>> Hey folks, we hope
you've had a glorious week. We're excited
to welcome you back with another round
of free Marketplace content. Our October set features
a fully animated combat set, handcrafted movement animations,
an easy to use multi save solution,
beautiful foliage, and a fully customizable
loading screen system. In addition to all the timed
to free content, a robust JRPG template
and modular sci-fi environment bundle have been added
to our permanent collection. Download them all today. Soaked with info on the state
of our oceans, 2D side-scrolling puzzler
Koral was designed to be enjoyable for players
of all skill levels, enthusiasts
and casual players alike. To execute the goal,
solo Dev Carlos Coronado designed Koral’s gameplay to use nothing more
than a single joystick. In our discussion
with Coronado, he details how scuba diving
helped him capture a realistic yet surreal depiction
of the ocean, elaborates on his drive
to be a one-person studio, and shares tips
for aspiring solo devs, looking to follow suit. Dive into his interview
after this. Already on the third evolution
of their car configurator, Audi is trailblazing the use
of real-time technology. With the most recent edition, vehicle changes
are continually updated to the Unreal Engine files, meaning Audi dealers
can now demo the latest version of the vehicle
almost immediately. These real-time ready Assets could also be leveraged
to create even more projects such
as interactive service manuals, custom user created videos,
or marketing Materials. All this vastly increases
their return on investment. Do you have questions about UMG? Well, you're in luck
because we've got answers. Our own Chris Gagnon,
lead programmer on Editor, assembled and provided
a handy overview of best practices for UMG. His tech blog covers
how to structure your UI elements
in both C++ and Blueprints. Be sure to walk through
his helpful example. We want to extend
our congratulations to Frogwares who recently released
their Lovecraft-inspired investigative
thriller on Switch. We caught up with the team
leading up to their launch to discuss
their development process, to find out how they
overcame design challenges, and to discover
what inspired them to dive into the dark
setting of HP Lovecraft. Now for this week's
top Karma earners, we're giving shout outs
to ClockworkOcean, Adnoh, Shadowriver, Alekann01, IndieGameCove, GarnerP57,
expose, Rumbleball,
Herr Edgy, and chrudimer. Thank you all for supporting
other folks on AnswerHub. First up in our
spotlight section, This Dead Winter
is a dark fantasy adventure game set
in an ancient woodland suffering through
an unending winter. You, a beautiful fox, embark on
a journey across ruinous lands to find your missing cub. Follow their development
on their page. Community Dev Bartosz Kamiński
has recently kicked off a series of four-minute
tutorials for beginners, teaching the basics
of procedural Materials, implementing footstep sounds,
or spawning meshes in a radial. If you're just getting started,
hop over to see how he may be able
to help you on your journey. Last up, this week is DAYMARE:
1998, a third-person survival
horror game set way, way back in the 90s,
where you explore and uncovered deadly creatures that lurk in sleepy
Keen Sight, Idaho. Alright, that's it
for this week. Thanks for tuning in
and we'll catch you next time on the Unreal Engine news
and community spotlight. >>Victor: Hey everyone and
welcome to the Unreal Engine livestream.
I'm your host Victor Brodin. Today I have invited evangelist, Chris Murphy to come
and talk to us about some of the adventures
with shaders you've been doing, working a little bit
on landscapes and other various cool topics. Just to start us off, what time
is it right now for you, Chris? >>Chris: It's 4:00 in the
morning. >>Victor: Chris is a champ
everyone. He asked me,
“What time is the live stream?” And I said,
“Well, it's 2:00 PM over here and on the East coast,” and he says, “Well, that's
4:00 AM my time. I'll do it.” So big up everyone in chat. I'm happy you're
able to do this. >>Chris: He’s
exaggerating a little bit. There was a definite groan
before the words, “I'll do it.” >>Victor: Okay, fair enough. There was a bit of
a conversation in between, but he said, “I'll do it,”
and you're here today. So, very happy about that. Before we get started,
I'd like to let everyone know that we will be handing out keys
to Daymare: 1998 in chat. So, keep a lookout for those. There will be a little link
to an image, so we actually have to
type them in manually. With that said, I think
it's time to get started. What is your first thing today,
Chris? >>Chris: I'm going to be running
through a couple of effects. I'll be honest with you,
I'm not running really at 100%, just due to the time
zone difference, but I'm going to step through
a few different things now. I kind of last minute
changed up some of the topics I was going to cover
but I'm going to begin with a couple of
just little landscape things that I've noticed
a lot of people don't do, that I would like people
to just be aware of. After that, I'm going to move into
the procedural building stuff that a few people
have seen me do on Twitter,
which I should be able to the remake from scratch
reasonably quickly. After that I’m going to move
into some kind of interesting Post Processing effects
and concepts. Then finally I'm going to touch
on the Pre-Skinned Position and Pre-Skinned Local
Bounds nodes that have come in
various fashions of Unreal but we haven't really -- I haven't seen too many teams
really taking use of them yet and I want to stress to teams
why these things are powerful. Because I'm unfortunately running on Australian
internet from home, I'm going to have to kill
my video feed while we're running this.
