azmaybe9
patreon
Finishing Neko vid music?
🕑 Added 2021-06-22 02:22:28 +0000 UTCNow working on sound and music
------
Ok, last time it took 2 hours to realize that I should've grouped the entire sides of the buildings together as I learned that I should cluster vertically so that the walls don't fall through the bottom walls. Will I need to redo the building? Eh, maybe. Or maybe I just haven't learned well on stuff falling throu- i'm just gonna get things done.
------
After a LOT and I mean A LOT of learning, practice, and studying, it's time to put this method to the real test. Chaos is the most performative choice so far with destruction but is slightly inaccurate due to no cache.
The plan of Blender destruction for room scenes then Chaos for whole buildings WILL work! The practice tests have been great after learning that I don't need to fracture too much for the back building parts! If this goes as well as I think it will, then the next step is adding particles to Chaos and then smoke with Fluidninja. Fluidninja is a whole class of its own and my next slowdown so... it actually might not make it into the video and I could just go with some normal explosion puffs for now rather than the thick ones where you can't see anything.
------
Alright, so we know that Blender can do detailed up close destruction really well but the performance gets absolutely destroyed once things get to a whole building level.
This thankfully won't require any asset backtracking. Epic Games has been updating Chaos ever since I had issues with performance of it before. Now the Chaos demo no longer has hitches and playing at half the framerate it should! (This heavily affected when I tried to use it for Giganta)
With that in place and me also realizing the limitations of Chaos, I've devised a plan I want to test today for potentially the fastest way to implement destruction with details for Blender destruction when needed.
The idea is to offset the cost of performance from destruction between the 2 programs.
So say I want to deface a building. Place the assets in UE4 then export the building to Blender.
CUT the building where it gets defaced in Blender then export both parts back into UE4. Chaos will handle the destruction there.
But what if I want it to be more detailed and inside a room?
That's where Blender destruction comes in. Chaos can't do Giganta picking up a building and chucking it at Wonder Woman. Only general stuff destroying it. Blender CAN do those kinds of interactions. So all the inside the apartment of Neko vid would be Blender, but the building destruction scene surrounding it should be Chaos.
Which brings me to think that the falling building from Neko vid should be in Blender with a cut near the bottom and destruction at the top once it lands. (Add rubble in UE4)
And now the only issue left. Chaos Cache still isn't working in UE4 in preview branch 3. But that's probably yet another reason why Chaos is stuck in preview branches. Anyways, this puts a hold on converting scenes to games at the moment. I still need to study that, but destruction is SO so important to me. What's more important is that it doesn't hold me back by turning the project into a slideshow and that's NO GOOD!
Still gotta figure out how FluidNinja will factor into all this but it will be ok to put that off to another small project for now. Chaos can give breaking data for particles to emit from and that's insanely powerful that I can't do with skeletal meshes from Blender.
-----
After all the crashing from the previous stream, I studied some more and now i'm gonna attempt the whole apartment when Neko crashes out of it. I might have to recreate parts of the building but maybe not.
------
Been nervous about doing a destruction stream but here we go
------
With character lighting getting much better now, I'm now replacing destruction and fixing any light leak issues or whatever around the map
