Looking at the design of the pipe being used as a radiator shell, would some of these issues be solved / reduced with a really stiff shielding plate forming a 90 degree arc around the rear face of the pipes or a specifically shaped shell to replace the pipes ? ?
I would think that would mess with the radiation too much to be helpful (unless it could somehow attenuate all frequencies equally and delay all rear reflections by 6ms).
Nothing I said about setup was an issue or problem, but more so just how I like to set them up for best results (in my non scientific opinion of course).
... replace those PVC pipes with 5mm thick steel pipes of the same internal diameter. Would the speakers sound the same after that ?
Yes
The pipe came from both being strong enough to stop box resonances and to have no baffle reflections and measure flat in the listening plane:
(The driver doesn't vibrate the pipe like a box, only creates an internal standing wave which is dealt with by first stuffing the pipe, then by measuring the wave and equalizing for it in the active or digital eq/xo)
IMO I would contend that all speakers, outside of ones like a Klipsch cornerhorn (uses the walls as a horn extension) or DSL synergy horn (broadband, sharply controlled directivity) would benefit with some space off of the sidewalls.
SL was just very straightforward to ideal placement and why. There's a wealth of information on his site:
http://www.linkwitzlab.com/index.htm going into all kinds of issues with speaker design, radiation, room response, hearing, and psychoacoustics etc, which I won't pretend to have a good enough understanding on to be able to speak at depth about. I can however copy and paste from his website
:
The above proposed/ideal setup is based on each speakers radiation pattern (dipole=lx521, monopole=lxMini) as it related to the listening position to get an ideal direct/reverb ratio:
So the LxMini would use the red line for listening distance vs the listening room's measured RT60. lt's interesting SL specified RT60/placement specs just like the Genelec and Nuemann figures someone posted in another thread; however he was a proponent of, and designed the speakers and setup for, a much more easy to achieve (IMO) RT60 of ~400-450ms. The 200ms that Nuemann (I think?) specs for listening distances is much more directed towards a dedicated studio with lots of absorption and treatment IMO, so the listening distances for each speaker listed would need to be brought in a bit to achieve the same direct/reverb sound ratio in normal listening room.
I apologize for the long post and all the copypasta; I can't say I have enough knowledge or experience do it any other way. Others here can expand these ideas/approaches and the trade off's way better than I can.