[VIDEO PLAYBACK] [STIRRING ORCHESTRAL POP MUSIC] [GENTLE GUITAR MUSIC] - In one of the fastest-growing
cities in the world, a love affair is blooming. [INSECTS BUZZING] Melbourne's residents
are talking to their trees. A conversation is unfurling. Dear 1517, I'm confessing
something very dear to me. I have fallen in love with 1583. Should I leave or should I stay? It would be great if you
could give me some advice. [BIRDS CHIRPING] Around the world, more people
are living in cities than ever before. - How's it going, 1517? I've been wondering,
what's the city like for you? - If we change our perspective,
we can see that trees are
nature's great connectors. [SYNTHESIZER MUSIC] When trees breathe, so do we,
each breath cleaning the air and restoring balance. Now past, present and future
are intertwined with trees, yet we have lost the
ability to see them. To thrive, we must fall in
love again with our city's trees and nurture our urban forests. - Hello 1517, we don't have a
lot in common, you being a tree and all. I'm glad we're in this together. Regards, a tree lover. [TREE FROG CHIRPING] - Hello, tree, it's your mother. How come you never call? AMANDA: Hey! Can you believe it's been a
year since we revealed the first look at Unreal Engine 5?! We went behind the scenes
of Lumen in the Land of Nanite to discover the inspiration
for those special features--and now, we're thrilled to
put them in your hands. Unreal Engine 5 Early
Access is now available. This build is for you
game devs out there who like to live on
the bleeding edge so you can start experimenting
with our latest features, including Lumen, Nanite,
World Partition, MetaSounds, and more. Keep in mind that UE5 Early
Access is not production-ready, but you can now download
the latest build and the Valley of the Ancient sample project,
featured during our YouTube premiere. Get all the details on UE5
Early Access from the feed, and share your feedback
with us in the forums. When you take a break
from giving UE5 a test drive, be sure to pop over to the
Unreal Engine Marketplace to download two new realistic
automotive scenes for free. In addition to an
expansive salt flats area, the Automotive
Salt Flats project features customizable
mountain ranges and special atmospherics. Or get away from the sun in
the Automotive Winter Scene with customizable levels of snow,
dirt, or wetness on a winding country road. And if you're looking
for something else, thousands of Marketplace
products are now on sale at up to 70% off during the Unreal
Engine Massive Marketplace Sale--arguably
our biggest sale yet. Find discounts
on everything from explosive sound collections,
modular castles, materials, plugins, to the entire Earth. Sale runs now through Saturday,
June 5. Ever thought about quitting
your day job to build a game? Solo developer and father of
three Jonas Manke shares what it was like to do just that. He's now working full-time on Omno,
a single-player journey of discovery through an
ancient world of wonders. Hear more about the Epic
MegaGrant recipient's journey from hobbyist to full-time
studio founder and developer. When AVSimulation set
out to improve SCANeR, their all-purpose automotive
simulator software, they chose to convert it to
an Unreal Engine platform. Learn more about
how that shift enabled them to improve the
software's flexibility, modularity, and realism while also
providing customers with a slew of new options. And now over to this
week's top karma earners: mindsurferdev, Everynone,
ClockworkOcean, Shadowriver, Fraps2020, Rene Zwaan,
w2lf, Kehel18, staticvoidlol, and chrudimer. Thank you all so much! To open up our
community spotlights, put your hands in the
air for the VizMo team behind this rocking
virtual concert! They're so excited to
finally be able to share it, so be sure to head
over to the forums and let them know
what you think. Next up is a Mammoth display
by Alessandro Mastronardi, complete with
dynamic fur in UE 4.26. They plan to create a short
film with it in the future--stay tuned to their ArtStation
page for updates. And last up this week is Kainga,
an ancient village builder, in development by
solo-dev Erik Rempen. Adapt to survive against
changing weather, lurking bests, and competitive tribes
as your culture and technology evolve. Wishlist Kainga on Steam! Thanks for watching this week's
News and Community Spotlight. VICTOR: Hi, everyone,
and welcome to another episode of Inside Unreal, a weekly
show where we learn, explore, and celebrate everything Unreal. I'm your host, Victor Brodin. And to help me with
the episodes that we have coming for you in
the next couple of weeks I'd like to introduce,
once again on the show, Chance Ivey. CHANCE: Hey, everyone. So good to see you. It's been a while. I miss you. I'm glad to be back. Glad to talk about UE5. VICTOR: And we're not alone,
though. It's not just me and Chance,
even though we will be on the stream
for the next couple of weeks. And for today, we'd like to
welcome you to Unreal Engine 5 Early Access. And to help us with this,
we've invited Michal Valient, engineering
director for Graphics. MICHAL: Hello. Thank you for having me. VICTOR: Nick Penwarden,
VP of Engineering. NICK: Hey, happy to be here. It's been a while since I've
been on the stream, so excited. VICTOR: It's good to have you,
Nick. Nick Whiting,
technical director. NICK: Hey. Been a while for me too. But happy to be back. VICTOR: And last but not least,
Simon Tourangeau,
engineering director of Tools. SIMON: Hi, everyone. VICTOR: Great. To kick off,
I'd like to start with a little bit of pretty pixels. And so let's roll the short
of what we did yesterday. [VIDEO PLAYBACK] [AMBIENT ORCHESTRAL MUSIC] [WHOOSH] [SHATTERING] [ROARING] [END PLAYBACK] VICTOR: We hope that
you all had the opportunity to at least start the download
of Valley of the Ancient and get UE5 loaded
on your computers. It is available
through the launcher. You can go to UE5 tab. Everyone should
be seeing it by now. If you don't,
just go ahead and make sure you restart the launcher. Yesterday was an
exciting day for all of us. We're finally able to
talk about all the things that we've been working
on since we announced UE5 earlier last year. And I would like to kick it
off with Nick Penwarden to talk a little bit
about what we've been up to since the
announcement last year. NICK: Yeah, sure. So the reveal last summer
was a really early look at some of the
technologies that we were developing for Unreal Engine 5. And so, since then,
we've had a focus on getting those
features closer to maturity, working on optimizations,
working on getting them ready
to be used in a number of different situations
and scenarios. And then, in addition to
what we showed last year, we've been working on
numerous other features that we're really
excited about for UE5. So between the two of those,
that's kept us really busy
for the last year. And I'm really excited to get
to share all that work with Early Access coming out. CHANCE: Whoa, that's awesome. And yeah, so I remember,
Nick, sometime last summer, chatting with you and Kim and
a bunch of the other Engineering folks about kind of the
lead-up to Early Access and kind of talking about we we're
going to release in 2021 and what that's
going to look like. Can you kind of
give us a little bit of an overview of what
Early Access is to Epic. Like what do we have in there,
what was our focus for defining the feature set that's
going to be in there, and what do we want
developers to get from it? NICK: Yeah,
I mean, I guess, first and foremost,
Early Access is the first look that developers are
going to get in terms of getting their hands on
UE5 and playing with it. And so it's the opportunity
to see what you all do with the technology
and then take your feedback, see what's working well,
what's not, and incorporating that feedback
into the ongoing development of Unreal 5. In terms of feature set,
I mentioned that we've been working
on a lot of other features outside of what we
showed last year with Lumen in the Land of Nanite. And so it was
pulling together all of those streams of development
together into a cohesive whole. So Nanite and Lumen,
I think everybody remembers those
from our demo last year. But then in sort of bringing
that together with the World Partition system that we've
been building for building and editing large worlds,
for streaming large worlds, we've-- last year, we showed a preview
of some of the new animation technologies that were going in. And those have really
evolved over the last year. And then the video
you got to see, in depth a bit more,
how those have evolved and advanced over time. And then some
exciting features that we didn't talk too much about last year,
I think, like MetaSounds and the future of audio in Unreal,
some of the developer workflow improvements that
we've been working on right, so game feature
plug-ins and how to enable the writing of good modular
gameplay code and gameplay features. As well as, of course, I haven't
mentioned the new editor UI, but we've been iterating and
working on that for a while, trying to really focus on
the developer experience. And there's a couple of
behind-the-scenes things that I'm glad a lot of
people have noticed, things like the asynchronous
building of textures and meshes so that can get into
the editor more quickly, trying to make sure we're
blocking the editor as little as possible, so keeping that
editing experience smooth. Those have all been focuses
for us over the last year. And yeah, Early Access is
really pulling all those strings together and culminating
in a still-early look at UE5, but an early look that is more
encompassing of the feature set that we are
planning on shipping. CHANCE: No, it's great. It's been a big pleasure
working on a number of different systems that
use a lot of these features as we have gotten ready
to move into Early Access. One thing I noticed,
when working on Valley of the Ancient
specifically for this launch was how at home I
felt in UE5 versus UE4. I came from the
UE3 world into UE4. And there's a lot of
conventions that kind of stayed the same between the two. But the tooling was advanced
enough for that generation jump that I had to relearn
a handful of things, which is good and exciting. A lot of us, internally,
working on this tool and working on this demo,
one thing I noticed was how easy it was just
to kind of pick up and apply a lot of knowledge
I already know, that it's not just that a
lot of the core frameworks are the same there,
but even with the editor UI, I mean, it operates mostly the same. It's just simpler,
more streamlined, and it's kind of out of my way. NICK: That actually reminds me of a another aspect
of US development that we haven't
really touched on yet. But one of the
guiding principles we had with going into
UE5 was making sure that developers could
bring their projects and their content
from UE4 into UE5 and trying to make that
as frictionless as possible. And so that was
always something that was on our minds
over the last year. We want to make
some pretty big changes to the engine, both new features
but also some of the systems under the hood. And making sure that
we do that in a way that we can preserve the
ability to take a UE4 project and bring it into UE5
with as little effort. And one of the things I've
been really happy to see over the last day as
people are starting to download and explore
UE5 is that a lot of people are reporting back
that they've been able to take the game
that they're working on and migrate it into UE5. and sometimes there's a
little bit of fix-up in game code. But by and large, maps,
assets, are all loading. And with some minor changes
for a couple of compatibility breaks,
it sounds like people are having a pretty good
experience bringing their UE4 content into UE5. And that's really
important to me because it means
that developers are going to be able to
make that jump from 4 to 5 and not have it be as a
big impact to development. CHANCE: Yeah, I remember,
early on, I was talking with I think maybe you
and Mike Fricker. And Mike's like,
we need to build a list of things that we know is OK to
kind of deprecate a little bit and find the appropriate
way to make sure that people don't run into too
much pain when they do that. And it was this really
regimented thing where we were
going kind of system by system to make
sure we understood what the impact for those things were,
just so people aren't blindsided by any of this. And I think that turned
into a doc we have that shows, hey,
here's some of the things that you might see that are
fundamentally different as we move forward. So one thing we
did too-- and Nick, I talked with you and Nick
Whiting early on, I guess. I say "early on." It's been, what, like a year. But about-- when we were
talking about Early Access. And that was the
drive to push out a sample project for a number
of different reasons there. And so what we had
ended up deciding was like we knew we wanted
to see what we can do with Early Access at that state. We knew the feature set. We had a good
amount of confidence that it was going to land. We talked with this
group here and then numerous other dev leads,
to kind of knock down the specifics of, hey,
these things we feel there's a certain
level of confidence that we're going be able to
land in the time frame to the level that we want people to
actually go and try them and give us feedback on. There's going to
