Any software/hardware or other methodologies for upconversion are based on 25 years of established practice in the film & television post production industry. I used to work for a facility here in Hollywood that pioneered these methodologies, and for a long time it was a dark secret. The cat has been out of the bag for some time now, however. I've used hardware, software, and good old manual labor to do this work. I can't comment on the very latest software tools, but I have used- and own - a TC Electronic System 6000 with the "Unwrap" package, and any more recent software-based systems are surely built on the same principles. "Unwrap" itself is really just a quick way to get to a result that one can make happen with any decent DAW and a few plugins. It comes down to these basics:
1. Stereo width management;
2. Center channel extraction/fabrication
3. Low frequency/bass management , i.e. do you have dedicated sub FX or are you simply re-routing everything below 80hz to the sub channel?
4. Surround creation, usually the primary stereo channels with temporal (delay) and frequency modification.

It's easy to do all this manually. The problem, especially with film/video material, is that point FX are not covered in any of the software-only solutions. I recently upconverted a stereo reality travel show to 5.1, and tested most of the discussed methodolgies. They all came out about the same, except that doing it "the old fashioned way - manual labor in a DAW - really sounded the best. But with ANY of the software or hardware all-in-one solutions, you'll still be faced with the problem of point fx. So, let's say that however you got there, you've done the upconvert. Now there's a scene where cars and motorcycles are moving across the screen. If you want to do follow pans, you'll need to do this manually. Of course, because the upconvert software & hardware can't see or interpret what's on screen. So for these point fx, it'll need to be done manually anyway.

I got down to about 4.5 hours for each network hour of material (47 minutes, more or less), and this was without specialized hardware or software and it included point FX panning and a full QC, subsequent modifications and 5.1 mix output.

The hardware & software solutions are ok for getting you in the ballpark, but, properly done, manual work will always be required. Because of this, it nay be worthwhile to just do it the old fashioned way in the first place.