Jonno said:
There are lots of variables here. First off, a mix that's been specifically designed for a 5.1 scheme and then presented in a discrete channel format (5.1 AC-3 or DTS) will inevitably be more precise than a matrixed 2.0 mix which only had four effective channels to begin with. So for most films post-1992, the 5.1 is the way to go; no argument there.
The reason these laserdisc projects are underway is chiefly one of preservation: that 5.1 DTS-HD mix on your Temple of Doom Blu-ray is not how the film's state-of-the-art (in 1984) Dolby Stereo originally sounded, whereas the 2.0 stereo on the laserdisc is a great deal closer. When these modern surround mixes are made for older films, liberties are taken in an effort to modernise the sound, with rethought dialogue/effects/music levels and often altogether new content; even in a relatively tasteful mix, like those of the Indiana Jones films, the feel is wrong.
Add to that the compression factor: laserdiscs with digital sound present their authentic stereo mixes as uncompressed PCM, whereas 'basic' Dolby Digital and DTS crunch down the audio information drastically in order to achieve manageable data rates. I won't get into the whole debate over whether this loss of information is noticeable, I will simply say that I find PCM stereo tracks to have a presence and intensity often lacking in compressed tracks.
So in summary: what you lose in channel separation you gain in integrity - not only in data, but also in authenticity. 'Immersive' is a subjective term, but where older films are concerned I trust the original sound engineers to present the mix as it was intended to be heard, and they can often surprise you with their strength.
I can't answer your question about PLII - perhaps someone else here is more an expert on decoders. However it works, it's pretty impressive; it can actually use the phase difference in a basic stereo track (e.g. music) to derive a full 5.1 spread, so a 4-channel source (which already has a dedicated centre channel) is ample for fairly convincing results. I try it out from time to time, though often end up switching back to basic Pro Logic for authenticity's sake.
Jonno covered it perfectly in this post.