Ok, that took quite a while to finish up - damaging stuff is fun, but quite repetitive and time consuming…
Feature creep will get me always… planned to have ~20 gravestones, have 85 
. Here’s where I’m at:
And as I spent quite a while doing this I did a quick promo video 
Looks amazing.
soo true!!
I like your enthusiasm.
The promo video is cool! I like the background music and the lighting.
What, no long low stone grave covers types? Or bigger box like ornate ones? 
Thanks!
No, no, no, no… those will come . Just the ‘headstones’ are done (for now 
).
Imported into unreal:
And Unity:
That took couple of hours to fix all the little issues with the models. As always - import to game engine as early as possible to catch issues as they arise.
Ah, and that would take even more time if I didn’t wrote python scripts to help with the exports from blender to both engines.
Quite a bit of time has passed, and I figured it’s high time to update this thread. I didn’t work much on this asset pack in December due to working on another project. But I did spend quite a lot of time on it. I managed to repurpose some of my old models, but I had to make a lot of new ones as well.
I did change my process/method a lot from my original assumptions:
Finally I reached a point that “vertical slice” of all kinds of models are done.
Here are some more screenshots. First the fences. They are similar to what I did in my 1 Month of Low Poly Practice Challenge - Finished! - #73 by bOBaN , but I had to redo many of them to be game ready, and have less triangles (this applies to all the models that look similar to what I did at that time):
All of those models are already imported in Unity, with neutral lighting and no post processing in showcase scene:
I’m at 305 models at the moment. Next step is to add variants to most categories. I should do this faster ( ). Still a lot of more work to do though 
Sounds unreasonably difficult and complicated. No wonder no one wants to do a Blender to game engine courses or tutorials.
Stay out of Game engines! Keep to pure Blender! 
Yes, it easing the easiest thing to do. But also learning Blender isn’t the easiest thing to do in the first place
Was feeling less creative in terms of adding models today, so decided to do a “reveal trailer” for the pack. More of a learning experience with making videos in Unity and Resolve, than actual PR material:
Though I would be interested in feedback that I could take into account when doing actual “PR” videos for the release of this asset pack.
Cool stuff! I also want to make some assets (mainly characters) that people can use in their games eventually. How did you go about learning to make this kind of video?
Thank you!
Go for it! It’s a lot of fun (just takes way more effort than just making the models).
The hard way . I don’t think my videos are especially good though, and I have a lot to learn still. What I did so far was watching a lot of tutorials and experimenting with things. I did branch out though, and was not searching in gamedev/3d area only. I watched for example tutorials about making movies in general or discussing how movies are shot or those weird videos about “how to become the best youtuber ever in 3 days” (they are mostly waste of time, but some of those people making are also showing some tips&tricks on how they edit videos).
If I may add a little feedback considering video. Try to avoid static images, where camera is not moving. If you have pan on 2nd shot, 1st and 3rd should have some movement aswell. And in this example Id try to add a little zoom-in. You can even do it in Resolve, by adding like 5% zoom in post if you dont want to make another recording. This is quite easy, and while it will reduce quality, going with small factor is fine and it adds a lot to overall feeling - its 100% worth it imo.
Also there is that heavy shake,when (i guess) the camera passes the fence. You probably know this, but that wont do in promo video 
in theory you can stabilize that in Resolve aswell, but this is not a thing Id recommend. Shake is heavy and Resolve will have to crop the image substantially.
But I like the framing and the overall idea is good.
There is also so much more you can do with that
Like make image correspond with text: Foliage - show Foliage, Gravestones - cut to pan to tombstones etc. Optimized - show wireframe on rotating objects etc.
The Burn transition is nice. As I see lately, rotating transitions are popular, if you want to make dynamic clip. (But that would mean digging info Fusion or finding something ready.)
Youd probably also want to show all the assets, even briefly, like in a showcase screen. (One thing to mention: custom (mesh) colliders are much heavier than basic colliders, so it might not be the best for optimisation in Unity. But its good to have them, dev can choose to use it or replace with Basic ones.)
