@simonmoon , It is not about the recording method Simon, it is about the playback. The symphony orchestra is so wide and so deep. At a live concert the sound of each instrument appears to comes directly from that instrument. The sound of instruments does not come from beyond the stage. In a full concert hall there is extremely effective sound absorption in the form of human beings. The reflections you hear are extremely late and relatively weak due to the distances. Up front in row 6 center- left where my dad's season tickets were the image of each instrument was very specific. I could easily close my eyes and point to each instrument.
https://en.wikipedia.org/wiki/Symphony_Hall,_Boston#/media/File:Symphony_hall_boston.jpg. The orchestra has a distinct size and in row six it is really large. On playback you are expecting two speakers to recreate the orchestra's size and the position of each instrument. Since you do not have an individual speaker for each instrument the position of each instrument is determined predominantly by the relative volume of the that sound coming from the two speakers. Phase and timing assists but only if you sit dead center otherwise that information is corrupted. The full size of the orchestra will be determined by the speaker separation and how close in you are sitting. If all early reflections are well dampened it is physically impossible for sound to come from beyond the speaker without studio tricks. Not only that but sound coming from outside the orchestra would be anything but realistic. What you are hearing when you hear imaging beyond the speakers is poor control of room acoustics. How the orchestra was recorded might affect the quality of the orchestra's timbre and image but unless studio tricks are applied it should not produce sound outside the limits of the orchestra which would be surrealistic anyway. My experience with systems over the years bears this out. This type of thing may sound cool but it is not accurate. On the bright side you have identified a problem to conquer and once you do you should have a much tidier image. I have many recordings of Seiji Ozawa at Boston Symphony Hall and most of them are surprisingly accurate. None of them cast an image beyond the speakers and in my front row it is about the same as being in row 6 at the hall, wonderful.
https://en.wikipedia.org/wiki/Symphony_Hall,_Boston#/media/File:Symphony_hall_boston.jpg. The orchestra has a distinct size and in row six it is really large. On playback you are expecting two speakers to recreate the orchestra's size and the position of each instrument. Since you do not have an individual speaker for each instrument the position of each instrument is determined predominantly by the relative volume of the that sound coming from the two speakers. Phase and timing assists but only if you sit dead center otherwise that information is corrupted. The full size of the orchestra will be determined by the speaker separation and how close in you are sitting. If all early reflections are well dampened it is physically impossible for sound to come from beyond the speaker without studio tricks. Not only that but sound coming from outside the orchestra would be anything but realistic. What you are hearing when you hear imaging beyond the speakers is poor control of room acoustics. How the orchestra was recorded might affect the quality of the orchestra's timbre and image but unless studio tricks are applied it should not produce sound outside the limits of the orchestra which would be surrealistic anyway. My experience with systems over the years bears this out. This type of thing may sound cool but it is not accurate. On the bright side you have identified a problem to conquer and once you do you should have a much tidier image. I have many recordings of Seiji Ozawa at Boston Symphony Hall and most of them are surprisingly accurate. None of them cast an image beyond the speakers and in my front row it is about the same as being in row 6 at the hall, wonderful.