This surprising similarity prompted an important clarification: the test didn’t rely on exotic or high-end gear. It used ordinary consumer gear like a laptop, a USB audio interface, and standard recording software.

Well this could be part of the reason other than the banana ,etc, just adding resistance.

But you’d have to vet your random participants. Like for instance my wife’s phone speaker is terrible and she turns it up to 100% to watch videos. She doesn’t notice anything wrong, but for me the distortion and clipping hurts my head. Some people just dont “hear” bad audio.