be other things that are in Early Access
that are a little bit more experimental than
what we have there. But it was really
exciting to hit the books and think through
what that might look like. So Nick, what was that-- Whiting-- I think that was what,
October when we started the conversation
on Valley of the Ancient? NICK: Yeah, October is when we kind of started
ideating on what should we do, what can we do. Like you said,
it was important to find the features that not only
could we ship a project with but that we would feel
confident people could actually crack it open and mess
around with and explore a little bit on their own. CHANCE: Yeah, it was crazy. Because we had a bunch
of different efforts going on at the time with a bunch
of the talented folks at Epic. The Quixel team was
thinking about what they want to do for
Early Access and how they can show off a lot of their
assets using Lumen and Nanite and try to solve some
of those problems. We were working with
our Aaron Sims Creative already on thinking through
some of the next-gen animation tech things. And we were like,
let's get together and assemble a
project with these folks. And it's funny,
we look back a lot of other demos that we've got that we've done
around some of these releases, and it was kind of important to us,
I think, across the board, that we didn't want
to build something that tried all these
things out and could show what was possible
without giving it to folks too. So it was, kind of from the get-go,
a new lens to kind of look through
building this project that we wanted to do. So yeah,
so we decided to just find the way that we can provide
some really nice, clean examples for all of the features
that you called out in the video and more. There's quite a
bit more in there that we'll be covering later on. We wanted to show specifically
how far we could push some of these things for today. And so that's kind
of how we ended up with the big landscape,
try to take things to the limit, fail a lot, figure it out,
understand how the tools work. Michal was kind of laughing
over there because we went back and forth
a lot on specifically how do we take Nanite to
scale and not paint ourselves into some weird corners. And we learned so
much in the process. It's been really exciting. And then also, too, it let us
really walk through, gave us a goal to get something out so
we can harden some of the tech that we knew we had there. The engineers,
the dev leads, were really excited to get
involved and have somebody using their tech day by day and
embedding with the content team to actually build something. And so it was
really cool to do that. Go ahead. NICK: Yeah, I was just going to say it was really a
good opportunity for us to experiment with using some
of these new features ourselves, again, pulling those different
strings of development together and making sure
that they all work together harmoniously, and then letting
us experiment with, as you say, pushing the tech to its limit. So what happens when we
try to build a 4 kilometer squared area entirely out
of Nanite meshes? What does that look like? What are the
workflow challenges? What are the ways we
can optimize those systems in the engine and really hone
in and improve those workflows? And so it was a really good
learning opportunity for us. Because this is all
new revolutionary tech in a lot of ways. And so we are learning,
on the fly, the best way to take advantage of it. CHANCE: Yeah, it's crazy. I think we'll have,
on a later livestream, we've got Brian Karis
to come on and talk-- VICTOR: Next week. CHANCE: Yeah,
we have Galen on too. So we can talk
artist versus engineer on letting somebody just go nuts,
Just go free with building a
Nanite environment, just hand-place everything
and see what happens and then tell us how
we're going wrong and finding the right
middle ground there. And so there's a lot of
really cool learnings there. Another thing, too,
I think is interesting-- and I think this
happens every time we go to release something-- the conversations we had
back in the fall about, hey, what do you think is going
to be ready for Early Access and what do we
want to commit to? That list, the tools got better
and they got more robust, but also we were able to add
more stuff because people were really excited about
seeing people touch and feel their tools for a while. And they were able
to just add two more and kind of commit
to more as we go. And it just turned into a really
rich Early Access experience even though it's
still super early days. There's really a lot to get in
there and poke around with. VICTOR: All right,
we're definitely getting a lot of questions here. I'd like to let
everyone know that we are going to do Q&A.
We just want to make sure that we go through everything
that we'd like to talk about. Chance already mentioned this,
but go ahead and check out the schedule on the
Twitch About page for what we are covering
in the next couple of weeks. If you're curious,
Nanite, Lumen, World Partition, we're doing a deep
dive on all the new animation features. And so if you are specifically
looking for some answers to those questions,
we will be covering all of those in detail. But I think we should definitely
talk a little bit about Valley of the Ancient,
which we all have a difficult time not saying the code
name for it here. So we're going to do
our very best to refer to it for what you all know it as. Chance, would you like to
show a little bit of pretty pixel for us. CHANCE: Yeah, sure,
I'll pop over real quick. I've got the project open here. We might not be going too
much into it in detail here. And I'm sure you've
seen the video specifically. But as mentioned, we've got-- what is built out
here is it's roughly-- well, it's much bigger
than a 2-by-2 environment. You can see here in our
World Partition mini-map. The actual space that's,
I think, 2 by 2 is roughly-- if I use my little
kilometer thing there, it's about these 25
cells in that space. But we've got a bunch of
items here in the distance that you can see
when you're lower. They're just always loaded. It's a hand-built map. It was done in just about,
I think, 10 to 12 weeks by a
handful of Quixel artists using our open-world tools,
which we'll get to in a second. We'll have Simon
talk to them a little bit. We wanted to kind
of see how far we could go with Nanite
and environment and using some of the
World Partition streaming. We could go way larger
than what we had here. We just were working within
some time constraints that kept us focused just
kind of in a specific area that we knew we could get
to a certain level of quality and density and time. It stars our character,
Echo, from Reverb. It's not necessarily
an extension-- I said Reverb. Ha, there we go,
it's the first one. Lumen in the Land of Nanite. It was the demo
from last summer. We have code names around here. So we have Echo
here at a campfire. And she can use her
remote to kind of explore the space and
kind of get into all the hyperresolution
geometry as we go. We have the data layer system. Real quick,
I'm going to pop these over. This might take a
couple of seconds to load since there's just
a ton of content in there. And I haven't loaded
it on this box quite yet. Let me get rid of this one. VICTOR: Chance, what hardware are you running the demo on? CHANCE: This one right here is-- let me get back down
to where we were-- well,
this is a 3080 on this machine here. This is one of the ones that
we use to record early on. But most everybody
that works on the project was anywhere from 1080 to 2080,
for the most part, as we built the entire demo. We used the other features
inside the new open-world tools to switch over into this
dark world re-imagination. And then we end up fighting
a big ancient robot, held together with
magic is what I say, since he's built with a bunch of
pieces of Nanite geometry that are all kind of
intertwined and everything. And so it's pretty cool. It shows off all the
features that we'll be talking about here
in some level of depth. And then we'll go
super deep on all of them as we move on to the
rest of the livestreams. VICTOR: I just realized that we don't have a
representative here today, but we will in the coming weeks,
but Quixel will played a big role in the demo. And we're seeing,
I think, almost-- not all of it,
but a lot of the environment assets are Megascans, right? CHANCE: That's right. Yeah, the team went out to Moab,
actually, back in the fall. And they used some new
drone scanning techniques where they fly a
drone up and they're able to scan massive
cliff faces and then take all that scanned data and all
those pictures and composite them together to make super
high-resolution geometry. And there's a new
asset type that they built that's basically a
combination of those things, almost like a prefabricated
collection of those meshes that can be easily used in the scene. And if you've seen the video,
it's outlined pretty well there. It's pretty awesome. The team is using
some spline tools to be able to dress up
cliffsides and things like that using some utilities
and Blueprint, and really,
quickly build out a lot of the scene. VICTOR: Cool. All right. Let's go ahead
and dive into-- so we will be diving into Topaz in
detail in the coming weeks, especially in regards to to
Nanite stream next week. Galen Davis will
be showing off sort of how Nanite was worked with,
the workflows, the implementations,
the stuff we learned, all of that. But to move on, I would like
to ask Simon a little bit about-- we have a new user interface,
which is probably the first thing that
folks see when they download UE5 and open it up. Would you like to talk
to us a little bit about it? SIMON: Sure. Yeah, so for UE5,
we really wanted to simplify the interface
and make it not overwhelming for new users but still being
empowering for advanced users. So as Chance said,
transitioning from UE4 to UE5 is pretty easy on the UI front. You won't need
to relearn the tools. You might look for
some of the things. But it's not major changes. We wanted to make
the viewport really a first-class citizen,
as over the years it seems like the viewport
got smaller and smaller. So now the content
browser is in a drawer that can be pushed down. You can collapse
panels into sidebar tabs. So you get more of the
nice pixels in the viewport. Now we have a dark
team for the editor. And you can customize it. You can make your
own team if you want, change the colors and all. In the future, we'll probably
offer a light team also. Now, in the future, we also
want to simplify some workflows so that you can
modify properties without having to
go into Details panel, like changing light
parameters and things like that. But it's not in Early Access yet,
but will be for 5.0. So hopefully you'll
like the new UI. CHANCE: Yeah, on the new UI,
when working on Topaz, one thing I noticed that wasn't
called out specifically to me-- so when I first saw the Content
Browser drawer, I was like, this is great. Because I need it all the time,
but I don't need it in my face all the time. Because I don't
do a whole lot of set dressing,
but I need to get access to a thing. The drawer is available in
pretty much every other window too. So if I'm in a Blueprint and I
need access to a file on disk, whether it be in a
public field or if I'm trying to spawn something,
I can pop up in the Contret Browswer drawer, find the thing I need,
and drag it into space or just reference it
directly from there. It's super nice. I've learned to use that
pretty much everywhere I go. If I've got a
multi-monitor setup, depending on what I'm doing,
I might pop it out and set it somewhere. But yeah, for this project,
almost always I left it in a draw and just
grabbed it when I needed it. And it was great. VICTOR: And the
drawer is not finished yet. No promises here, but I did
hear from some of the developers that they're working on the
possibility of adding more tabs to it so that you can, say,
have-- because it was one of the biggest requests, right,
being able to add the output log and so that you can just
pop that open and look at it as you're playing in viewport. Exciting stuff. And it's Control-Spacebar
for those of you who haven't found it yet. Should probably be one
of the easiest key binds that you might
accidentally touch. Even someone did tell me
that that's what happened. Simon, were there any
other details on the UI that you were
interested in covering? Or should we move on? SIMON: I think we can move on. We have lots of
stuff to go over. VICTOR: Sure do, sure do. CHANCE: And one more thing I will say about it since I saw
the question come up online, you can set it up, the viewport,
the same as you had in UE4 if you prefer that. Nothing there is enforced. I think the intent
was to make sure that you can kind of
customize in the way that best fits your
workflow needs. VICTOR: There's that Dock in Layout,
I believe, is what the button's called,
on the right side of the Content drawer. New flashy features,
Lumen and Nanite. Michal, you have been the
director for the development of these features. I don't know, where do we start? MICHAL: It's, I mean,
tough to bring something new. Because we've been talking
about this for quite a while. I think one part that I
liked about the demo is how it's shown that, with
these new technologies, actually, that allow you to get
