They don’t require you to author lightmaps per asset at all, so this should be quite different than lightmass I believe.
Yeah but we all know Epic won’t take the time to put this in the engine, they are too focused on Fortnight right now
I didn’t say that it used lightmaps. “Online” SVOGI =/= Offline SVOGI… I’m assuming that by them calling it an offline version, it means that everything has been precached/baked and doesn’t need to be calculated during runtime, hence the description that goes along with it in the picture…
And I was referring to how it is not putting you in the same boat as lightmass does with lightmapping, lightmap seams, bake artifacts, etc. They turned SVOGI into a perfect solution for both real time and visualization use.
Epic saying SVOGI didn’t run well is still ringing in my ears. They dropped a gold sack because it was heavy to carry.
So sad, we should start a petition or something cause something has to change
But it is… You’re obviously missing the point of what they mean by offline SVOGI. If you have to precache the SVOGI, due to it being an offline version of it, then how is that different than sitting around waiting on lightmass to bake a scene…? I’d be willing to bet that it’s not just some magic little process that takes seconds to perform.
Completely baked SVOGI would have problems dealing with dynamic objects however, more so than standard lightmaps.
I’m simply referring to the fact that SVOGI in cryengine is a dynamic G.I solution that works fine for real time uses, even in most dense scenes runs pretty fine on relatively old hardware, works fine for viz uses, without needing you to spend precious time creating lightmaps, without requiring you to fix/prevent lightmap artifacts and things like that, and scales well on consoles. It has almost nothing in common with Lightmass. Something Epic didn’t want to look into. Not because “SVOG doesn’t run well”, but because they just didn’t want to have it due to paragon/UT/Fortnite not needing it. Which has been the case with spline decals, etc, etc. 
Complete Laziness, how sad, Epic needs to get their ■■■■ together
Can you please stop with your laziness BS, it’s just disrespectful to the hard work the devs at Epic put into this engine!
Epic people aren’t lazy, it’s just the general direction they take with UE4 which has been making it more and more useless for creating high end PC/Console games.
I know Epic is not lazy, they make a wonderful product, but when people like DICE Technical Lighting Artist Tilmann Milde are complaining about the lighting in UE4, you would think Epic would maybe think “Huh we have a very very experienced lighting artist who is noticing some major flaws in how we do lighting, We should listen to him and communicate with him and make some changes that benefit everyone.”…Sigh
It is pretty apparent, isn’t it? Omitting pre-processing step for the assets and faster iteration times.
I know right. Let’s talk about that.
As soon as you turn manual metering mode on you’ll see different lighting components getting broken here and there. I even reported black reflections caused by skylight when using manual metering mode, using the bug report form, and they confirmed the issue, and said that is not something they’d be trying to fix for now but thanks for reporting it! which basically means = physical lighting in UE4 is not something they made for people to use. It’s just going to be there for the sake of being able to say UE4 has physical lights regardless of if it works or not. I can list a bunch of problems with UE4 physical lighting right here, but why would I do that if they’re not listening? Basic issues such as not having physically based lighting makes me wonder if we should continue hoping for some high end dynamic GI haha.
I agree that there is a weird inconsistency in how Epic responds. For some issues and feedback we get an answer right away but for others, especially on the rendering side we have to wait sometimes for years to get an answer. Quite a few times I found myself in a position where I thought about asking for a feature but didn’t, because chances of it being a priority for Epic were too slim.
I also wonder how the fact that Epic pretty much makes more money now in a month than in a whole year prior to the release of Fortnite: BR, influences the development of UE4. I hope that the bigger budget will allow them to incorporate more features that go beyond what they need internally for Fortnite.
I thought it’s obvious.
Feature priority:
- Is it needed for Fortnite at the moment ?
- How much it is needed for Fortnite.
- Are you huge licensee to prioritize feature for you ?
There is nothing more to it.
Unless some big industry player needs dynmic GI, we are not going to get it.
There is a ton of other useful graphics features, that require incomparably less manhours to implement and far less effort to maintain, other than dynamic GI, yet even they receive low to zero attention. Should Fortnite been a title, on the cutting edge of graphics realism, things would be different.
One of biggest reasons why Paragon failed was because moba players have potatoes, but Paragon required “real” graphic cards to run on.
Unfortunately the more you invest in “graphxxxx” the more production costs rise and the more your target player base shrinks.
On the other hand the less you invest in graphics the less market chance you have of “causing an impact” in terms of marketing.
Unreal 4 is right on the “sweet spot” of great graphics also capable of down-scaling to run on potatoes.
That working so well, for a game where Epic already made around $500KK in such a short period of time, that leaves almost no room to gauge interest on graphics for games made for the average gamer whilst said gamer cannot afford the hardware required to run said game;
Specially when major player base are regularly reporting crashes on Consoles due to graphics already consuming so much resources in Fortnite…
They are actually cutting corners to stop those crashes, not thinking about expanding the engine (unless you’re Disney with a demand lol).
So, from the PoV of engine manufactures that makes very little sense to invest a lot of effort on lighting when the landscape is something like that.
This leaves an open opportunity for competitors such as CryTech, Unity and Amazon to bite a few costumers here and there if they ever manage to reproduce someday the toolset we have for granted in Unreal Engine.
“Epic Mega Games” can now go ahead and rebrand again to “Epic Fortnite Corp.” xD
I honestly do not get the obsession with these BEYOND CUTTING EDGE graphic features, especially dynamic GI. 99.9999% of people will just play around with it a bit, maybe do one render and that’s it. The VAST majority of best-selling and most successful titles in the past decade weren’t on the cutting edge of graphics, so in general it’s a useless feature and the (probably) hundreds of thousands of dollars that would have to be invested in it are better spent anywhere else.