So – >>Victor: We were not able
to do both at the same time. >>Chris: Switch
off my screen now. Yeah, we have to pick one
or the other in Australia. Here we go. The feeds coming through now? >>Victor: Yeah, we see it.
You're all good. >>Chris: Rad. Okay! Video games. So, we have a very
basic environment that's coming through
at the moment. We've just got
this kind of empty area and I'm going to use this
as just kind of a test bed for what I'm going to be
developing from here on out. Now the first thing
that I want to do, and I'm going to try not to move
the camera too quickly because I know we're running
at like one frame a minute, but, the first thing that I really
want to touch on is landscapes. There is a very simple
little option that if you're going to be
using a lot of foliage -- I'm not sure if you'd be able to
see it very well in this video, but you should be seeing a lot of popping going on
with this grass -- >>Victor: It's coming through.
>>Chris: -- as I move forward and backwards.
That's coming through okay? Oh, what a world.
Just be aware of folks that to fix your LOD
popping on Assets, is literally as easy
as opening up the Material and clicking this
Dithered LOD Transitions and then just hitting apply and then moving on
with your life, because seriously,
that's all it takes. What we will get in return
is your grass will now actually do like a nice screen
door fade between each version. Now because I just applied it
to a grass that has like four
or five variants, it'll take it
a second to build shaders but hopefully you'll see where
I'm kind of going with that and why that's advantageous. I provided b-roll for a couple
of different effects that we'll be doing
today as a just in case, but you'll now notice
that as I move forward and backwards, we're no longer
kind of getting that effect. I know it's a really
simple thing where Chris’s first thing
to show you is a checkbox, but, for anyone
that's unfamiliar of what my role
as an evangelist is, basically across Australia
and New Zealand and Southeast Asia,
my job is helping teams. So sometimes it's obviously just
meeting people at conferences, but other times
I'm on site for a day here and there for different teams or dropping in for an hour
to check out things. There's a lot of commonly
reoccurring issues that we really see popping up. So, it's just little things
like that kind of checkbox that I'm always
kind of flagging. If you're following me
on Twitter, you're probably used to me
generally harassing people
over these checkboxes. Obviously, the heather in here
has not been set up to do it, but you'll notice
that we are no longer just getting that pop
of the grass. I just really wanted to stress
that as the first thing. The next thing that I wanted
to go into was, quite often this language
of "Smart Materials" gets thrown around and anyone that’s talking about
Smart Landscape Material, it sounds like a really,
really complicated thing and it sounds like there's going
to be more work in there, but it's straight up,
it's really straightforward. I really want to show people
about how to approach a little bit of logic
in your Landscape Materials to get you a long way. I'm going to load my simple
Landscape Material that we've got running
at the moment. This was the one that we're
seeing in the world right now. It's nothing too fancy,
but I want to run through a couple of things
that it's doing. First off,
when I opened this up, what I've got is
I go to my paints and I just paint
in a little bit of dirt. You'll notice that as
I painted this in that grass is disappearing and reappearing
and plenty of Unreal users are going to be
already aware of this one but I just want to make sure
that everyone is, especially if you're new
to the Engine, or you've just started looking at
things like these live streams. In my Landscape Material,
I scroll down here, you'll see I've added this thing called
the landscape grass output, and what this actually
lets you do is the values that get read
into this, if it's a one, it will automatically spawn
Grass Meshes at that location. Because that's coming through
with a value of one, it's spawning grass. A little cheeky thing
that I'm doing is I'm doing a one minus value
over here applied to a Texture, you can see the Texture here, which is essentially just
swapping it every now and then. So sometimes it's grass
and when it's not grass, it's flowers, and both of them are looking at the
landscape layer called grass. So, this will actually go ahead
and just spawn those Meshes and this just feels, I mean, just for perspective sake
of what this looks like, if I didn't have grass enabled,