to this level of quality with fewer people and
short amount of time, if you think about
how polished it is and how much detail
we were able to get in, I don't think it was-- obviously Nanite enables
a lot of more detail. But just the flow
and ability to create this with a small group of
people in the time given, it's amazing. It's pretty revealing,
I think, for what is achievable with
the new workflows. When it comes to Nanite,
I mean, we already talked about it a bit. It's our new
micropolygon renderer. It's a combination of software
rasterization, hardware rasterization. And basically it does allow
you to let loose significantly more when it comes to the limits. You can use a ton of instances,
ton more polygons. Basically throw it onscreen,
it will work roughly at a constant rate. There are some performance
scalings that obviously we can run into. But you don't have
to worry about LODs. You can just take your cinematic
model from your DCC tool, put it into Unreal, and go. You have it onscreen. It runs a performance. We have still a lot-- it looks very polished. It looks great. But we still have a
lot of work to do here. Right now we're working
with opaque and rigid meshes. We have plans to add scanning
in the future to add translucency. That's something that's not
in the Early Access build yet. So yeah,
it's one of these enabling features that will and are changing the
way people are making content. Lumen is-- I've seen
some great images people creating with Lumen,
like the day Early Access was released. I've seen awesome images. It seems like this is one of
these very accessible tools. It's a true GI. So you take your scene,
disable lightmaps, put it on,
turn on the lights, start to play with emissive materials. It's great. Everything reacts to it. Time of day actually
suddenly looks so much greater because Lumen reacts to it. I think the great
thing we're doing here is how widely accessible
Lumen is in terms of hardware. You don't really need to
have ray tracing hardware to run Lumen. It's great. It's supported. And we are still
working on improving it. But we have the whole
software tracing solution for the hardware
that doesn't want to use hardware ray
tracing or for the titles that make the choice. That's actually
working just as well. And that's what we
used in the demo. It's a mix of screen space
and world space tracing. So to keep things smart,
we try to gather as much information from screen already as possible. And then we add even more
detail from our world space Lumen structures. Another thing I think we
didn't give too much attention yet is our new temporal
super-resolution technique. So it's the next step in
our temporal interlacing that's been available
in Unreal for a while. It's great. There's a lot of
magic behind it. But it basically allows you
to render almost true 4K image at a cost of 1080p. It's very easy for people to
try if you just download Unreal. Download the demo. You can turn it on and off. And you can see how the
frame rate actually changes and how small the
difference in quality is. And we're still
working on this feature. I'm pretty sure it's
going to go even better. Yeah, and oh,
I should not forget, actually, Groom and Hair. It's not completely new feature,
but I think it's shown prominently
with Echo this time around. It uses our custom
rendering pipeline to actually be able to render
individual strands of hair. There are several dozens
of thousand strands attached to Echo's head. And we rendered with
our custom technique. It's properly lit. We have voxelized technique. We use voxelization
there to make sure that shadows work really well. It's fully integrated with
our virtual shadow maps. So a lot of improvements
around here integrated with
volumetric cloud lighting. We reduce the footprint here. And the performance is
greater than it was before. So we are moving much
forward with this technique. I mentioned virtual shadows. I don't think I give
them credit before. It is one of the key
techniques actually that shows off the Nanite. It might sound a
bit counterintuitive, but the amount of detail,
shadow detail, we are able to get with
the virtual shadow maps do bring up all the
fine geometric detail coming from the Quixel
Megascans and the render to Nanite. It's sort of a sibling to the
Nanite in terms of enabling all this detail on screen. That's about it. VICTOR: We're getting a
couple of questions in regards to what Nanite can
currently and will support, one of the main
ones being foliage. Folks are applying
it to tree assets. Could you go into
just a little bit of detail there of where
Nanite currently is in regards to what
it supports right now and what we might
see in the future. MICHAL: OK, so Nanite currently supports opaque
and rigid geometry. That means what
you've seen in the demo, actually even the Ancient,
was Nanite, even though it's moving
but it's all rigid parts. So it allows to be
rendered this way. We will be adding,
in the future, scanning support. We want to add translucencies. That's what, basically,
you would need for foliage, animation and translucencies. And also we have some ideas
and some research ahead of us on how to render
foliage with Nanite. Because it's less about
what Nanite allows, the amount of performance,
the amount of polygons you do. But with foliage,
you need to think about how to reduce the
geometry and complexity as a tree or a bush moves
further away from the screen. So that's, again,
some bit of an R&D we want to do in
the future to enable. Currently, you can take a tree,
make it opaque mesh. So you can model
the leaves as polygons. And it will render. It might just not
look entirely correct as the tree goes further
away from the camera. Because we don't do
all the right calculations considering the sparse
nature of how tree looks. VICTOR: Thank you, Michal. For those of you interested
in more details in Nanite, I'm going to keep
repeating this, but I do want to
let everyone know that we will be
covering Nanite in detail next week on the stream. And if you go ahead and
look at the Twitch About page, you can see the full schedule
for the next almost two months of UE5-related livestreams. And I'm sure there will be
more coming later in the year. Let's move on to
our large-world tools. They involve quite
a few new tools and features and
workflows in Engine. Simon, would you like to
go ahead and talk a little bit about it. SIMON: Yes. I'm pretty happy that we
finally have World Partition out in the world. We've been working
on this for a few years at least,
with a fairly large team. The reason why we wanted
to work on World Partition was to greatly simplify
workflows for large worlds. Like if you think about
building a massive world, it means you probably
have lots of artists and they all want to work
collaboratively and be as efficient as possible. And the way Unreal
was set up in Unreal 4, all the actors were
stored in level files. And it brought contention where,
if you want to move an actor, you have to check out level
that was already checked out, you're blocked. So with World
Partition and one file per actor on the editor side,
we split the actors down to the five levels. So each actor is its own file. And what World
Partition does is that it handles dividing the whole
world into a bunch of editing cells that you can decide
to load or not load to work into your big world. So instead of loading a 4
by 4 kilometer world that's going to take a
long time to load just to work in a small
corner of the map, you can just load that
corner directly, that cell, and probably load 100
times faster than you would have if you loaded everything. In some cases,
you can't even load the full world because it's just too massive. So World Partition
becomes critical. The other aspect-- well,
let me back up a bit. Before that,
to handle the contention, what people would
do was split the data into multiple sublevels. So if you want to have
a level arches working and you have to have lighting
arches working and sound, you would split
that in sublevels. And then you won't
have contention problems. But this brought data
management problems where you have to think
about splitting everything, you have to think, OK,
I'm going to have four arches, I'm going to have
four sublevels. And then if you move
actors between sublevels, it adds complexity. And since the sublevels
are also used for streaming, you can end up in cases where,
whoops, the actor is now in the wrong level and
streamed at the wrong time. So it just becomes complicated. And we really wanted to
make it easy so it just works, so you just place your actor,
place your stuff, do your gameplay. And it's just going to
work without thinking about splitting data. On the runtime side,
what World Partition brings is, again,
no need to worry about sublevels. No need to worry about,
OK, I'm going to load that part of the world. And then when I
move to that area, I'm going to load
the second part and unload the previous part,
and then have to worry about
memory budgets and all that. So World Partition,
when cooking the data, it splits the world
automatically on a grid and creates, basically,
mini sublevels for each cell. And then we use a streaming
radius around the character or around any point
of interest that you would want to use to stream
the necessary cells as you move. So as you move through the map,
it's loading the cells in front of you,
unloading the cells behind, it's managing the HLODs. It just makes life much easier. And then you can customize,
you can change the size of the
runtime cells if you want, depending on the
density of your world. And for very advanced users,
when closing and shipping a game,
you could, if you want, to create multiple
grids having a coarser grid for the buildings and
such and a fine-grained grid for details on
rubbish and things like that you would want to
have different streaming policy, different streaming radius. So yeah,
what else for World Partition? Yeah, with World Partition now,
World Composition is deprecated. And if you have a game that
already uses World Composition, we have a commandlet that can
be used to convert automatically. It does most of the job for
every World Composition levels that were distance-based if
you add some logic to handle the streaming, then there's
a little bit of tweaking to do, but it's not too bad. And the last point to note,
World Partition is not enabled by default.
So if you create a new project, it won't be World Partition. You need to go
into Project Settings, enable World Partition. And then,
if you create a new level, it will be automatically
a World Partition level. Make sure to read
the documentation. It's a little bit different
working in that system. You need to figure out
how to load the cells and all. But once you understand the basics,
it's pretty simple. Maybe last point,
for the one file per actor, we've also greatly improved
our source control integration so that when you get to submit,
you see all the actors that you modified. Previously you would just
submit the whole level blindly, not really knowing
what was changing there. So now you see that,
oh, touched this light, I touched this actor. I should not have
touched this actor. I won't submit it. And then you can manage,
directly from the editor, your changes
and then submit it from there. VICTOR: Thanks, Simon. One of the questions we received
in regards to World Partition came from alleya but
a few others voiced it. "What is the state
of multiplayer support for World Partition?" SIMON: Yes, for multiplayer, there's a few things-- let me check my notes. Yes,
there's a few things that were not fixed in time for Early Access. So if you play around
with data layers, it will not be replicated. And on the server currently,
HLODs are getting loaded. So that's probably going to
be fixed in a future release. And replay on the
server also doesn't work. But other than that,
it works fine. CHANCE: And that's just
for Early Access, right? There's teams working on-- SIMON: Yes. Yeah,
we already fixed most of those. But we just missed the cutoff. CHANCE: So that's great. These tools have been
really game-changing for us and working on Valley of the
Ancient-- not the code name-- we were able to,
as I think Simon said, there's a little bit
of change in thought around how we build-- similar to Lumen and Nanite,
some of the more traditional
methods of just building games or whatever, you
have to rethink them a little bit. But once we got the
teams up and running, which was within a handful
of days of us just talking through how it's going to work,
we had like a dozen artists
from the quixel side and from the Aaron Sims and
the content side on the engine team, kind of like
putting things in the world all the time,
all day, every day. And we worked
around the world too. I mean, we had people in I can't
remember how many countries working on the project. And no one had the,
oh, I have the map and I have to give it to you
unless it was like a Blueprint thing that we had to handle. And that's completely separate
from a number of the things we were talking about. And we would just sync,
and then just more and more and more stuff showing up in
the map all the time. I don't think,
in the time frame that we had, that we could have
really built it that way. I mean,
I can't remember how many-- it was something like 2 million
mesh instances in the Moab scene. And there's a ton in
