This is exactly my experience. In an ideal recording, for example recording a single source of sound in an anechoic chamber, or playing test tones as Rob mentioned earlier, results in "pinpoint imaging", I've also heard it called a "Spotlit Presentation". A very resolving and well set-up system can create this "spotlit presentation" to too high of a degree, with the main complaint being the images sound too small vs real life.
I've found it's possible to effect image size in various ways... amplification devices such as vacuum tubes, some class-A SS devices seem to manipulate perception of space and image size. Interconnect cables can make massive differences, believe it or not... a very accurate cable like a UPOCC silver (round wire) IC cable will have little effect. However, switch the wire shape to foil (width>>thickness) and you can get larger than life images that are too diffuse and run into each other, so there's no "air" (space in between images). A ribbon cable (width>thickness) can split the difference and make for larger images vs round wire cables, but not so much that they are larger than life. There seems to be little difference in resolution between these types of cables so fine detail may be preserved, but foil is doing something to the sound to "unfocus" it. I think it has some effect on phase that is very subtle, but has an effect on our psychoacoustic perception of image size.
Ribbon shaped wire has a lot of other, even larger differences vs round wire such as capacitance and inductance so I wouldn't say this image size difference is a main difference, it can be subtle unless you have a certain kind of system. Also, many electronics have enough character to them that they define the presentation more than the cables. So like anything YMMV as far as experiencing this phenomenon in your own system, I've found it more with hard-cone driver/SS amp systems.
My own system uses tubes, so cables don't effect image size nearly as much as many other systems, but I do find SET amps w/ efficient drivers and little to no crossover have other qualities that are hard to beat, so that's what I like. I do think tube based systems can present space in a pleasant and natural way that doesn't interfere with resolution much, no PPI, no 12 ft tall woman singing in front of you.
Ideal recording? Anechoic chamber, test tones...
What does that have to do with real music recorded in real space?
Sorry, I'm not buying it.