look at that thing. What is this, amateur hour? So, having this kind of set up
just feels so much nicer. It feels like a complete product and it's really
not that much work to just get it following
a simple couple of rules and have
something running there. Do me a favor and just make sure
that those things are in place. Now the other thing that really
want to stress is go back to your landscape. Now at the moment, landscapes
can end up at giant mess but what I often find
is a really nice way to handle these things is, you'll see here that they have
a thing called ML_Dirt and another one called
ML_Grass. This isn't anything fancy. I've literally
just created a Material layer and what I mean by that is
if I open this up, you can see that
this is the Material for grass and I'm calling to set
Material attributes at the end and I'm actually
passing this around as its own little mini data type
and by doing that, it means I can just kind of crunch down
all of the dirt logic and all of the grass logic
into a single instance, like a single node
that I can just drop in. Now, not only does this allow me
to have fewer nodes causing chaos in my landscape, it also gives me the added
advantage of being able to have, if I ever need to use dirt
in my project, I can literally drop
in that ML_Dirt layout and share it between
different Assets. If I drop in a tree and it needs
to have dirt or grass or something at the base of it,
I can re-sample those things, use the same Textures and not have all of extra
complexity for no good reason. The other reason
that I bring it up is because also you can then use this Blend Material
Attributes node, which literally acts
like a Lerp, but it Lerps
entire Materials together and we're going to be using
a little bit of that today. To give you a quick example
of what that means, let's add in an entirely
new Material layer. I'm going to use the same
Material attributes and I'm just going to create it
as a spot here. I'm not going to create it
as a separate Material function, but to do this, I'm just going to quickly
drag in some Textures. I know I have some sort
of rocks available to me. >>Victor: Hey, Chris,
if you're able to, when you're zooming
really quickly, it's coming through
in a little low frame rate. So just try to be
a little slower then we can follow along,
alright? >>Chris: Yeah, no problem.
Thanks for the heads up. Please direct all concerns
about my internet speed to dm.gov.Au
and we'll go from there. So, with this here, I'll try
not to do any sharp zooms, but we’ve got this rock
[inaudible] Texture and just this Normal Map. I'll just preview this
on the Assets, you can see what's up.
Nothing too fancy. I'm going to get this
and I'm going to add what's called
a World Aligned Texture. When I add a World
Aligned Texture, what this essentially does is, if you imagine a projector
was placed along this axis and it projector
was placed on this axis and another one on this and they were just shining their
light at whatever was there, it blends
those three images together. I'm actually making hand motions
right now that no one can see
while I describe. >>Victor: Next time, next
time you'll be in the studio. >>Chris: Next time,
oh thanks, man. I'm going to get this and the first thing
you need to know is that when you're handing a Texture
to a Material function, they need to be handed off
as a Texture 2D type. This typically outputs
just the color values, but I need to actually
output the entire object. So, you just can’t plug it in, but to do it you literally
just right-click and hit Convert to Texture
Object and just plug that in, okay? Because I'm going to need
to start previewing this, I’m going to go ahead and I'm going to call
a Set Material Attributes like that node you saw
just before. I'm going to add three channels. I'm going to add a base color, a
roughness, and a normal channel. For the time being,
I'm just going to plug these in and I'm just going to
preview it over here. When I preview it over here, I'm just going to go
to this sphere. You'll notice here
that unlike before, this sphere actually has no pinching going on
at the top of the sphere and that's because it's just
projecting this Texture in space at this thing.
And the reason I'm doing this is because
what we can actually do is, we're going to use this Texture as a way of projecting cliffs
into our landscape. I'm going to right-click and I'm going to call
Blend Material Attributes and I'm going to blend between
the existing Landscape Material where we have dirt and grass
and I'm just automatically going to add cliffs
any time it's sideways. Next thing that
we need to know is, well, how do we do that?
