive limiter is the gain knob with a range from 0 to 12dB. As you raise the dB on the gain knob, the margin meters on the right get less negative, so I thought. Let's say the gain knob is at 3dB. The margin meter might read -2.4dB. Then when the gain knob is at 6dB, the margin meter might read -0.5dB. I do know that with this scenerio, when the gain knob is at 8.0dB, the margin meter reads 0.0dB. However, the gain knob goes up to 12.0dB, so when it is at 12.0db, strangely enough, the margin meter somehow gets more negative and reads at -0.8dB. Why is that? You can obviously tell by looking at the master fader that the output ceiling has been lowered to a reading of -0.8dB when it is still set to 0.0dB on the Adaptive Limiter. If you want to recreate this scenerio, please remember that when you take the gain knob past 8dB, remember to click the number on top of the margin meters on the right to reset them, because the number has gotten more negative. If someone can explain this crazy thing, please do.

Haha, no replies in 6 years. Did you ever find an explanation for this? Or better yet, a way to fix this? Cuz now I've encountered the same anomaly and it's pissing me off. I need the extra dB of gain for perceived loudness. I too find that 8dB on the gain knob results in the highest output. I cannot logically figure out why this would be. I've googled and googled, read all the relevant manuals, but can't find any explanation. Very strange. A glitch in the algorithm? It has just occurred to me that I could put 2 adaptive limiters in a row, each set at or below the 8dB weirdness threshold. Tried it. It works. Alternatively, since the adaptive limiters are on the individual regions, I could probably achieve the same results by putting another on the whole mix, rather than a 2nd one on each region. Yup, this works too. I shouldn't need to do this though. Grrr.