the dark world scene. We were able to have people
work in the same virtual space next to each other. We'll probably talk
about it in a livestream. I think we have Jean-Francois
and myself later on. But the Quixel team actually,
in the world, because you don't have
a sublevel that you're checking out,
they put these big white planes that says, OK,
I'm working up to this white plane. So don't worry about
going past that point. And I remember Whiting is like,
what is with the planes? Is this like some debug thing? And it's like, yeah, it's their
bookmark, just to kind of say, this is what I'm working to. And everybody
could just go crazy. The "one file per actor"
support is really, really, really something
special with that, being able to see exactly
what you changed, why you changed it is really awesome. I can't tell you how
many times I've worked on a project
where the map file's dirty and I don't know if I did it. And so I don't know if I'm
losing a change that I need. But I don't think I did it. But I'm a designer. I'm a gameplay person,
putting stuff down. And I'm just not sure if there's
something I'm not referencing. And so I just have to
revert and hope for the best. Not anymore. It's phenomenal using
that tool for the project. SIMON: I think
the first thing that will be difficult to adapt for
seasoned UE4 developers is stop thinking
about sublevels. It's like the first instinct. It's like, OK,
I'm going to create a sublevel for this, a sublevel for that. Don't worry about that. Just put your stuff in the map. And stopping that
reflects takes a few days. CHANCE: [CHUCKLES] SIMON: But they
fall back to that. But with World Partition,
you don't create sublevels. We don't even allow it. CHANCE: Yeah,
we were using data layers upfront to learn how they work. And then just to generally
be able to, across the board, turn things off to see
how they impacted perf. A good example is foliage. We didn't really know
how some of the bushes we were going to
put in there we're going to affect our performance. And so we put them
on their own data layer that we can toggle at runtime. And I put it to a hotkey so
we could do some fly-throughs and just do perf recordings
where it just turns on, turns off,
it turns on, turns off so we can kind of see how
you spike and move down. And at the end,
we collapse those down to only the data layers
we need to actually switch between the two worlds. So they can be a
good organizational tool if you want to,
but they don't have to be how you actually operate. And Simon, too, on that note,
if I have a data layer disabled, is it like the cells,
is that stuff just not in memory anymore also? So I'm not paying for things
that I don't need at edit time? SIMON: That's a good question. CHANCE: I wasn't really sure. SIMON: I would believe so,
but I'm not sure. CHANCE: OK. I think the fact that,
if I come in here and I disable it in the editor,
the campfire replace one is
the one that has all of the super-dense geometry. We turned a bit of that off. The fact that it
takes quite a bit to pull that down
and turn it all off, it's not just flipping
hidden flags, I'm assuming that it's
[INAUDIBLE] a lot of that out. SIMON: Pretty sure it is. CHANCE: But that's
the same thing too, like with World Partition. Because people had
their dividers in that space, they knew the cells
that they could just not have to ever worry about until
they come up to that divider. And so that means you don't
have to have more resources locally to understand the context
of where you're building. You just kind of take what you need,
build what you need, submit what you need. And eventually you
might add features like being able to lock a cell
so that changes can't be done in those areas or basically
like checking out the checking out the level if you want. So maybe in 5.0. CHANCE: Yeah. Super cool. VICTOR: And now
we definitely don't want to forget about audio. Because if we do,
then Aaron McLeran will come and let me know. Hopefully might be
in chat today as well. In UE5 Early Access,
we have MetaSounds, which is getting closer
and closer to what one who is familiar with
working with DAWs might be more used to in
regards to how you can shape and dynamically
modulate audio in real time. Whiting, would you like to
go in a little bit about what MetaSounds involves? NICK: Yeah, but now I'm scared that Aaron McLeran is watching
and judging everything I say. VICTOR: That's the case anyway. He's going to watch
this in the vault. So just-- NICK: Hi, Aaron. VICTOR: [CHUCKLES] NICK: [CHUCKLES] No, I think that's a really good analogy. MetaSounds is really the
culmination a lot of work that the Audio team has been
doing over the past few years. Because sound cues, which
are kind of the old system that's been around since early UE3 era,
is a little bit antiquated. There wasn't a lot
of things you could do with them other
than take a sound and then kind of modulate
the volume levels on it and do some very,
very simple stuff. But over the course
of the past 2 to 3 years, you've seen the new
Audio Mixer come in. That's allowed us to take
some of the audio processing and put it on to
other threads so that we have a little
bit more headroom to do cooler things with it. And so MetaSounds,
which they're kind of making their debut in the
Valley of the Ancient project, are really our attempt
to bring the power that you would get
on the rendering side from writing something like
a shader or a post-process into the realm of audio. So when you look at what you
can do with a MetaSounds graph, it's not just change the
volume of something, but you can actually do
sound synthesis on there. You can playback
samples on there. All the stuff that you'er
really used to working with, in the audio side,
with a DAW, you can do now in
the engine directly. So one of the cool parts in
the video that we premiered was the sound of
the Ancients' laser is partially synthesized
in game and then it's partially has some
samples overlaid on top of it. So it just really
kind of opens up, in the same way I think
programmable shaders and materials did for rendering,
we're kind of opening up those same
sort of workflows with audio. And then Quartz, on top of that,
is a system that-- I think, to simply put it,
it lets us talk back and forth between the audio that's
playing in the background and the sound effects
and the actual game world. So in the battle with
the ancient golem. The music in the background
is actually talking back and forth with gameplay so
that as the music starts to build and whatnot,
the golem starts charging up and animating. And then once we get
to a point in the song where it's like now would be
the perfect time for you to play that animation and start
the laser charge going, we can actually call
directly back into Blueprint to say, all right, go ahead,
play your next animation and do the attack. So it's really
opening up the door for a lot of really cool
audio features that weren't possible before
and a much tighter integration between things like the
foley and the music going on and the actual
gameplay systems itself. And I would be remiss
and Aaron would yell at me if I did not mention
it was sample-accurate as well. So if you want to start
making music games in UE5, I think now's a good time
to start playing around with those systems. VICTOR: And make sure
you let us know and show us what you're working on,
because that will be very exciting to see. CHANCE: Actually let us
hear what you're working on. Because it's really
hard to look at audio. NICK: I think one other
thing to mention too is, originally,
Chance was talking about, back in October and November,
which systems are we going to show in there? And we're like, OK, maybe
we'll have one or two MetaSounds as a sample in here. And then Dan Reynolds, who's
one of our technical audio guys, was like, we're just going to do
the whole thing in MetaSounds. And so it was a little bit
shocking and terrifying. But it's really,
really cool that we actually pulled off every sound in there
going through the new system. So really,
really happy with what the Audio team
has done with that. CHANCE: Yeah,
one thing I thought was super cool about
MetaSounds is how much data you can pass in
about your world. I mean,
it's kind of a theme in general about a number of
these new features, is that it's not just
a bunch of content that you bring in that
has a minimal set of kind of injection points to
pass data or do things with. It's just a lot more flexibility,
a lot more deep integration into other things. And yeah,
the Ancients laser attack is kind of a really
big one that's got a bunch of stuff
going on in there. But you can look at numerous
other MetaSounds in there that do very similar things. And we're already seeing,
again, in the community, people already poking around with that
and making some really, really cool things. NICK: Yeah,
and the laser example is kind of a simple one
we wanted to do just so it's easy to pick apart and
people can see how we make it. But there's no
reason that that's the upper bound on complexity. [INTERPOSING VOICES] CHANCE: Yeah, and I think,
too, talking to Aaron-- we'll get him and Dan on
at some point in the future-- but the widget library is
going to be just a bunch more modular effects
and different things you can do that you would
expect in like a sound library. And the thing will grow,
similar to the Blueprint API, over time. Super exciting. VICTOR: Sure is. Almost too much. But the show must go on. And like Chance said,
we will be covering MetaSounds and Quartz in
detail later on, in July, I believe. Check out the schedule. If you just scroll right below
on Twitch, it is right there. CHANCE: And real quick,
too, we said it in the script, but on MetaSounds,
I just want to make sure it's in there,
if you're not an audio designer or you don't understand audio,
this is an incredible point to try it out. Just like me--
I'm not an artist, but I can get around
in the Material Editor and I can make some
shaders that do something interesting in my games. MetaSounds brings a lot
of that same kind of power to audio generation. And so it looks very similar
as if you were to hook up things, like kind of ins and outs,
into like a stereo. And so if you can think in those
terms as far as node-building and stuff,
even if you don't understand audio or you're not a big audio nerd,
it is an incredible tool
just to go poke around with and learn from. We've had a lot of fun with it. Sorry, Victor. VICTOR: There's no
apology to be made. I mean, I even left my
camera here to get my box. Let's go ahead and talk
a little bit about animation. It is sort of highly featured
in Valley of the Ancient demo that is available. And there are a
couple of new systems here that we would
like to highlight. Simon,
I will give it over to you. SIMON: Sure. So we're introducing our
brand-new full-body IK Solver, which is significantly better
than our other older Jacobian solver. So it's inspired by
constraint solvers and rigid-body simulations. And it offers a much better
performance, artistic control, and with a simpler interface. We also have the new
motion-warping framework that allows you to
dynamically adjust the root motion
animations to adapt to dynamic or
changing environments it matches specific targets
that you place in your world. You can find examples
of that in the demo. And you can also find
examples about the slope warping to adapt the flat locomotion
to the rough, rugged landscape of the world. It's written entirely
in Control Rig. And it's using the
new full-body IK solver. Finally,
we have the blend space improvements that allows you to
blend arbitrary animation graphs instead of just
simple animation clips. CHANCE: Yeah, awesome. FBIK is used a couple of
places throughout the demo. It's again, a way we can
all kinds of meaningful data into the system and make system
A and system B work really well in tandem. For us, it kept iteration
time really nice for that battle sequence specifically. We were able to get
the results we wanted without having to
go back and author a bunch of new animations. We just had one base
thing that we wanted. And then we were able to
really refine that, iterate on that as long as we
needed to to make it feel just right for the battle. Super cool. Again,
these are tools for designers like me that aren't specialists
in animation or art or audio other things like that. I feel like I can go
in and do something really impactful here. And I feel like I don't always
need an engineer or an artist to go build new things for me
to be successful, which is great. It keeps disciplines working really,
really tight together too. There's been so many times
when it's like two or three folks would get in a room,
in the engine, and just kind of hammer
out the details of what we need to implement. And that's what
we end up shipping. And FBIK is one
of those things that's been really nice for that. VICTOR: And we have
Jeremy Grant and a couple of the folks from his team
back on the livestream in a couple of weeks to
talk about these in detail. If you are curious about
some of the features now, maybe it's worth pointing out-- and this is something that I've seen,
a question that's been coming up
for the last year, you can still use
features in UE4 to learn the tools that
will still exist in UE5. Obviously some of
them are not there. But in regards to when it
comes to our learning content, if you're new to Unreal
