So for a finished product, the way it is now is better.
Plus storing extra data in the dark areas for display is pointless. You only do it because you want to manipulate the data and maybe bring something out of the dark area, but in the final product, dark is dark.
If you, for whatever reason, wanted to give the user more data, then you would provide a 10-bit or 12-bit linear file.
Do I understand correctly, that this idea wouldnt work because you either end up with the same quality, but bigger files or with the same file size, but worse quality?
I had this idea that you could increase the amount of black colors a monitor could display if you combine 4 pixels. Having 3 black, and one of them "one bit" less black. This way you could have a more fine grained black-white transition.
I thought a "log" video might be able to contain the information to allow this without to much overhead.
Right now we can transmit levels of black with a regular 8-bit PNG file. The issue is that you can’t display it unless you are using an OLED screen, which your phone probably has but your computer monitor doesn’t (usually due to $$ for your average person).