We need a like button on the forums :)…
I used the painting feature dropping some grass or rocks, and i confirm the number fps displayed gonna down quickly. Today UE4 has not some instacing system like what we can see in some DirectX demos where you see hundred of instanced characters running at same time.
The best you can do is make your game using simple terrains without too much trees and grass , so you’ll be able to do play tests without any speed issues.
I can’t say if it will take a long time or not as this is related to some core rendering parts , you can ask Epic directly in the feedback thread.
Well the game is based around an open world sandbox, sure I can cut down on the foliage for now. But it’s going to come sooner or later, I agree no point in putting off the main portion of the game waiting for lush trees and what not. But it does need sorting, as I don’t want my level designers chucking paper planes and drinking coffee all day.
Haven’t tested on compiled vestion, but pretty sure FPS will be much higher.
I’m unclear as to what exactly you are trying to demonstrate here? If there a specific setting you are using that is helping the issue or what?
Could a vector graphic be used for a grass sprite? Has that ever been done?
If the problem is overdraw then that wouldn’t help, unfortunatly. It still has to be rasterized.
Here we are in almost 2019 and Epic has made zero progress on fixing this issue. All the other engines are lightyears ahead of epic when it comes to foliage. Such a shame because I love this engine. Its sad when Skrim mods with ten times the amount of foliage nets me more fps than UE4 
Well it shows that Unreal Engine was originally build for small, arena style levels, which would mostly be indoors or with very limited view distance. The engine wasn’t intended to do open-world stuff, and therefore it’s not optimized for it.
Considering that Epic is now mostly catering to Archviz, industry (like automotive) and is still working on improving static lighting, I doubt this problem is going to change in the near future. It’s up to us to build workarounds for large foliage scenery.
Quite a necropost that is. For the record and speaking without any bias or personal experiences, nobody so far had provided a test foliage comparison scene in a fully controlled and matched environment where engine X would perform better. To me, explanation behind this is simply the fact there isn’t and cannot be any major difference. If you are intending to argue against that, it would be preferable that statements would be backed up by some evidence.