Engine and you're looking sort of for
basic tutorials on how to use Blueprints, for example,
or just our gameplay framework, you can still learn all of
those workflow frameworks, naming conventions,
et cetera, in UE4 and apply them in very
large quantities to UE5. And so don't feel like
you cannot go and search for questions and problems and
tutorials in UE4 and then apply it to UE5. You will find out and see that,
in many cases, the user interface and
the tools are very similar, except for the ones that
didn't exist in UE4, of course. But we are going
to cover and try to do our best to fill out
documentation as well as learning samples. While I'm on that topic,
Unreal Online Learning does have lessons and
tutorials specifically for UE5. So if you haven't
checked that out, it's learn.unrealengine.com. And you'll find it right
there in the Course section. CHANCE: Yeah, Whiting,
let's talk about game feature plug-ins and how cool this is
for designers and programmers working together
to not make tarballs. [CHUCKLES] NICK: Yeah,
game feature plug-ins are something that kind
of came out of the learnings from making a game like Fortnite,
that works on such a massive
scale with a massive team with a ton of gameplay
programmers And designers. And it kind of fits the
same role for programmers that the one file per
actor in World Partition does for level designers
and artists in that it's a way that you can kind of
keep pockets of functionality separate from one another. So if you look at the demo,
when Echo starts out, if you drop the Echo
Blueprint in the world, she doesn't do a whole lot
other than just walk in there. But during the
course of the demo, she's got two modes
that enhance her abilities. One of those is when
she releases her little moat drone at the beginning
that can fly around. And then once she
transitions to the dark world, she can throw the light dart,
kind of charge projectile, to blow stuff up and find
the Ancient at the end. And both of those features are
implemented as gameplay feature plug-ins. So basically what that is a
plugin that self-contained that has all the
assets it needs, all the Blueprints
scripting it needs, and any native C++ code
that it needs that has a little descriptor file that you
can open up and say, OK, when I load this plugin,
I want to stream in this content, add these components to
these types of actors and whatnot. It really lets you create the
scripting and functionality orthogonally to the
actual Blueprints itself. So if I was working
on the drone moat and chance was working
on the light dart ability, we wouldn't be stepping on
each other's toes necessarily. And the other
nice thing about it is it means that it's
more easy to transport that kind of functionality
between different projects if you want to because it's
all kind of encapsulated in that module itself. So it's not the sexiest
thing to show off, because it's just a
bunch of menus and kind of an organizational thing,
but it really kind of forces us to do a lot more
implementation of functionality in components that can
be added to Blueprints rather than in just monolithic
Blueprints themselves. So it's a little bit of a
technological feature and a little bit of a
philosophical shift at the same time. What we found was really cool-- and we'll have Michael Noland
and I think Mike Beech on to talk about it later-- is that we were able to
pipe a lot of the functionality through the gameplay
ability system. And with that,
we're able to give players new gameplay abilities
with a lot of this content that, again, doesn't exist
in the core game at all. So there's a one-way reference. And using the gameplay
ability system, we're able to say, this turns off these other
things in a tag-specific way. So the main game might manage
some tags for specific things. But that allows us to make sure that,
if Whiting's working on something that
takes these input controls, and I need those and we know
that they're never going to be on at the same time,
I can just make sure that that whenever the game
is appropriately loading this, it turns those things off. And so it's a really
powerful system. It's funny, we built the drone--
we pulled the drone straight out of the [? Kite ?] demo. And we kind of
repurposed that a little bit. We knew we wanted to explore
the space with something. And we kind of wanted to
see what we could do there. And so the C++ and the
Blueprints and the VFX and everything for that is and
the Hover Drone plugin there. And then the actual
light dart blast, energy blast she's got,
we originally built that as a separate game
feature plugin to just enable to give her a new ability. And then we kind of
took it a little bit further for our demonstration. If you see, it's called the
Ancient Battle plugin now too. And not only does it
turn all of that on for her, but it also adds the
destructible pieces, all the targeting stuff that
the actual attack needs, and it brings in the actual
ancient for you to fight. So it might be like
in a world where you've got a map for some time where,
oh, there's this place that some future
content is coming down the pipe but it's not ready yet. It maybe a limited-time thing. You're timing this with
some event in your game. And then so you kind of
release this game feature into your game. And you enable it
whenever it's ready. And then, that way,
now that it's enabled, your player can do things
that they couldn't before. But it's not like
you had to release a whole new client build,
so also their new stuff with it. It's something that you can
just enable separately and go from there. I think we've learned a lot
using this feature in Fortnite, like Nick mentioned,
to see how do we not build that, where you have a
character, if somebody needs to check out the
character asset, and because of that there's
all these other things that it has to touch to receive or vice versa,
if I have to save the light dart, well,
since it references Echo, it thinks it needs
to resave Echo. So I've got to make sure
that I got to do that to. And Nick happens to
have Echo checked out. And he went home for the weekend
because that's what he does. NICK: During a pandemic,
I'm always home, Chance. CHANCE: That's right. Oh, yeah, sorry,
you went upstairs for a week. NICK: I did. I did. I think you did touch on
something else that was-- folks that are cracking open
the Valley of the Ancient project, the interplay of the
game feature plugin and the gameplay ability
system is really kind of important. And if you're used to traditional,
super-deep Blueprint code,
it looks like there's a lot of indirection
going on now. But really that's kind of the
philosophical portion of it, where we build things using
gameplay abilities and gameplay tags and whatnot to keep
them kind of generic at the base game level. And then we can layer
that functionality on top of it. And that's how we can
have a generic character at the base that has
their animations extended and their abilities and
their input extended. And we can make sure that,
if you load the hover drone ability and the battle ability
at the same time, when you pull the trigger,
the right thing is going to happen. It helps you kind
sort out what's going on the input
stack and things like that. So those two really are
important to think about in tandem, I believe. CHANCE: Yeah, and it works
with the input action system too. So that makes it
really easy to just find new input
actions/remove new input actions from pretty much
anything you want in there, just to stack components
and actors in your world. It's pretty great. I'm a big fan. VICTOR: I think we're at
eight pages of questions. And I will let everyone know
that we are gathering them all. And thank you so much
for sending them in. We will not be able to cover
all of them in detail today. But we are gathering them
and we will prep all of the guests, in the upcoming week,
to go over some of the questions that you have. But we still have
some time to go. So perhaps you'll
get lucky and we're able to catch the ones
that you've been asking. A lot of good ones. All lot of good ones. Something I've seen
come up over and over and over again that I think
it's worth covering real quick, I'm seeing UE5 and VR and AR. What is the state of
stereo rendering and all of these new shiny features? Is there anyone on
the call who would like to tackle this question? MICHAL: I can. So Nanite,
in the early access, is not is not supporting
multi-view rendering. Is it something we'll
be adding soon? So unfortunately,
the shiny new feature doesn't work. Lumen works. VICTOR: All right. And there is a new VR template. So you can go ahead
and check that out. [LAUGHTER] CHANCE: Just a shameless plug there,
Victor. VICTOR: Shameless plug there,
yeah. NICK: What would you know about that,
Victor? CHANCE: Yeah. VICTOR: No idea. CHANCE: Oh,
my lights came back on. That's good. NICK: Fantastic. CHANCE: I'm like,
the only person here. And so every once in a while-- NICK: We paid your
power bill for you. CHANCE: Oh thanks, buddy. Appreciate you. Hey, I see the next thing
we have on here to discuss is with animating in the engine. We've had control work a bit. Simon, you want to give us
the TL;DR of where we are there? SIMON: Sure. I mean,
it's pretty awesome to think that the engine was all
completely animated in-engine. No external DCCs
were used for that. That's pretty cool. CHANCE: Very cool. SIMON: Obviously it's
all set up with Control Rig. Echo is also set
up with Control Rig. So you can download the
demo and play with that. Basically,
animating an engine, we're using Sequencer in Control Rig. And we have a few tools like
a Pose Library, a Snapper tool and a Twin tool. With the Pose Library,
you can create reusable poses and then apply
them to your libraries. So you want the end for
the robot and things like that. With the Control
Rig Snapper tool, it can pin objects or
controls from your Control Rig to other objects in the world. And it makes keyframes
over a duration of time. CHANCE: Oh, that's right. SIMON: Yeah. CHANCE: Yeah, that's awesome. SIMON: And the Twin tool, you can create the new in
between keyframes and weights and weight them
based on the keyframe or pose that is either ahead
or behind the current keyframe. And Control Rig
also now supports encapsulation and the ability
to share and reuse modular rigs, if you have function libraries. CHANCE: No, that's awesome. Yeah, and you touched on
it for "Valley of the Ancient." You know, The Ancient is
100% animated in-engine, all of it. Everything he's got there,
using Control Rig, which is a super cool. Working with Jeremiah, he showed
me how some of the tools work and how it's used. I've never used it before and again,
I'm not an animator. But I feel like I can go through
and do some things there, that I could not do
before and at least try out ideas and whatnot,
using these kinds of tools without having to load a
different piece of software, which I don't have a
license for nor how to use. SIMON: Control Rigs is really,
really make it easy to do and play
around with the character and if you had on there. CHANCE: Yeah. SIMON: Very powerful. CHANCE: Yeah, the FBIK
node is basically a Control Rig node, right? That you can pass
all your data into? Yeah, super. Super powerful, man. Great, great stuff. Wanted to call out to
that's the high-level number, or I guess a high-level list
of features that were UE5 specific, that we wanted to
make sure to showcase well in Topaz-- I'm sorry,
in Valley of the Ancient. I mean, it's 5 now. [LAUGHTER] I just wanted to
also call it too, that there's a number
of UE4 features that we use that have
really good representation in the project as well. So if you are new to UE,
or if you want to see how some
of these things were done, there's some good examples
of other things in there. Like, White,
are you going to talk a little bit about Destruction? NICK: Yeah. With the destruction
that's in the demo plugin, we wanted to have a few pieces. And it was really
important to us to put it through its
paces running on console, in terms of performance
and memory. Because I mean, it's doing
destruction that has Lumen-- or sorry, not Lumen,
Nanite assets within it for the destruction bit,
so that they actually match the environment. There was a lot of time
and a lot of hard work, both by the Content side of
the team in the Simulation team, getting that up to speed. I think we went from
gigabytes of memory in use on the console to have
those pieces in there to much, much, much less memory
that properly streams in and out. That was, I think,
a huge behind-the-scenes piece that I think we'll be seeing more
and more of as UE5 becomes more mature, now that it's
kind of up to snuff on consoles. CHANCE: And a really cool
way to take a look at that too, is if you load up
the Dark World, you can open up the robot,
the Ancients' sequence for emergence. You can kind of scrub
through it and watch all the individual caches
break and whatnot, and, take a look at the
assets specifically from there. Some really great
examples in there. Same thing with Niagara. I mean, on the Ancient,
who's got the big old spinning red
volumetrics in his head there, that Ryan Brooks put together for us,
which was super great. As well as the fire,
there's some really great things in there. I haven't spent enough
time with Niagara yet. It's the first place
I'm going to dive in, just to see how we made
