Performance above 20Khz is frequency neglected with the assumption that people would not be able to hear any difference. It is often assumed that sounds that by themselves cannot be perceived by humans will remain inperceptable when they are mixed with lower frequency audible components but is this really true?
It is worth noting that humans can actually directly hear single tones as high as 28Khz but that’s at over 100 dB (101.3 to 111 dB for 3 of 32 test subjects) which is much louder than the ultrasonic components in typical music (not even close).
Multiple studies have indicated that ultrasonics that by themselves are not audible (and doesn’t seem to be perceptable in other ways) can still make a difference with music likely due to these higher frequencies interacting with lower frequencies inside the human body.
https://journals.physiology.org/doi/full/10.1152/jn.2000.83.6.3548
https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0095464
https://www.frontiersin.org/articles/10.3389/fpsyg.2017.00093/full
But when people are asked to guess which version that contained high frequency audio they typically struggle, this remains true even for experienced professionals.
https://www.researchgate.net/publication/257068631_Sampling_Rate_Discrimination_441_kHz_vs_882_kHz
When we design our speakers we will focus heavily in the 100hz to 10Khz range since that is very much audible but there isn’t much of a reason not to choose a tweeter able to reach well above 20Khz with decent performance.
Tweeters like all speaker elements have a range where they perform the best and when that range reach 18Khz+ the performance will typically be acceptable well above that.