It isn’t even the same for all of the shots. Some have smaller but still statistically improbable blank spaces. Some are just notably sparser but no huge gaps.
My guess as to what happened here: They first cut out all of the “real” shots with slightly different timing. This also cut out the noise. They then added the pristine shot waveform they wanted to use (eight times) at the spacings they wanted. Then they added some noise to mask the process. But the noise wasn’t the same as the real cabin noise. And in some places, they didn’t get it everywhere they needed it. Not easy to see if you are looking at the whole composite signal, so out the door it went.