ty ranfan for finally commenting on the article that has some measurements!
However small the measurements showed the changes to be (and maybe the manufacturer sent already burned in headphones as the sample, let's assume they didn't) there were still changes. Do we have measurements on what the limit of a trained listener can detect before we toss the findings of measured change aside as non-detectable to a trained listener?
The example of the listening device is, imho, also not proof that burn-in is pointless -- some devices are designed to not need burn in while others are. The idea that burn in must cause continuous changes over time isn't even the hypothesis here. The idea that burn in causes the device to better show its true colors is the hypothesis. And it could be you just want burn-in so a device that does have a weak component will break within the warranty period.
Are we tossing a mass of people's perceptions aside because we haven't bothered to measure properly?
I've found burn-in changed one set of headphones while it didn't change multiple sets of speakers, headphones, or IEMs I've owned -- surprised me on the one set that did change. However that same set of headphones also had a sticky driver even after the burn-in period; the right speaker would slowly have less decibels then stop emitting sound altogether sometimes, but giving it a huge amount of volume would make the speaker come alive again for a time. Eventually that right driver worked 100% of the time -- was that burn-in? It was definitely a mechanical issue.
Is that one example, anecdotal as it is, enough to say that
sometimes burn-in helps? And maybe it is even recommended just to get the 'dust' out?