Actually, just before I do that, I'm going to quickly
set up the roughness value so that
it doesn't look terrible. In a situation like this,
this Texture that I'm reading in is pretty close to gray scale. If I was to grab just
the red channel from it, I'm going to use that
as a black and white mask and I'm going to run
that into a Lerp and I'm going to
tell the Lerp that hey, whenever the value
is a purely bright, I want that to have
a roughness of like 0.8 and whenever that value
is super dark, I want that to have a very high
roughness value because I don't want to have
the shiny crevices going on. This is just a quick way of me
just getting around, not having a roughness
for that Texture. That's going to be the rock, the cliff face
that would kind of bring in. Actually one of
the little things I'm going to add is just a scale and I'm going to plug this
into Texture size and
this actually kind of controls how big
that projection going onto it is and if I was to set this to 512, you'll see that --
give it just a second. Sorry, Victor, I know
in micro zooming, it's a habit. You can see here
that if I set that to 512, it's quite large, if I was to drop it down to 20,
we'll see a tiling Texture. In this situation,
I'm just going to set up the 512 and move on with my day. The next thing that we need
to do is work out how we're going to blend
between top and sideways. To move between the -- Sorry everybody. To move between this
and this version, what I'm actually going to do is I'm going to use what's called
a normal, in world space. So, I'm just going
to preview this node and hopefully I can explain
it sufficiently to everybody. Now for just one moment, if you're
unfamiliar with using this node, that's fine.
What I want you to do is look down on the bottom
left corner of my screen and you'll see that we had X,
Y, and Z. Okay. Also, it's really nice using
the word "Zed" instead of "Zee," while talking
to an American audience, but that's what you get. So, we've got our X axis
our Y axis and our Z axis. You'll notice that those colors correspond
to red, green, and blue. If I was to look around here
as something as along the Y axis,
you'll see that it goes to green and then it goes into
what seems like black. But actually, what this is doing is giving me a value of negative
one on the green. Same thing goes for here whereby
the X is a value of one, and over here it's a value
of negative one. All of this basically tells me
which direction the surface of the sphere
is actually pointing in. The more blue it is,
the more upwards it's pointing. By knowing that the more blue
something is, the more upward it's pointing,
I can get this, I can say to it, “Hey, I just want your
blue channel or your Z axis.” That's now going to give me
a value that when it's completely
facing upwards, it's white
and when it's down here, it's going to approach a value
of zero, value of black, which is actually pretty close
to what we need here. But the truth is,
is that I really want to get this rock kind of blending
in nicely up top so that it doesn't just do like
a smooth interpolation into it. One nice thing that I can do
is I'm going to get this Texture that I used
earlier for this noise and I'm just going to bring
that across to here. What I'm going to do is
I'm going to subtract this value from this one. Even that won't quite be
what I want yet. I'm also going to lower
the scale here so they don't match up.
So, watch what happens here. If I was to get this and I was
to just subtract this channel and just preview it,
you'll notice that we're now kind of
getting this pulling away at it and that's a lot closer to what we really want
to be seeing here. See how that's subtracting
through? Just a quick little bit of logic
that I want you -- Excuse me, hiccups. Just a quick little bit of logic
that I want you to be aware of is if this value is white,
if it's at the brightest point and I'm subtracting it
from this value, what this means is that
a value of white here even subtracted
from a pure white value up top is still going to be equal
to zero when they line up. So, whenever this happens
to be a value of one up top and it's subtracting one,
one minus one is zero. What is good to do here
is to actually multiply. If I was to multiply this
top section by another value, what we're essentially given is a way to kind of push up
what these numbers are equal to. I'm just going to call
this Cliff Blend Multiply, and hopefully this will make
sense in just a little bit. As I plug this in, what we're
actually going to see here is, notice that -- actually I'm going to add
a saturate at the end as well. If you notice
what's going on here, watch what happens when
I change this value, see where that line
is ending up? It's getting pushed further up
and down as I change this. 54 was too high. But what that's
essentially doing is this value was previously a value
of zero to one, right? Zero, one. Now this value
is a value of zero to two. By being a value of zero to two, it means even if this value
is equal to one, it's only what was initially
a value of 0.5 that is now going to actually
accept any change in its value because two minus one
is still equal to one and what a saturate is, it’s a clamp
between zero and one. It stops any value
being above one just makes it equal to one and it stops any value
being below zero, it just makes it equal to zero. I've done those two and then
the other little thing that I would really like
to do in this situation is I'm going to use
a Power node. A Power node multiplies
any value coming into it against itself.
