Clipping is generally something considered to be a fault. When the red lights come on, you bring down the gain or level to prevent clipping.
But it is a fact established by the BBC that the ear can tolerate brief periods of clipping perfectly happily. This is the basis of the analog PPM meter (Peak Program Meter), the needle of which has a rise time of 4 milliseconds. This period was set because few people can detect periods of clipping shorter than this.
This leads to the question of how noticeable a longer period of clipping would be. What about 8 ms? What about 15 ms? The longer the duration of clipping, the more noticeable it will be.
There is therefore a degree of subjectivity in this. The engineer can raise levels into clipping and the sound will change from undistorted, to unnoticeably distorted, to perceptibly distorted to grossly distorted.
Somewhere in this range there will be an optimum where the subjective level is maximized but the distortion is still at an acceptable level.
Still on the topic of clipping, it is essential to compress first, and then clip. Otherwise very few peaks will be clipped, and hardly any additional subjective level will be achieved.
Compressing first can produce a signal that is close to 0 dBFS almost all of the time. Clipping will then improve the subjective level of the whole of the signal.
It is also worth saying that different pieces of equipment produce different sound qualities when clipping occurs.
Often it simply sounds bad and cannot be used. In analog tape, the 'rounding' of peaks can be very pleasant. Other equipment will fall in-between these two ends of the spectrum of clipping subjectivity.