Speaker Imaging - Do you hear a line, or do you hear an arc??

with a lot of recordings … the vocals are placed further back than the instruments

On pop records this is probably incorrect except for solo guitar, sax, etc.
On live opera, correct.

A fair number of LPs/CDs contain a photo of the group/orchestra playing while being recorded.
When listening to the recording, have you been able to correlated what you hear with what you see?

Only valid if the recording is mic'd stereo. If the session has several microphones, all bets are off there is any correlation, but some multi-mic sessions could be correctly panned and mixed to reflect player positions.

Mic-per-chair orchestra recordings don't usually image well and suffer the additional defect of 'strolling' wherein the perspective changes when a player or section is featured.

While I agree with the room and speaker placement within it doesn't explain the differences in imaging from different speakers placed in the same place. The revel salon 2's were the most particular in my room, imaging was always recording dependent, but the soundstage was elusive with them. the big ushers imaged wide and deep but very little height. the revel f52 were the most intimate imaging speakers i've ever had in this room placing singers and instruments solidly in space while capturing panning effects intensely. There's been several quality speakers down here with their own personality,   the current residents manage to erase the walls and put me in the middle of the performance.  Roger waters, amused to death for example the phase thing had different results with each speaker. A speakers ability to capture my imagination and allow me to create the image can hide many other shortcomings and I don't know if there's a measurement for it yet.

The basics are pretty simple, in my experience and based on audio science.  I've been fortunate to be able to experiment with a wonderful array of spaces:

*  Sidewalls -  the closer they are to the speakers the more your image is anchored on the speakers while forming a "U" between the speakers (speaker output is reinforced with short delay reflections).  For best imaging, you should for all intents and purposes have no side walls within many, many feet of the speakers.

*  Wall behind speakers:  The further away, the more "depth" you will perceive when it is recorded on the recording.  The closer behind, the less depth.  (Of course, you have also to pay attention to room nodes for accurate bass reproduction.)

*  Prevention of center "cross-contamination".  Experiment with center reflections from the wall behind the speaker.  To experiment, put a six foot high stack of empty boxes with one corner ("v") centered and facing out behind the speakers.  This breaks up reflected diffusion from between the speakers.  In a real room, this can be approximated many ways, including a "v" shaped two panel screen if needed.

This is the simplified basic.  It can be approximated different ways with room treatments.  If you are lucky, you can even arrange a living room this way.