Which component is most responsible?

Awhile back I started a similar thread. For me, the holy grail was a clear spatial separation of the first and second violins in string quartet recordings. My premise was that the problem was inherent to my speakers (magnepan 3.7Rs) due to the side by side arrangement of the ribbon tweeter and the quasi-ribbon midrange.
I got a variety of suggestions from power treatment to other speakers, even headphones. In the end, the solution was realized by a preamp upgrade. You may find that you will find the biggest improvement by replacing the weakest link. In the end, every component in the chain has to be up to the task. Speakers can't recreate a spacial relationship that your CD player smeared, but the reverse argument is equally true.

This has to come from the source or else nothing downstream can put back what was lost. Massed chorus music and complex full orchestra passages are extremely challenging to the entire audio chain so that it all needs to be right or else muddiness results... much easier to get distinctness with a few "voice" lines than many competing voice lines

Which component(s) is most responsible

Components is correct, all of them. As Davide256 said, "This has to come from the source", then the preamp/amp must do its job to drive the speaker with this quality created by the source, and the speaker must be able to reproduce this quality. Just having all high quality components/speakers will not assure that you will achieve this quality. Synergy between components/speakers is extremely important as well.

It's how it was recorded. Either the recording has or it doesn't. If it does, even a moderate quality system (with attention paid to setup) will reveal this effect.