Hi
The issue is complicated. Speakers have a very hard job to do: Reproducing as faithfully the electrical signical that is fed to them. They mange to get some to a point where we can recognize what comes out but their degree of faithfulness to the signal is abysmal... Yet we make something out of it to the point that it can be preferred to the real live sound.
I think the issue is more complex than this. What we want the speakers to do is create a sonic experience that communicates the artistic intent of the recording artists. (Take this to mean the musicians, composers, arranger, engineers, producers, mastering engineers, recordings are a team effort usually.)
An electric signal only has a relationship to a sonic experience if it has been crafted to be such. Overtime mics have been developed that capture sound pretty faithfully, and if that original signal is captured, and input to an amplifier, which is connected to speakers, then some representaion of the original acoustic event does, miraculously, emanate from the speakers.
But without considerable effort, this acoustic representation will not be interesting. The "producer" of a recording is responsible for creating a signal that will communicate the musical expression with conviction.
To craft this expression, in all but the most purist recordings, the various captured signals are worked on in a studio, while listening to monitors, in the hope of creating a compelling product. This type of communication is complex, and the "accuracy " of the speakers is only one component of the process.
I record music, and one of the vexing challenges of this process is creating a recording that "translates" well to all the different playback systems that might be used to express the recording. What I usually do is rely one main set of monitors. But I will switch between various sets of speakers, as well as check mixes in other playback environments, in the hope of creating an expression which transcends the specific playback system a listener might choose. It can be shocking how different mixes sound on different systems.
There are specific qualities of a recording that help it "translate" like this, largely concerned with having a balanced dynamic spectrum, as well as overall well controlled dynamics. (I think of dynamic spectrum loosely as having a well-balanced frequency spectrum at the different dynamic levels in the signal. For example, if a recording has a ton of low-mid content on the quiet sections, and the loud sections have a lot of high-frequency content, this will be hard to represent on an average playback system.)
In any case, we all benefit because of the skill of producers/team in creating recordings that translate well across different playback systems, which means we can utilize "less than optimal" systems and still have a great listening experience.
I've been thinking about the recent discussion here about "evidence based speaker designs" a lot. The research being referenced (as evidence) points to speaker designs that in my experience I do not enjoy listening to. T
For studio monitors, I do prefer "accurate" speakers, just because it makes the project of crafting a signal that translates well more doable.
But for fun listening, I hate the sound of this type of speaker. I tend to prefer "boxy" designs, that have some cabinet resonance.
I have some provisional theories on what accounts for my subjective preference, but I'm not sure. One issue, is that speakers that are "colored" in a distinctive way can be more or less flattering to specific recordings.
I listen to mostly "rock music" (broadly defined) and I listen at low volumes. In my informed but limited experience, "accurate" speakers do not present this music very well.
It's making me inclined to do some informal blind listening, just to check what I might discover in that context.