But the most important is to know in what style you want to keep the vid: basically dynamic (showreel) or slow, that shows spooky climate of the graveyard. This dictates length od shots, transition types (fade for example might work for the second type, while not so much on dynamic cut), music, cut rythm (good edit is done with internal rythm), but well, I think we delve a little too deep
Little? Wow, I would like to see your feedback that you don’t conisder “little”
Thank you very much for taking the time! This is really valuable and will help me a lot! All of it is on point! 
Oh, I was agonizing over colliders a lot. I wanted to set it up in a way that covers as large number of projects as possible. I was really annoyed when I spent money on asset pack and I saw that colliders are missing or that something was wrong with them. I’ve also seen people complaining about it a few times.
I spent a lot of time trying to figure it out. It’s not easy, really. I even found conflicting descriptionshowing profiler graphs where simplified mesh colliders were actually more performant than a bunch of primitives one. And I came to conclusion that there is basically no right generic answer. So when I wrote “custom” colliders what I really meant was that I’m doing them “custom, per object”. If I can cover it good enough with primitives, I do that. If not, than I do both: colliders out of primitives and a collider from custom convex simplified mesh. Here is an example:
This will be probably good enough for most games, but some might want better collisions, so you can simply disable the box collider and enable the mesh collider:
And you get this:
I do compromise at times, when I cannot cover a mesh with one convex collider, I just leave it as primitives only, like for this one:
Though for this specific example I might go back and add alternative from 2 boxes and one custom convex mesh, but at this point I don’t really know if that is worth it.
Oh, not for me, not at all!
Once again: thank you for your feedback! I really appreciate it!
Glad to be of some help.
Your research on colliders is great and that’s exactly what Id do if i had desire/timer/skill etc. Having both types is very, very valuable, so good work with that. This just simply saves a lot of time for devs/designers .
What I did mean is that single basic colliders is better and usually enough - exactly as you pointed. If you need to use more basic colliders than one for a single mesh, than custom mesh would likely be better. But you are spot on this subject from my perspective.
If you want to check/practice editing, and want to spend some time on it, you can download some stock footage and try to edit a reel from that. (Whatever it is really) Than get a music track (any style you prefer be it classical, rock, disco or 8bit) and use it with your edit. You can than adjust your edits to fit with music rythm and see how much it changes if they work together. (Dont do any fancy transitions. Simple cuts. You can change them later if you want.) You can also mute the audio later and compare both edits - done flat and done with rythm and see the influence of rythm on edit. (Just dont try to meet every beat in music, or every bar. It cant be too uniform ) if you prefer you can skip first, flat edit and do it to music from the start. Bear in mind that im not professional video editor, so take it with a grain of salt.
Sometimes yes, sometimes no . Seen data supporting both and Unity docs suggest using ‘compound’ primitive colliders should be better. The most interesting and surprising thing I found is that someone needed to use a lot of rotated box colliders for his use case (i.e., rotated child game object with box collider). And at some point they switched to normal ones, as the overhead from having so many game objects was too big. Someone else pointed out that MeshColliders were affecting start up times too much for his scene and they switched to compound primitive colliders. Those are probably very niche use cases, but if someone bothered to write about their experience I’m guessing there are more people having similar issues, but just don’t write about that.
That’s really good to hear. I was beginning to wonder if anybody aside from me would care about it
I will do that! I should probably do those official learning lessons from Black Magic Design too finally
Maybe you aren’t a pro, but don’t sell yourself short! Your advice is really helpful! Thanks once again!
Small progress update: ground pieces, open graves, background mountains.
Viewport view:
Mountains rendered in Blender:
Quickly done map in Unity (more for checking if the modules fit together than to showcase it):
Will the set come with Zombies? Vampires? Ghosts? Grave robbers? Just to keep you busy.
I actually have a zombie and skeleton: https://vimeo.com/776028332 . I even figured out how to rig them to standard animation systems in both Unity and Unreal Engine (so that people would have easy time using them with animations on stores for both engines). And I planned those and more . But in the end I decided not to add another month to the production time . Might still do it for version 2.0 or as a separate asset pack though in the future.