some of those things work. As mentioned before,
a gameplay ability system is peppered
throughout pretty much anything we could get onto for Echo. I've seen a number of questions
scroll through chat about it. If you're interested
in learning, we've got some really
good docs on it from UE4. And you can see it in use with
these other gameplay systems that we've got. Of course, too,
Blueprints are still here. We're using them for pretty
much everything in here. So if you're used to
digging around in there and want to see how any of
this is set up, on the Ancient is one big Ancient. Or I'm sorry,
one big Blueprint called, I think, BP Ancient. So you
can find that pretty easy too. NICK: And a bunch of
gameplay ability Blueprints. I saw in the chat,
they were saying how does Echo play different animations and stuff? CHANCE: Yeah. NICK: It just doesn't
happen in the base game. If you look at how the gameplay
ability Blueprints are set up in there, that's how we do it. CHANCE: Yeah,
I think there's like, what, the Dodges one? The Walk Overrides one. NICK: Mantling is one. CHANCE: Light/Dark one,
the Vault, yeah, the Mantling is one. NICK: The touching
of the Dark Rift. CHANCE: You interact with that,
that's right. NICK: Mhm. CHANCE: We tried to go ham. NICK: We went totally ham. VICTOR: Yeah,
I was going to say in a lot of areas. Not only animation. CHANCE: Yes. Yes. NICK: It's a very
ham-rich project. CHANCE: Yes. We've talked about a
number of these features that are well represented in
the project in different areas. I'm sure there's a
handful of other things that are in UE5 that
might not be here. Is there anything,
just want to make sure there's anything
we haven't covered, we can before we
move onto what's next. VICTOR: Still
gathering questions. CHANCE: OK, yeah. VICTOR: I don't even know
what page I'm on right now. CHANCE: It's all good. I just didn't know if there's
anything, it's like, oh yeah, we didn't have the
appropriate time to highlight it here
or actually use it. VICTOR: We're
doing good on time. We have gone through
a whole bunch of it. So I think we can just continue. And then we'll have a good
amount of time for questions. CHANCE: Yeah, sounds good. Nick Penwarden-- I have to
specify because we have so many Nicks-- what's the road
to 5.0 look like? Are we going to do hotfixes? Any supplemental releases
between now and then? What can developers expect? NICK: Yeah, the road to 5.0, we are looking at putting
out a hotfix for early access. But what we want to do is
gather up crash reports, feedback, stuff like that. Focus on getting stability
up in early access a little bit. And seeing what
iterative improvements we can make over the next couple
of weeks, maybe next month. Then our focus is going to turn
towards the 5.0 release itself. So that's all going to be
on taking the features that are in early access,
continuing to improve workflows, continuing to have a
really big push on polish, on optimization,
on stability, to really get those buttoned up and
ready to use in production to make games. The other stuff that
we have in the future is we're going to be shipping
Fortnite on top of UE5. That technology change is really,
our goal with that is to use it as a proofpoint. Like, UE5 is ready. You can ship a game on UE5,
because we are shipping our own games on it. That will give us a
first nice proofpoint of shipping Fortnite on it. Then we'll get
it all out to you. CHANCE: Awesome. Anything in particular from
the engineering directors here? Specifically for what's
next on the way to 5.0? Anything you think
we can talk about that's going to maybe
land at that time? There's a handful of things
that we've mentioned in the past. But just anything that's
not already covered? MICHAL: I think I
can add a bit more. We're obviously polishing
and making the features easier to use. Specifically to
call out I think, Lumen, and the extending of
the hardware ray tracing support is something we are
still aiming for 5.0, I think it's going to be a
great alternative for people. It hasn't been mentioned,
but we are working on making platforms,
the non-Windows platforms, easier to work with,
easier to iterate on. But that's the whole effort
we're pushing forward. Things like Vulkan
ray tracing might come. We're working on optimizing
the renderer, and so on and so on. CHANCE: Right on. NICK: I'll throw at that the
MetaSounds' workflow is they're working on just
making it way more feature rich right now, and adding more
visualizers and nice debugging and workflow sorts of things
to really help you implement MetaSounds,
and help people get up to speed with how to use them. I know they're very
keen to see what people are doing with them publicly. That would probably
affect which kinds of features that they're adding to it,
based on what demand is. Post your stuff,
send it to Aaron and to Unreal Engine on Twitter,
and I'm sure Victor can
help us aggregate. VICTOR: I was going to say-- Aaron
McLeran, Dan Reynolds, @unrealengine. Those are the three
that you want to ping. Adam Block too, actually. CHANCE: Yeah. On that note too, it's funny,
working with MetaSounds, it's very reminiscent-- Whiting, you might be able
to speak to this-- of Blueprints at 4.0, right? Where there's really
great functionality in there. There's a lot of
really cool little things that have been added
to Blueprints over time, that I think that they
want to bring in there too. Like re-route notes, for instance,
and things like that, to really make it nice
and clean and neat. And yeah. Go. NICK: Yeah,
Blueprints were basically a result of a lot of
people walking over behind Nick Donaldson's
desk and seeing what crazy stuff he got up to,
and then making features that help keep him happy. So I think the same
thing will happen with MetaSounds,
where it's really based on what people
are trying to do with it, is where we spend our time. CHANCE: Yeah. Cool, cool. VICTOR: Cool. Are y'all all ready to tackle
some of the questions that we received during the
last hour at 17 minutes? NICK: Yeah, let's start the Q&A. VICTOR: I want to hear it. Great, yes. Yeah. NICK: Let's do it. NICK: It looks like
there's one or two, maybe. VICTOR: People have questions? Let's kick off with one question
that came in quite recently, from AE Surowka
"Is the preview going to be updated throughout
2021 until the full release? Or is this the only version we
have access to until the 5.0?" CHANCE: Yeah, I can cover that. I mentioned briefly,
we are going to look at getting a hotfix out. Maybe one or two,
trying to similar to preview builds as we release them,
before UE4 releases, looking to get stability up,
looking to fix the very common problems
that people are running into. But after that,
the next release will be 5.0. VICTOR: And you can
follow us on GitHub as well, if that wasn't mentioned. I was reading questions. CHANCE: Yeah, I actually
sent a tweet out about that earlier too,
main is on GitHub too. It's really cool. I don't know how many
people had recognized that. But we have not just
early access build. But if you really want
to go like into the depths of the dragons and
fight them with us, you can live on the absolute
edge with main as well. NICK: Yep follow along
as check-ins go into UE5. You can see it all live. CHANCE: I put a tweet up. And I was like,
I wonder what the last check-in was. It was like, an hour ago. I was like, OK, yeah,
it's getting updated. So pretty cool. VICTOR: Next question
comes from loraesh. Who's wondering
"What's the road map for tessellation
replacements for things that Nanite does not support?" MICHAL: We are working
on a knowledge-based article that will cover this in detail. But we intend to replace
the tessellation with support within Nanite. It's still something,
some work to come. I, unfortunately, cannot go and
say we have a roadmap for it. Because it's a bit of an R&D.
It's a bit of exploration there as well. So yeah,
it is something on our mind. And something we
are working towards. VICTOR: Next question
comes from ChickenWeed. Oh go ahead, Chance. CHANCE: Oh no, it's fine. I'm tracking you and the dog. So following. VICTOR: Oh. [LAUGHS] Next question
comes from ChickenWeed, who's wondering,
"Will Nanite support transparent and animated
meshes in the future?" MICHAL: Absolutely, yes. That's our goal. We want to make sure that
every part of the pipeline that deals with meshes will
be supported by Nanite. So we are looking
at translucencies, at animation,
we would like to get, as I said, tessellation,
displacement working. VICTOR: Yeah,
I remember another question, a little down further below. Don't remember who asked it,
but someone was curious-- oh, here we go. From RonCon,
who's wondering, why did you implement your own rasterizer? MICHAL: I think this
will be best in detail answered by Brian,
in the next livestream. It's a very passionate
subject for him. But I think the short
answer is because it enabled us to make Nanite. It's somewhere where we
were able to move much further forward than anybody else. NICK: Yeah. I mean, two words, it's faster. But Brian can expand
on that in more detail. VICTOR: Tune in next week,
same time as today, same channel. We'll have Brian Karis. CHANCE: We got one here
from AlveroRealtimeMayhem "I noticed the
gameability system was used in the demo project. Does this mean we should
expect newer updates to it?" Wanting to know
the future of a GAS. NICK: Yeah, you know,
game playability is something that
kind of got incubated on Paragon and Fortnite. Then this is really
Valley of the Ancient, because we were using
the game feature plug-ins, was really the first
time we were trying to productize it in the engine. But we've got some
efforts spinning up right now that'll come
out around the 5.0 launch. That will really put
it through its paces and make some more
complicated examples. And based on the learnings,
we will definitely be updating that. CHANCE: I can grab
this one right here. This is from thc_ilia. "Are there any plans
for a tutorial series for Unreal Engine 5?" I think just as we move forward,
we should expect everything,
same behaviors we had around UE4 throughout its lifecycle. That's not just a continued
support release, over release, with new features
and functionality. But that comes
along with us being able to put out resources
for pretty much anything that we build and
add to the engine. So yeah, I would say
it's pretty safe to say so. Then another one from Anteaus,
"How often do you plan to release updates
and fixes to the release?" I think we had
already discussed what we're going to do for early
access and between now and 5.0. But Nick Penwarden,
this is an assumption. I assume it would
be very similar to what people expect from UE4 as well. NICK: Oh, sorry? Say that again? CHANCE: All good, that's good. NICK: I cut out for a second. CHANCE: You were texting like me,
earlier. It's totally fine. Nah, I'm just kidding. "How often do we plan to release
updates and fixes to release?" I know we talked about
early access to 5.0. But I'm assuming that 5.0,
going forward, is going to be
very similar to what people can expect with UE4 now. We're just going
to kind of continue that trend, as we move forward. NICK: Yep. Yeah, that's the plan right now,
is once 5.0 gets out there, we'll be releasing frequent
updates throughout each year. So yeah, expect a release
cadence similar to UE4. VICTOR: In regards to Nanite and this new workflow, just,
we're going to go back to this. I know Brian Karis is coming on,
but there are a lot of questions. And I think this one
is worth covering. When you now import
movie-quality assets to the editor,
you are left with files that are very large compared
to the traditional much lower-quality workflow. What are we doing to
make this workflow normal and easier to work with? MICHAL: There are a
bunch of technologies underneath the surface. I don't know if people tried it. But if you try to take a big
mesh and turn it into a Nanite, you actually notice that it's much,
much smaller than the original mesh. We have really good compression
underneath the surface. We're still working
on improving it. We're looking at how we
can make the assets smaller by obviously not distributing
things you actually do not need to distribute. So it's a work in progress. But I think it's
worth pointing out that the Nanite mesh
is because of what our rasterization allows,
are so compressed compared to the traditionals that you're
probably actually saving space. I would add,
on "Valley of the Ancient," too, the bigger offender,
in terms of this size, was the texture,
stores, because you need so many high-res
textures to fill things up. But we did squeak in storing
the original assets as JPEGs. Because a lot of the photos
scanned assets from Quixel were JPEGs anyway. There's no need to
store them lossless, as the raw data assets. So that's a feature that made it
into early access and saved us a ton of disk space. I probably shouldn't say it,
but a month and a half ago, we were at like, what,
320 gigs for the project, before we cleaned everything
up and compressed it. But that does show you,
the JPEG compression was a huge portion of that too. NICK: And goes to
show how much texture was a part of that. CHANCE: Yes. NICK: Also-- CHANCE: Oh, sorry. On that note,
on "Valley of the Ancient," if you look through the content,
I think it's something
like 90% of our content is that folder, right?
