Those are just 2 examples of "strong" DRM, but since even those two have been so thoroughly broken, we can just assert that any DRM that is based on W3C EME is gonna be childsplay. I don't think anyone actually believes that this W3C crap is gonna be "secure".
There is still the question of the analog hole[1], but that's a separate matter (unrelated to open source vs. proprietary binary blob).
[0] https://blog.mozilla.org/blog/2014/05/14/drm-and-the-challen...
Of course, then there's the software integrity problem -- I can't imagine a feasible system that prevents bypassing the software verification completely. Or, for hardware checks, I can't see a regulation enforcing "All TVs must have this enabled" (i.e. you just buy from an open brand).
This would be more interesting though for authenticated video streaming. Imagine every user is required to reveal a real identity to retrieve content. Then they can not only watermark the content but point to the exact user responsible, as long as the content has enough degrees of freedom to support it. Makes file sharing a lot harder if you can be held responsible after an indefinite period.
For bypassing the software verification, it can be made quite hard although it's kind of irrelevant today because it only became mandatory (via BluRay Consortium "regulations") in 2012. So there are still lots of players around that don't do it and this will be the case for the forseeable future. Verance is doing a big push to get it into TVs and other things but I'm not sure how successful they are being. The technology works without a doubt but of course when you add up licensing costs, etc, it's not always necessarily a win.
How can an open source software prevent users from saving the content? If the CDM decrypts to plaintext, it should be trivial to modify the open source sandbox to save the plaintext data. (Obviously the software can prevent saving the content by default).