I don't think that most studios target for a non neutral direct sound during mixing.There are many views on this. In short, music is usually mastered for the least common denominator. The more popular the music, the lower the denominator becomes. Nowadays many engineers balance with mixcubes
https://www.soundonsound.com/reviews/avantone-active-mixcube
There are only a handful of studios that even attempt to do 'neutral' - they tend to produce with a full 24bit/96kHz chain and the albums lack compression. Problem is they don't sound like much on the average car radio.
For the rest, what is the point of neutrality after applying compression and filtering?
Also, what is neutrality if half of the sound is generated electronically and mastered to sound good on laptop speakers and Beats headphone?
Sure you can agree with the production engineer but the next remaster will make drastically different choices ... and why not?
Just my 2c.
The use of mixcube or so is to ensure the music content don't too bias to the HF or Bass for it to sound nice, just making it sound ok when a consumer mainstream speaker say phone speaker still sound nice, mostly be focusing more on the midrange and the bass or so is for extra enjoyment when the bass or treble response is there. and noticeably usually they roll off the extreme highs where bad speakers usually have nasty peaks there.
As the non-neutral speakers' response are all over the place it's virtually impossible to target a response to sound best on "general bad speakers", and if so, hifi manufacturers would go bankrupt as then the $20 flee market speaker would actually be highest fidelity as most studios tuned for it to "sound right" where the hifi speakers would be dark, boomy, lack clarity... I would agree studios do those car checks or so, but usually that is to highlight those midrange problems which would be masked by nice highs and bass in a really good speaker, or target in taming the annoying high frequency peaks if it were there in some bad speakers to make it less intruding.