it's the Megascans folder. We were able to delete I
don't know how many normal maps, to cut all that out too. So we're using fewer
textures in general. We probably could have
been a little bit more deliberate over time,
to pick the actual resolution of each of these textures
that we actually need for specific spots. We also have a very broad
array of different objects that are in there. So there's very many, many,
many different unique meshes that require unique
textures as well too. So probably some optimization
to be saved there as well. NICK: The last thing I'll say on that is if you check
out the Nanite documentation, there's a pretty
good analysis there, of disk footprint and
memory size of a Nanite mesh versus a traditional
static mesh, as well as some thinking
about what textures you need and how that
potentially impacts size. So it's a good
place to go to get an idea of what you can
expect using film-quality assets in your game,
how it will affect this footprint for your project. NICK: I think it's worth
everybody hounding the Quixel guys when they
come on the stream. But they already
have a lot of learnings based on what they did here. And they think they
can get away with using a lot fewer textures,
and more detail textures, that are reused all around
at higher resolutions to really chop
down the footprint that we just didn't have
time for on this project. So it'll get a lot better
and we'll have a lot more best practices in the future. NICK: Yeah, that goes back to the "Valley of the Ancient"
being an experiment for us. Right? This was us learning how to
try to build this type of content. With what we know now,
I think we'd built the environment
a bit differently. And we'd be able to
do it more optimally and achieve the same results. So I think it was is an
awesome project for us to get an idea of how to
build content with this tech. And as we learn more and more,
we'll want to share that with you all. And as you explore the tech
and have learnings, please share it back with us. CHANCE: Exactly what
I was going to say, yeah. VICTOR: Go ahead, Chance. CHANCE: No, I'm just saying, that was exactly what
I was going to say. There's so much that we
learn through the process here,
that we hope to be able to share. VICTOR: Something
that I didn't actually know before we released it,
but some folks in the community packaged the demo. And it turns out that the
package billed is 25 gigabytes. Which is quite the change
from the size of the project itself. Add everything that comes
with optimizing in performance, is usually what comes after
the tools have been built. So there's a lot to come there. I hope everyone's
excited to participate in it. Mayne on GitHub, be vigilant. NICK: Yeah. It's also worth pointing
out one of the reasons for that size difference is that
when you download the project, you will get both the
original source assets, like the source JPEGs for
all the textures, the raw source triangles for all the geometry,
as well as the optimized versions. That way your first
experience opening the project isn't processing
all of that data. So you're effectively
downloading what's in that final package,
what you would shift to players,
once you've made your gamer experience,
as well as the source assets themselves. All of that is encompassed
in that initial download. VICTOR: On that note, there have been a few
questions in regards to hardware,
and to what kind of computer do I need to work
with Unreal Engine? What kind of computer do
I might need in the future, to play Unreal Engine games? I think we can
touch a little bit on the difference between
"Valley of the Ancient" and Unreal Engine 5 in general,
when it comes to the plethora
and range of hardware that exists out there? NICK: Yeah. I guess I can speak to that. So with "Valley of the Ancient,"
we were really targeting
how do we push the limits of the technology
that we built on the high end, for next-generation consoles? We were targeting
next-generation console hardware with this demo. And we really built all of
the content with that in mind. The short answer
of what hardware do you need to
develop with UE5 is, depends what type of
content you're wanting to build, and how you want to build it. When you want to build a game,
that's going to say scale across
console generations. You can still take
advantage of Nanite and take advantage
of its capability to render extremely
high-polygon meshes. You just need to build your game
content with the idea that hey, I'm going to scale
across platforms. When we were getting Fortnite,
which was originally made for consoles running on mobile,
part of that effort was figuring out how
do we scale down? Or how do we take that
content and scale it down to run on mobile phone hardware? It's kind of the same idea,
when you're going to approach this, is really,
you approach building your content by thinking about all
the platforms that you're going to want to ship it on. And then take
advantage of how you can scale up and down that
content on different platforms, to optimize the
final results for it. CHANCE: Then, I would even say, we've been working the
"Valley of the Ancient." I mean,
game development is a series of those kinds of compromises. You choose where to spend that
budget based on what you think is most precious to
you and your game, whether it's frame rate or visuals,
or reflections. So you might spend some
points in bucket A. Really, a lot of it comes down
to being content specific. VICTOR: Thank you both. All of you. Thanks to all of you. [LAUGHS] Without you,
we would not be here and there would not be
UE5 available right now. Putting that out there. Moving on. NICK: And The entire
Development team. We're just a couple of
people who helped make this. VICTOR: When I'm saying you, I guess it's just all of
you developers out there. There are more. Believe it or not, this is not
the only one folks who worked. CHANCE: Yeah,
I'm just the guy they asked to talk. MICHAL: It's the royal you. VICTOR: Same here,
just the face. Like I all the
intricacies in UE5 yet. Trust me,
I've not had nearly as much time working with the tools. I see all of you out
there working with it. I'm getting a
little bit jealous. Because we're in the
weeds with everything that's involved around community
and preparing all this for you. But eventually,
we'll all get there. It's a journey that we're
all traveling together. Next question comes from o_dz_o,
who's asking, "Will there be other
sample projects released? "Valley of the Ancient" is
good as a complete sample. But it's quite overwhelming
to jump straight into it. Other sample products
that are less complex would be good for me. CHANCE: Funny,
that Penwarden and I were literally discussing this,
I think what, four days ago? Right? And what the plan is next. We don't have anything really,
to share at this point. But we do recognize that
there are numerous ways that we can provide
value to folks that's not a high-end example project. There are plans for 5.0,
for a number of the different various ways
that you would get access to examples,
such as content examples. Or other sample projects in
the works that are on their way. So stay tuned for
more information there. But it is something that we're
discussing in the shorter term too. VICTOR: And I would
say that we can probably expect a lot of content to
come out from the community? CHANCE: Oh yeah. VICTOR: Just,
I've had several requests from folks, tutorial creators
around the world, who is wondering if they're
allowed to start creating content with UE5? And they are. There is,
the same end user license agreement applies to UE5 as
it applies to Unreal Engine 4. So go ahead, produce content. Make all the crazy things. And we'd love to see it. Next questions come from JB,
who's wondering, "How easily do plug-ins
transfer over from UE4 to UE5? Should we expect to see
some plug-ins updated to support UE5 soon?" That's going to be multifaceted. I think maybe it needs more,
a little bit more info. There's UE4 engine
plug-ins that are in UE5. There may be some
that made the jump and may be some
that we plan not to. Other plug-ins to be
updated to support UE5, that might be things
like Marketplace or whatnot. Nick, Penwarden,
are there things that are in Ue4 right now,
like plug-ins or systems that we know, that we haven't
yet ported up to UE5 yet? NICK: Well,
everything that we ship with UE4 is available as part of UE5. Marketplace
plug-ins are different. So individual plugin
developers are going to have to take
the time to update their plug-ins for UE5. It's another thing
that I've noticed, as I've been trying
to see how developers are working with UE5. And a number of
plug-ins developers are noticing that they
were very able very easily able to get their plug-ins
up and running in UE5. So I hope more and more
plugin developers have that kind of experience where it's,
more or less, easy to get them translated over. Depending on
the specific plugin, it may be easier or
more difficult to get it running in UE5. But similar to getting projects in,
I mean, the goal is that upgrading
shouldn't be that much more difficult than updating
for a new UE4 release. CHANCE: Cool. VICTOR: The next question comes from Polter_,
who's wondering, "What are your plans for
the Water plugin? Will there be any changes
in Unreal Engine 5?" NICK: Yes, for 5.0, we want to make the Water
plugin production ready. Right now, I don't remember
if it's experimental or beat, but it's not
production ready yet. One part that is missing
is the making water work with the World Partition. So splitting the
water across the grid. So that's in the works. Yeah, just quality of life
improvements, and another thing that we want to tackle
is the brush limitation. Which right now,
the brush is for the whole map. If the map gets too big,
then we run out. The brush basically
makes the engine crash. So we want to have a
better approach for that, so that we can scale
up to much larger worlds. VICTOR: Thank you, Simon. Next question comes from
Richard Sun who is wondering, is there an established
pipeline to convert UE4-lit scenes to convert towards
and to support Lumen? MICHAL: That might
be a better question for Daniel, who will be,
I think, live in two weeks? CHANCE: That's right MICHAL: Two weeks. But right now, I think I'd
summarize it the way I read it, somebody tweet about it. You go in. You disable Lightmaps. You turn on Lumen. And it looks awesome. I haven't tried it,
but people tweet about it that way. CHANCE: Yeah,
it's just a dropdown in the Project Settings at this point? MICHAL: Correct. VICTOR: It's enabled by default,
in all new projects. Which,
if you are using the VR template, you actually do have
to disable Lumen. Because otherwise,
the lighting that's baked there does not display it
in the correct way. This will be fixed. This will be
hotfixed in the future. Some special
template shenanigans that had to be done to make
that work when you create a new project,
that it's default and it shouldn't be in that project. Just to note. Next question comes from
Mateo Merchan Lopera, who is wondering,
"Can it be possible to produce a full video game only with
Blueprints in Unreal Engine 5?" NICK: I would say
as much as possible as it was in UE4. I mean,
there might be some rough edges if you have to cert on console,
potentially. But we've had people ship
games that are Blueprint only. So there's no kind of
fundamental limitation to it. We tried to use as much
Blueprints as possible in "Valley of the Ancient,"
just to show it can be done. And I always harp
on this example. But Robo Recall,
which I noticed Victor, you are repping back there,
was about 85% Blueprints. There's no limit on
performance and whatnot, as long as you are diligent
about what you do and write good code in Blueprints
the same as you would write good code in native code. There's ways to get around it. So I don't think there's any
fundamental limitations for it. If there are, we should fix that,
so let us know. CHANCE: Working with
folks over the UE4 lifecycle, there's numerous games
that launched on Steam, right? Or other PC stores,
that are Blueprints only. Typically, yeah,
when it comes to console, generally the
native aspects there are operating with SDKs mostly,
right? Things that the hardware itself
needs that native layer for you to attach to you. But yeah. VICTOR: Moving on
to the next question comes from ShadowAimai,
who's wondering, "How big can you make maps now?" SIMON: Well, the limitation comes from the
floating point precision. So 10 kilometer by 10 kilometer,
you're going to be fine. Basically, if you move away
from more than 5.2 kilometer from the original, you start
getting to the world position world. You can have a
16-by-16-kilometer world if you want. And if the outer rims are
more like fake meshes that you don't have gameplay,
that you can't navigate in that area. So 16 by 16 would work. But yeah, 10 by 10,
if you want to have gameplay. CHANCE: Yeah. On that note too,
what we shipped in "Valley of the Ancient,"
we did 2 by 2, is the place that we focused on. But we had artists hand
dressing a lot of that, for a dozen weeks or so. With more people and more time,
we could've done the same thing and gotten much larger,
even with that process there. Because of streaming
systems like World Partition. Being able to get what we
don't need out of memory and what we do back in. Wasn't a limitation of the tech,
at that point. For us, it was just
limitation of the schedule. VICTOR: Next question
comes from Simon Finnie who's wondering,
"How important is fast-access memory,
such as in the PlayStation 5 solid-state drive,
for performance with this Unreal Engine 5 technology?" NICK: It certainly helps. Whether you're streaming in
content for levels or streaming in geometry with Nanite,
streaming in textures with Virtual Texturing. The more I/O bandwidth we have,
the more that you'll be able to
stream in that detail without noticing the difference. However, the technology
does scale surprisingly well. I actually have
"Valley of the Ancient" on a spinning-disk hard drive. I've been playing
around with it a little bit. It actually works
surprisingly well. MICHAL: Yeah. I think it's worth noting that
virtual textures in Nanite, for example, they are much
more gentle when it comes to the amount of streaming. It's not like you move and then
you get megabytes and megabytes of data suddenly, in one frame. It's much more continuous
and much more fine grain. So actually, yeah,
it's working pretty well on all kinds of I/O. VICTOR: Next question
comes from Biotorial who's wondering, "Are landscapes
planned to be supported by Nanite?" MICHAL: Huh. They're planned to be
