Thanks for the comment. I can’t wait for handplane to get updated for UE4 also
However, I have to say I’m a data oriented kind of guy and it might depend on the models used for my testing. However, for this set of data, the Maya 2015 bakes did come out better than Xnormal for the UE4 viewport. My modeler had a hard time believing also as he is a long-time Xnormal user as am I, but the comparison screenshots showed the difference around the edging in the UE4 viewport.
I also find in my own testing that it is better to average out the lowpoly normals prior to the baking process. In essence, I have less overall distortion for my canvas to bake the normals on. This makes the low poly model shading look like garbage in UE4 without the normal map which can throw you off. However once you turn the normals on, its very good.
The down side is that Maya 2015 is veeeerrrryyyy… slow at map generation and crashes more often than I like, so I use Xnormal until i’m ready for a final bake.
Again, I don’t want to negate your statement but this was a long project of meticulous research on my part since normally I hear Xnormal, Xnormal, Xnormal… and since they are never going to release a new version of it due to the WoW expansion(lol) I decided to test out Maya 2015 and this was my findings. I will continue to use both programs and will do the same test and report my findings as I do them. =)
I would be interested to see others’ data from similar test instead of just saying its better.
I was very happy to figure subUV functions. I can take a set of images such as bullet hole variations on one texture and randomize them easily. Here is a pic how to set it up.
Then all you have to do is to create a dynamic material instance and set the frame index as shown here.
It may seem easy, but I was sure happy to get it working Plus I tied the decal template based on weapon type and impact surface for even more fun!
Have fun Deving!
P.S. One note that tricked me up for a bit was the Z setting for the Decal size. This is actually the Z-Depth at which the decal will penetrate past the collision to the model surface. If you set it too low, the decal will not show up.
I haven’t used Maya for baking any normals for UE4, so that could well be true, I was only going by Epic’s comments that xNormal was closer. If Maya’s tangent space is close to UE4’s, then that is clearly the best option.
When it comes to tangent space issues, though, all I meant was that no program is objectively ‘better’ as such, it’s just, some programs are going to be closer to UE4’s tangent space than others. And if Handplane ends up supporting UE4 (there’s no reason it can’t, since the UE4 source code is available) then that will end up being the best option.
I totally agree and I heard through the grapevine(maybe the polycount ) someone said that handplane UE4 support is coming. I also sent an email to handplane3d a couple of months ago to find out, but no response
On another note, your project looks awesome and I look forward to seeing more about it
I had some fun making burned decals for the laser and energy weapons! I also just released it for voting on the trello board. Check it out here VOTE if you Like it
It really does make movement feel great. I tried getting something like that to work in UE4 when I first started out a couple months ago. I knew I’d have to use curves but I didn’t know you could use curves with a custom x value. Is it possible to graph with an x value that doesn’t represent time? Either way, the implementation of your graphs are excellent!
Thanks again! You could put custom x time length in for the curve, but for modularity I like to “Normalize” the curves to 0 to 1 or -1 to 1, then you can just control the magnitude with the Animblueprints. The curve editor is a really poweful toolkit in which can be applied several different ways for any number of game systems.
In that way, X can be anything you want time, direction, speed… and the same goes for Y or for any relationship between two variables. =)
yours decals efects are awesome
I’m thinking to this,I want to keep the damages and replicate thru network,
if a player will visit the same area he should see the damage,I want the damage to be persistent,
do you have any ideea how this can be achieved?
If you want to replicate the decals down to the network, it would just cost the initial transmit bandwidth. Then you can put “0” for lifespan for an infinite duration. Then you just need to do performance tests to find out how many decals get created before you need to worry about framerate. Thats down and dirty…
The problem is when you have people join in after decals get placed, they would not see them because they weren’t there to receive the initial transmission. You would need to build a decal manager to handle this. The decal manager would need to store an array of active decals placed and always transmit them to newly connected players on first join.
Then you would want to optimize further like by building an aggregate system. This would be if I shoot in the same spot 10 times, I don’t need 10 decals drawn ontop each other for performance cost. You would just delete the underlying decals.
I’m just thinking out loud here, but you could go really in-depth on a system like this, but you really need to understand your performance limits for max decals and decal shader complexities which I would urge you to do first
thank you for those good ideas :),
if every building will get 1000 bullets probable will be heavy to simulate all for a new player or for a new building,
I thinking if there is a efficient way to alter the original building material,apply like a mask maybe with a parent material so the user will not “repaint” all the history ,
I not sure how I can replicate a material I have many things to learn
Yeah, i 'm not sure on that. It is something I’ve thought about before. Currently, you could blend a texture dynamically with a material variable but that is over the whole material file and not point specific. However, if you could somehow use the worldspace coordinates maybe a possible way to do a blend to a single point radially. I’m not sure how that would work for multiple points/bullets… if that was even feasible…
I do like your idea of masking but I have no idea how to implement. Rendering is one of my weaker areas. It would be nice if you could have a damaged layer blended under the normal texture layer, you could use the shape of the decal as the alpha mask to apply to the spot you want uncovered. Imagine photoshop masking, you only have to deal with a greyscale channel buffer. Then even if you worked out a render solution other problems would probably show up for sorting… hehe
Even though I couldn’t help you with a solution, thanks for the thought experiment. =)
Looks great! I really can’t wait to play the game, it really has the oportunity to be a next great RTS
I would love to see more videos, pictures of the models and ingame
Thanks Zackstefy, the guys and I appreciate the support! We going through a big push here putting UMG in and taking Slate out. We hope to be showing more screenshots soon.
Our next character, The Reaver, is starting the modeling process very soon. I can’t wait to use his roll mechanic. Rama already has a great roll system going =)
I’ve taken some time and did my first UMG & Matinee project. I figured the Main Menu would be a good way to learn them. Once I saw the StarFx on the marketplace, I knew what I wanted to do.
This was a quick proto-type and lots still to do, but I was happy with the UMG + Matinee synergy. I had massive fun doing it!
Oh God i’m so excited about this game, looks AWESOME SO FAR.
Btw, can you show some environment work? Since i’m more an environment artist i’d love to see some game-props or making-of tutorials
