pretty much what Eric [bdp24] said.
System: BiWired Spica TC-50 polyprop recap. Lead shot filled spiked stands. Speakers attach to stands w 4 Isolate It 50 Duro 0.25 x 2.25 circular pads. Kimber 2.5m BiFocal-XL. cd6006, passive LC1, passive XO1. 2x ML 800x w 10kg anti-skating.
Specifics:
Test discs never told me anything well recorded program couldn't tell better. Attend plenty of live acoustic music, get the image right for a well recorded large orchestra at full steam in a good hall and everything else is a doddle.
On a well set up system depth should approximate the space, be it hall or digital. It should draw you into the music as if you could walk up on stage.
An excellent set of a wide variety of well recorded material in good halls with a fairly consistent sonic is Louis Frémaux - The Complete Birmingham Years. Of course you have to like French music in English halls.
Note also, when a sound seems to be coming from farthest away, it’s pretty much always centrally located (center back).
IMO, this is a system failure. Back in the 90's, we did tests with 32bit digital processors, adding even or odd harmonics @ the CD 16bit bit 0 level. Even harmonics spread the backstage and push the image back, odd bring the image forward and triangulate it back to a point. The even push back can be solved by speaker placement, the odd triangulation cannot.
Other failings, mostly time coherence, cause the image to be too tall, tilted fore or aft.
IMO,
a great many systems AT ANY PRICE are incapable of any semblance of correct imaging because their time coherence is non-existent.