supported by Lumen. That's what my answer. NICK: I think Brian will
have opinions on that. MICHAL: Yeah,
Brian will have opinion on that. But yeah,
it's certainly working on the demo. We were looking at
all the different ways we could shape this technology. So it's one of the
things we need to take into account,
as an experience coming from the demo. CHANCE: Yeah, currently
the virtual height field work would be the way to scale
up geometric complexity on the terrain. But we're exploring all
sorts of options for the future. VICTOR: Next question
comes from Ryan VanMeter. Even though this
question is specific, I think we can answer
a little bit generically. The question is, "Is multi-user
editing functional or planned? I'd like to start
prototyping with a friend, and it would be cool if
this UE4 feature made it into this early access demo." Other than the clear
deprecated features that is listed in
our documentation, unless it's there,
these features will be made available in UE5. If they're not
currently working, they will be in the future. So if you're curious
what sort of was in UE4 and what will be in UE5,
go ahead and head over to our documentation,
where you can see the entire list of features
that have been removed. If any of you know specifically
about multi-user editing in UE5, feel free to chime in. But I have not yet tried that. SIMON: Yeah,
I have not tried it either. But as far as I know,
it should be working. NICK: Chance, you need
to pay your power bill again. VICTOR: I know. CHANCE: Yeah,
I don't know what's up there. Need to have like, a ceiling
fan or something around here, always moving. VICTOR: Let's see, next question comes from AE Surowka,
another Nanite question. "Vertex painting on
instant Nanite meshes, coming at some point?" MICHAL: Yes. Although I am not
sure if we will call it vertex painting at that point. Nanite meshes are huge,
in terms of the amount of polygons and vertices you can make. So we are not yet sure if
what we will be doing soon is going to be precise vertex
painting or something more volumetric. We made some
experiments over last year. And it looked very good. Now we need to wrap
it up and productize it. So yes, it's something
we actually are working on. VICTOR: Next question
comes from MOA Development, who's wondering,
"Are there limitations to the density of
meshes when importing?" MICHAL: Not really. We had meshes that were
tens of millions of polygons. So I mean, at some point,
people can run out of memory somewhere. I would encourage
everybody to report the cases where this doesn't work. VICTOR: The user did note
that they have noticed crashes when trying over
5 million triangles. MICHAL: That sounds-- that's unusual. It would be great
to have that mesh and test it out here,
to see what's going on. VICTOR: MOA Development,
you know where to find us. Send us your Crash Log. Next question comes from Handkor. Handkor's wondering,
"Are Marketplace C++ plugins able to publish early access versions or do
they have to to wait until release?" For now, we have no choice in
regards to Marketplace support. But stay up to date
on our Twitter account, as well as the forum thread. And we will update you. Right now, UE5,
or the access, it's meant to be played around
with and experimented with. But feel free to
prepare your plug-ins. Make sure that
they work and there's no limitations for you to
not be able to distribute them elsewhere. CHANCE: I've got one here,
Victor, that I'd like to see if
Penwarden can answer. "What is the plan for future
updates to UE4, for now?" NICK: Yeah,
I can talk about that. We will be releasing
4.27 later this year. Currently that is the
last planned release that we have for UE4. We'll play that
by ear a little bit. If there is a good reason for
releasing a small update to you UE4,
say for SDK updates or something like that,
that's something that we'll consider. But our plan is to jump
right from 4.27 to 5.0. CHANCE: Yeah. Also on that note,
if you still are on 4.0 and you don't want to move to UE5--
or 4.27 and you don't want
to move to UE5, since the source is available,
which it'll remain available, if you need to do SDK
updates or things to support live games that are
already shipped there, that option will be available
to you going forward as well. Right. VICTOR: Next question comes from Sonali Singh,
who's wondering "If the Control Rig sample
project is available in UE 5.0?" I believe this is a reference
to the livestream we did with Jeremiah Grant. We weren't able to
say this after the stream. I know we did
mention that we were going to provide the
slope-warping example project. The reason why
we couldn't do that is because that
was being developed for the UE5
early-access release. So the "Valley of
the Ancient" demo actually contains a
much better version of the slope-warping example. So that is essentially
the example project for slope warping. Was it motion warping now? Am I saying that right? CHANCE: FBIK handles
the slope warping and motion warping,
as for the motion animations that are driving those. VICTOR: Next question comes from Emomilol1213. I've seen this, a couple of
questions in regards to this. They're a little bit confused
on the deprecated tag on ray-traced reflections,
et cetera. assume that hardware ray tracing
would run faster than software. Is there some hidden
toggle to use RTX in Nanite? Could we go into
details a little bit, what is going on there? MICHAL: Oh, yeah, that's
probably didn't come out right. The reason why it's
marked as the bracket is because Lumen is taking
over that part of the pipeline. You can use it still,
the optimized reflections that are
coming from UE4. But Lumen, Lumen is
integrating a better version of it. So as we work on the
hardware rate of sync support and extending it,
it will fully replace the exclusive reflections part. NICK: Yeah. Maybe a different
way to think about it is that Lumen is a
system that is combining a number of
technologies to provide indirect lighting of all forms. Whether it's GI via indirect diffuse,
or reflections, you know, indirect specular. Basically, the hardware
tracing is a component or a tool that Lumen has that it can
use for the purpose of reflection. So hardware ray tracing is
still alive and well for reflections in UE5 with Lumen,
and will continue to be. But that's the path forward,
basically. Ray tracing will be
integrated into Lumen. VICTOR: Thank you all. I know that a couple of you
have some important things to deal with right
around soon-ish. So I wanted to give all
of you the opportunity, before we end the stream here,
I wanted to make sure that if
there's anything else you thought of that you wanted
to bring up for our audience here today. Then I'm going to move
over to our little outro. They've said
everything they wanted. Go ahead, Chance. CHANCE: Of course,
I have something to say. No, the only thing I have to say is,
thanks for coming along
on the ride with us. Community, we're really excited. We want you to let us
know what you think. Download the tools, poke around. We've seen some really amazing
photographs come across social, or across the forums and whatnot,
videos. As always, you're doing
incredible work out there, with our tools. So we really like that. Continue to give us feedback,
as you always do. We're all really stoked
about this next generation. This is just the start. We have a ways
to go ahead of us. But we're all excited
and we're really excited to have
you here with us. NICK: Yeah, I'll just echo that. The early access release is
the culmination a lot of hard work that the Development
team have been putting into getting UE5 ready. So seeing all of the
really cool projects that you all are
working on is really inspirational for the team. And really looking
forward to what else you all create with it. CHANCE: That's right. Then also,
shout out to all the UE engineers and all the folks that
have been working on making this possible. It's been a pretty
massive effort to get here. It's just been really amazing
seeing it come together. NICK: And all the Slack
threads we have going on, with all the people posting
links of all the cool stuff that's being done. I've seen the Infinite
Dog quite a bit, and a bunch of the rebuilding
elemental with Lumen and stuff like that. That's the stuff that keeps
everybody going and inspired. Please post stuff. Especially silly stuff. That's my personal favorite. More Infinite Dogs in Nanite. CHANCE: Yeah. A volumetric dog,
if we can get it in there. NICK: Yeah,
if we can get volumetric dogs, so that we'll know
we've made it. CHANCE: Right. Then also too,
I wanted to mention, I'm sure that Victor will
probably ring us in with this. But there's a ton of
questions we didn't get to here. But we'll be taking these
out and parting them out for a lot of the
future livestreams we have that are
a little bit more contextual on specific topics. So we'll try to get to
everything in those as well. VICTOR: On that note,
for each of the livestreams that we do every week, not just
in regards to Unreal Engine 5. We post the forum announcement
post a week or two prior to the stream itself. And that is a great place
to go and ask your questions before the stream. Now, we have gathered
all of the ones from today. So we will try,
as to the best of our ability, to prep our future guests
with these questions. However, if you didn't
get a chance to ask them and you're watching this,
the Vault, on YouTube, or Twitch afterwards, when
those posts go live in the forums, head over there. Ask us your questions. Give us as many opportunities
as possible to answer the questions that you have. Because occasionally,
they might require a little bit more thought than something
that you can do on the spot. That's the best
place to go for that. Of course, you can always spin
up other threads on the forum as well. But this is a
special opportunity that you have to
specifically prepare the guest who is coming
live for us on that day and then tune in. That said,
all of our livestreams are available on
YouTube and Twitch. Immediately after we go offline,
you can find the entire Inside
Unreal YouTube playlist available right below on Twitch. If you're on YouTube,
it should be right there already. If you started watching
us today and you are new to the world
of Unreal Engine, go to www.unrealengine.com. Actually,
you don't even need a www. Maybe on some platforms you still do,
I don't know. But unrealengine.com is where
you can download the Epic Games launcher. From there, you can download
Unreal Engine, whether that's 4.-- 4.26, .27, or .0. Actually,
I thought it was kind of funny when Chance mentioned 4.0. If you're still on 4.0,
you should probably upgrade, at this point. That would be my recommendation. No? Chance says no. CHANCE: No, it still runs. VICTOR: It still runs, yes. Anyway, going off tangents here. That's where you can
find Unreal Engine. There is learn.unrealengine.com,
where you can find over
90 courses produced from professionals
in the industry, on how to use Unreal Engine. And you can learn everything
from lighting to animation. And like I mentioned previously,
these skills will transfer over
to Unreal Engine 4, just like your projects will. So if you'd like to learn,
there's no better time than
to start right now. We do transcribe all of our
livestreams on the channel. That script usually is
made available about a week after we go live. It does contain timestamps. If there was
anything in particular in the stream that you don't
remember when we talked about but that you
want to dig into it, you can download
that transcript. Or just turn on Captions for YouTube,
which is the ultimate and best reason for them. And you can find that,
Control-F, and you can find that
timestamp in the video. We do timestamp all the
videos on YouTube as well. Honestly, the best place to
watch the VOD is probably on YouTube, because of that. There are completely--
excuse me. There are communities
around the world that are still throwing virtual meet-ups,
even though we are not seeing each other in person yet,
as the pandemic is still raging strong. communities.unrealengine.com
is the place where you can find groups
of people in your local area to get together with. Some of them are still
throwing virtual meet-ups around the world. That's a good place to find
other like-minded people who are working with Unreal Engine. Whether that's games,
virtual production, media, architecture, you name it,
they are all over the world. If you can't find a group,
there's a button to become a leader. Fill out that form. Get in touch with us, if that's
something you're interested in, and we will get back to you. Make sure you visit our forums. There's also our
unofficial Discord channel, unrealslackers.org They just spun up last minute
UE5 channel's there last night. So if you have those questions,
you can head over there. There's also the new
category on the forums, where you can head over
and ask you questions. We'd love to see
what you're working on and get all your feedback,
so that we can make you UE5 the
best engine there is, can be. We're still looking for
new countdown videos. If you have them,
go ahead and shoot them to community@unrealengine.com. This is 5 minutes
of development. Whether that's sped up,
whether it's real time, you be our guest
and fill that out. We'd love to see all the
stuff you're working on. Make sure you
follow us social media. And if you stream on Twitch,
make sure you use Unreal
Engine category. You can combine that with our
Creative or Game Development. They're both good. It allows us to find the
stuff that you're working on. We try to come in and hang
out with you as often as we can. Notification bell on YouTube,
and now I'm about to be done with my outro. Next week we
have Nanite going on with Brian Karis, Galen Davis,
and Chance Ivey, on the stream. I hope you all are
excited for that as I am. And with that, I would like
to give our special thanks. Please, chat,
give it up for our guests today, for all the time they gave us,
the knowledge, and also all the work
that they've done and will be doing in the future. Thank you, all. Now I want to see
chat just go crazy. OK, there you go. CHANCE: Great. VICTOR: That's
a little bit better. CHANCE: It's silent applause,
but we can watch it. It's like watching audio,
like we talked about earlier. Thank you all for making the time,
Engineering folks, for this. This is super, super awesome. Real excited about UE5 and
where we'll take it from here. NICK: Yeah. Thanks, everyone. MICHAL: Thank you. NICK: Thanks. VICTOR: Thank you all. With that said,
it's time to say goodbye. We'll see you all next week. Let's roll the outro. Take care, everyone. [MUSIC PLAYING]
