It is smoothing yes, but it is different than logarithmic frequency scale smoothing, such as 1/3 octave or similar fractional octave smoothing.
Smoothing result from gating is on the linear frequency scale, this is the difference from more common logratihmic smoothing. And this combined with resonance Q's bandwith's being dependent on the resonance's frequency causes gating to obscure the low frequency data more than the high frequency, as long as you also assume that along the frequency response any resonance is likely to have same average Q value.
With the software you are using, is it possible to make the frequency response to be drawn on linear frequency scale instead of the common logarithmic scale? If you look at the frequency response curve with linear frequency scale, you will see that the low frequency section is squished, and high frequency section is widened; compared to the log frequency scale curve. The lower frequency the more squished, the higher frequency the more widened it will be.
So if you take any of your measurements and make them drawn on linear frequency scale, and try to imagine a smoothing done on this scale on the curve, you will notice that this smoothing will make more changes on the lower frequency than higher frequency. Because higher frequency is already much more smooth than the lower frequency.
This is what you see when you are saying changing the gate is not touching the high frequency region. It is not touching because those areas are already smooth enough on the linear frequency scale, and gate is a linear frequency scale smoothing operation.
Consider this, a peak at 100Hz with a Q of 1 will have a bandwidth of 100Hz. Similar peak at 1Khz with Q of 1 will have a bandwith of 1Khz. They both will have same peak to tail value, of 3 db. Now if you imagine the FR curve on linear frequency scale, the peak at 1Khz wil be 10 times wider than the peak at 100Hz, which makes it 10 times smoother already. But if you look at logarithmic frequency scale both peak appear to have same width and so same smoothness.