It happens typically around 200Hz, depending on speaker height and listening distance.
Isn't that missing the point of the discussion so far? A reflection isn't represented by a frequency. A reflection is a delay, and can be represented by a
time.
On a
continuous tone, the reflection causes peaks and troughs in what a mic and laptop will pick up and analyse with an FFT - which is what you're talking about. But our hearing is tuned to detect the delayed reflection, and separate it from the direct sound based on the multiple cues that
transients provide.
This is why floor bounce, basically, is not audible, as has been mentioned earlier. It isn't because the brain expects it and is ignoring it; it is that the brain has removed it on-the-fly from its perception of the direct sound. There is literally no point in trying to remove it yourself electronically because, to the brain, it is not there in the first place (except as has been mentioned earlier as a sense of 'spaciousness' that is separate from the direct sound). If you do try to remove it by altering the signal, the brain will just hear an altered direct sound.
If the recording incorporates floor bounce (having been recorded using a mic in the audience for example), the brain can't separate it from the direct sound because the recording doesn't capture the complete sound field and allow the brain and ears to do their thing. In this case, the floor bounce will simply be heard as colouration - hence the reason for close-mic'ing of instruments and separate addition of 'ambience' in professional recordings.
Everyone surely knows that a tape recording of a school concert does not sound like it did when they were there in the audience. Now we know why