Firefox, which is open source, both sandboxes the Adobe decryption module—giving the browser control over the world that the Adobe module sees—and fetches it at startup from Adobe's site. It seems almost impossible that you'd be unable to hit the same URL from your own browser and put it inside your own sandbox such that the module can't tell it's not running inside Firefox. You have the hard constraints that the module can't examine the outside world except through the sandbox, and the sandbox itself is open source.
Edited to add: Widevine, the (Google-owned) decryption module Chrome uses, has no license fees. http://www.widevine.com/
As an actual example of the problem, there are millions of PowerPC Macs out in the world that run perfectly adequately with Linux and are capable of playing HD video. Is the opaque binary blob hardware-independent?
If it is then I can't even imagine what its purpose is supposed to be. If my "hardware" is a virtual machine that captures the video output to a file then the blob isn't even doing anything. And if it's not hardware-independent then there are obviously going to be innumerable minority platforms that it doesn't support.
I think they have given up to prevent you from capturing the output. The binary blob does decryption and decoding in one step, which prevents you from replaying the compressed stream.
I don't understand how that is even supposed to be useful. The output is still digital HD video. The higher the quality of the video the fewer compression artifacts there are going to be and the less having to re-encode once will make any difference whatsoever. And it would only have to be re-encoded once regardless of how many times it's played back or copied subsequently. If they're hoping for some kind of VHS-style degradation of the video quality they obviously haven't actually tested it.
I suspect you could coax 720p out of a 1GHz+ G4 if the code was sufficiently optimized -- another obvious problem with putting the video decoder in a black box.
But more importantly, the point is not the exact quantity of particular models of PowerMac in existence, the point is that if I have one which is otherwise capable of it, this nonsense interferes with me doing it.
And the same goes for every other thing that doesn't fit a mold. What about Linux on a PS3? What if I have a PA-RISC/Itanium/Power/etc Unix workstation? What if I'm using X forwarding so that my "browser" is really running on an UltraSPARC server?
It breaks anything the central planning committee didn't contemplate a priori or can't be bothered to fix. And it's not just old hardware, it's anything new or different. It keeps new platforms from getting off the ground.
For Widevine, it looks like the following would be required:
Completion of legal agreements
Delivery of a Widevine documentation package
Technical discussion to understand the device type, chipset, and review the Widevine robustness rules
Exchange of libraries, SDKs, and integration documentation
Keybox request and fulfillment (as needed)
Client integration testing
For piggybacking on the Firefox binary, maybe that would work but sounds like something you might get a cease and desist for.
Anyhow, if everyone agree to make the binaries 100% free as in beer (like Cisco did with H.264), then that's better than nothing and makes writing new browsers easier.
But the future is in new devices, operating systems, and virtualization. Being provided binaries limits freedom to make new things. And of course who knows what backdoors are in them.
Anyway, even if you can download and sandbox the Adobe module it might be a violation of the license agreement. This can even be a criminal offence in many countries because of special laws about circumventing DRM!
What about operating systems and hardware platforms not supported by Adobe or Google? PowerPC, MIPS, Tizen, *BSD, etc.
Edited to add: Widevine, the (Google-owned) decryption module Chrome uses, has no license fees. http://www.widevine.com/