I have made several tests, so I don't know which tests you refer to. Most of my testing was done with Vuescan on the Minolta Dimage Scan Multi Pro, but two test were scanned at a scanning service in Germany using a high end drum scanner at 6096 PPI.
In my experience, correctly exposed digital images have incredible amount of info in the shadows. It's more like the high end that may be problematic.
I mostly shot Velvia in my film days, and I have never really worked with negative film. The reason I tested with negative film was mostly to find out if that would be workable for me.
My tests were done with Ektar 100. One trouble I may have ran into could be that one of the channels my have exceeded the dynamic range of my scanner. It is specified at 4.8 but it's a gross overestimate for sure.
The digital camera I used is a Sony Alpha 900, always at 100 ISO and generally striving for ETTR. Processing is done in Lightroom, naturally from "raw".
Lets put it this way. I have done a lot of testing (around 10 rolls of 120 film just for testing), and I have done the findings I have. Very possible that a better scanner or scanning program may produce better images.
I have been shooting MF and Velvia for around 10 years, before migrating to digital, and I have been scanning film for something like 15 years using differeent kinds of CCD scanners, so I'm not without experience.
Erik your testing of DR comparing digital with film is different from any tests or practical results I have found.
I scan with a Coolscan 8000 and make 48bit dng scans, I could easily expose for shadows with film and recover highlights, plus all the colours stayed fairly true with a bit of post.
Digital gives me very little room to exposure for shadow detail, I tested on a very flat day. If it had been a sunny day I don't think you have any choice but to let the shadows go. Any attempt to allow for shadows does not just blow the highlights it messes with colours tones etc not many stops above mid grey.http://treewithoutabird.com/untitled-2.html