AAA Grass

Very interested as well with the caveat of performance. Also unlike most people I really like the existing grass XD please include it as an option in the final product!

Every picture I ever post on the UE4 forums with with dynamic lighting. Not a fan of static lighting at all. In this case, there’s skylight and directional light (casting shadow of course) both set to movable. Generally UE4 doesn’t even handle 20 dynamic lights painlessly so I assume in any scene you add a few hundred dynamic lights you’d end up with ALT+F4. :slight_smile: I didn’t touch any quality settings and Cvars other than setting the settings to Epic. It’s been a couple months when I click on Shader Complexity mode editor crashes, not sure what has went wrong with my editor but waiting for next update to come out.

Grass looks good, i have been spending whole day today going forwards and backwards trying to get the best visual/performance ration on grass. Few weeks ago made grass followed Vega games tutorial from youtube which uses custom modeled grass in 3d software where 2 planes lean/cross eachother, but quickly found that that method produces too much overdraw.Thus today redid whole thing in Speedtree/max combo, advantage is better wind than UE4 material editor can produce, much easier LOD navigation (altough all LODs were manually made), and perhaps the most important one billboards for distance. Havent finished yet, but here is the preview:

LOD1 has still room to improve and reduce red shading. Some wierd black spots in the grass aswell, not sure what is causing that.
The main thing to do is to create LODs to reduce red complex shading. LOD1 and 2 are fine so are billboards especially. LOD 0 is going to have overdraw there is nothing you can do about it, otherwise ground will be visible for sure. If only unlit grass would have fit with the rest, funy enough it looks better even unlit at first place.

Looks good :slight_smile:
Until now I’ve always used this free customizable grass from Aaron Neal:

It looks great, but one of my biggest issues with grass for games is its only ever just blades, when there is so much more to “grass” than some basic blades.

heh, oh and remember we’re not allowed to call things AAA here anymore, so you’ll have to come up with a different shorthand for how amazing it looks :wink:

Thank you all for the feedback!
I might actually just bump the shader instructions a bit more. I don’t like the simple grass wind and the panning normal map like in the kite demo doesn’t look right under all lighting conditions. (It’s a disaster there’s no unified wind system in this next gen AAA game engine yet) :rolleyes:

The test I did was really beyond what actually happens in games IMO. There are no billboards yet as well so I think a little more instructions wouldn’t hurt.
@Ironbelly, Really? I assume instead of fixing the sorting issues, they banned the AAA term? might have to go with 3A Grass then! :cool:

It isnt needed since Speedtree is perfectly pluged into UE4. Their wind in either case best that you can get out of engine.

“AAA” doesn’t mean what you think it does; tip: has nothing to do with quality.

Yep, Epic renamed all of our packs to remove the AAA from the titles.

How do you know Max doesn’t have a 15 million dollar grass making fund?
Sure, it refers to the size of the development and marketing budget, but it’s fairly synonymous with quality. I mean it is possible to make a 50 million dollar game that looks like ***, it’s highly unlikely, if someone’s spending that kind of money there’s a good chance they’ve got a solid crew behind them. Again, not a guarantee but it’s synonymous enough with games of the utmost visual fidelity that I think we can all use the term here to save ourselves the time of writing ‘Super duper beautiful high quality Grass’ again and again and again :slight_smile:

I’ve worked for marketing company for 5 years and I’ve seen a ton of nasty things while I was there… They do anything to sell anything, they do mostly nasty bad things, so every time I see a small marketing gimmick around here I try to tell people they don’t really need to do these things to sell what they want to sell (because I like this community I react like this).
I want to believe that in today’s world is still possible to provide services/products without attaching empty cliche words to whatever your product is; but maybe that’s just me btw.

Some tall grass.

Hey guys,

There’s so much to share.
Let’s start by saying I’d include only 1 grass in this pack, which helps better demonstrating the real purpose of this package.
What this package is intended to do is to let you produce infinite variations of any single grass. For that reason, all the images I’d be posting here from now on is the same grass, but a different variation created using the function included in this pack.

So many more variations can be made in no time! for any grass!

(Going to update the OP today. Documentation is also WIP).

The most important thing to me is always the performance :slight_smile:
If you manage to actually create the most performant grass, i surely will buy it. Maybe even very short grass?

I would not even care about any additional features! Just a variation of meshes and materials … and the best possible performance.
… then I would not only be interested, but buy it asap :smiley:

Performance without and with grass:

Nice! Would create kindly do a short video with shadowplay? I dont know how many others would care about performance, but i do … alot :slight_smile:
And if this screenshot actually shows it as-is, then it would be alot more performant then other available assets.

I really do respect this work alot … since I’ve also tried to create grass like this countless times. And it was pretty much a failure every time. Sometimes the performance was great but it looked like ****, sometimes the other way around. Maybe I’m too picky, but i need really short and high-performant grass for a pretty big landscape …

Will play around with shadowplay and see what happens.

2 New screens of same grass.

8ms with and 8ms without? Surely something went wrong in test there.

what would you assume something is wrong, if something is performing very well?! :cool: