Well, I was going to wait to include this in another update post, but since you asked, and since I’m barely containing myself from posting more stuff, here’s some more stuff
Procedurally shaped lightning (bolt-build speed is adjustable, slowed down for this video to showcase it) :
Re-worked volumetric-cloud backlighting:
Proper moon-cycling with controllable self-shadow intensity:
And here are some screens of different times of day since time of day based variable automation is now complete.
Lightning is now physicalized, shouldn’t pass through solid objects
Volumetric clouds are finished, any remaining tweaking can be done via exposed parameters (as system is intended to do)
Height fog parameters are now exposed, if volumetric fog isn’t possible due to performance constraints/you don’t want it at all.
Extra features for post-release updates list on the Trello has been updated, includes stuff like light pollution, auroras, meteor showers, dust storms, and some other stuff.
After I’m finished with lightning, I have wind setup to do, should be able to finish that this week if all goes well, then it’ll be time for rain & snow, REALLY excited to bust those out and share media for it, I know several people have said they’re itching for it, I promise it’s on the way
Thanks for your continued interest and support, you all rock!
Yeh that lighting strike is really fantastic. I can see already buying this asset and completly overusing stormy weather just to get lighting strikes goin.
By the way do you have some performance numbers for current version? Since you are using volumetric clouds i am guessing it should be much more than Ultra dynamic sky, but just wondering where the numbers are landing.
You are correct, the volumetric clouds are fairly performance intensive, but I’ve included some quality sampling options to try and reach a wider range of hardware. Here is an image indicating my specs and the performance I get on each quality sampling option playing in a standalone game (play-in-editor FPS rate isn’t as accurate due to extra processes) as of the current build.
Thank you very much for detailed FPS pictures! Its great to see the scalability. Are you perhaps planing to add 2d clouds as third fallback? For example using it the way Ultra dynamic sky does it, becouse i wouldnt mind a little less quality of clouds, as long as you have everything else in the package that you are providing (weather etc).
That’s something that I had certainly considered, but it’s not currently on my to do list. If there are requests for it I could probably add it in as an option, though I don’t know for certain that it’d be in and implemented by release day, as testing everything and making everything scalable and consistent at least in pattern generation to the volumetric clouds would put me past my projected submission date of August 7th. I’ve added it to my list of extra features/post release updates on my Trello page to revisit once the other core pieces are finished. If you or anyone else would like to see the non-volumetric clouds or any of the other features there added in future updates, be sure to vote on their cards on the Trello.
If there is time before August 7th (projected submission date) I’d like to go back and add branches/cloud-cloud, if there isn’t time before then it’ll be added to my post-release update list, found here.
It looks awesome, but I’d be somewhat concerned by those performance costs. If you’re dropping down to 80fps on a 980Ti, I’m worried what that would look like on something like a 660 or lower, especially as this is an empty scene.
Sure thing, this is due to transparency overdraw on the volumetric clouds. Lowering the sample count combats this, as it lessens the number of transparent layers being stacked and rendered, though it’s not an end all-be all fix. Unfortunately I’m not a talented enough programmer to find a way around the overdraw issues, but I’ve tried to maintain the ability to still have decent looking clouds all the way down to the low quality settings, and I’m looking into adding 2d clouds for lower-end hardware as well, that way the rest of the system can be utilized without forcing people to forgo the dynamic clouds.
WAITING IS KILLING ME 