Yes, that does work. Though it has some issues. One is that it eats up a lot of unneeded file space. With your method, every image essentially requires twice as much file space. It saves a lot of file space if you do the color manipulation in the engine for basically free. And if you want to change the look of it a little, it's a lot easier this way, since you don't have to re-export every image. Albeit, the color part of this does require using Renpy 7.4, which doesn't work on some older devices.
The other issue is that then in your script, you have to write out every 'show' and 'hide' statement manually to swap between the two. Which is a fair amount of work and means you have to be sure you do it for every time who's speaking changes in the dialogue. Which might be a lot.
With my method, and a bit of setup, it does both of these effects automatically without any additional work needed or taking up additional file space. While I commend your method and think it is clever and certainly good for its time. I think this is a nice alternative for people going forward, if they wish to do it as well.