What that means is that, I want you
to think for a second, one times one is equal to one,
right? But 0.5 times 0.5
is equal to 0.25. The reason I bring this up is because
what that information tells us is that when we multiply
a number by itself, that's between zero and one,
the closer it is to zero, the quicker it gets darker
but the closer it is to one, the more likely it is
to stay at the same. It's essentially a way
of providing contrast to something
that we're looking at. I'm going to get this and I'm
going to create a new value called Cleanse Contrast,
or, apparently, "contast." I'm going to set this
to two just initially. Now having these two values
available to me, I want you to look
at what happens to the white and dark values
on this as I increase the contrast.
17 was too high, but see how that's actually
pulling it up into it. That's actually kind of
allowing me to kind of create a really nice little blend here.
Now I have no idea of what's going to be
a good value in this situation. But that's okay. We're just going to plug it in
and see what happens. In this situation,
I know that a white value is going to be
standard landscape and I know that the black value
is going to be cliffs. I'm just going to get this Blend
Material Attributes node that we've got.
I'm going to grab this. I'm going to say,
“When it's white, I want you to be that
and when it's black, I want you to be that.” By the way, a little thing
that a lot of people don't seem to know is if you hold control
and click on here, you can just grab
the end of a pin and plug it into something else. I say this because quite often
I see people disconnecting and going back
and plugging things in and that would be
a waste of time. I've got that plugged in now,
that'll do it for the moment. I'm just going to hit apply
and we'll see how this works. Victor is that feed resolution
coming through okay? >>Victor: It’s coming
through alright. I do know that when you're
compiling shaders it definitely takes
a little bit of a hitch. >>Chris: In that case, I'll try
to just keep things still whenever
I'm compiling shaders. With those things
kind of plugged in, I'm just going to go back
to the landscape and we'll let this thing
just do what it does. Give it just a minute. >>Victor: You want to take a
question? >>Chris: Sure. Let's do it. >>Victor: What are the benefits of,
or differences between using Material layers
versus Material functions? >>Chris: Essentially, they all get
compiled down to the same thing. So, your Material layers -- Sorry. Let me start again.
Can you clarify with the person whether when they say
Material layers, they're referring
to the Material Layer System that's experimental? >>Victor: No, when you made
the, what I thought was some Material function,
you said Material layer. >>Chris: In this situation,
what I'm actually doing, is a Material function -- a Material layer is a Material
function in this situation. So, as we've been bringing in
the experimental Material layer, I think that might be
a different system for that, but essentially it is
just a Material function. That's all I'm doing,
I always use the prefix of ML to remind myself that the output
of whatever this is and also for quick selection,
that this Asset is going to be returning
an entire Material type. So, I know all of this stuff
that's kind of coming out of it. Sorry about that. That that was me
not clarifying very well. In this situation, I'm just
going to go back to here, let's just grab this over
to here and I'm going
to get the sculpt tool and I'm going to just flatten
for a second and I'm going to pull this out. I'm going to turn off grass
for just one minute, we'll fix that up
in just a moment, but you'll notice
that as I pull it out now, what that's actually
automatically doing is it's creating
that Rock Texture whenever this thing
goes completely vertical. Do you see that? >>Victor: Yeah. >>Chris: Having that
kind of logic set up actually lets us create
more natural landscapes that as we kind of
just build cliffs, the cliffs are becoming cliffs. What I'm going to do
really quickly is just a tweak these values so that it doesn't
look terrible. The scale looks like it's okay.
So, 512 was fine, but what I will change
is just this Blend multiply to being something
that's a little bit nicer. So, I'm going to pull that
to about maybe there, going to get my blend contrast
and you'll notice that again, if I was to get this contrast,
as I increase it, we're getting a hotter edge
so that when it goes to rock, now it's just going
straight into it. But if I was to drop
this contrast right down, it's now going to be
a softer blend between them. If I was doing this in like
an actual professional capacity, I would use a Height Lerp
to kind of get at blending into the cracks
of the rocks first. For anyone that wants to know
how to do that, literally just typing
in the word Height Lerp and just using this,
it's pretty great. You can just use a Heightmap
and it'll blend between them from most values
to highest values. It's a really good little node, but I didn't really need to set
that up to prove a point. With this in mind, and this kind
of generally set up, the next thing that
I'm quickly going to do is I'm just going to go back to where my grass
sample was earlier. I'm just going to reorganize
this a little bit, grass stuff. I'm going to get this
and I'm going to multiply the existing
grass value against this and I'm going to add a value
here called Grass Contrast. I'm going to set this grass
contrast to like four, but now if I re enable grass,
this should now make it, no grass will appear
on the cliffs either because it's now using
that same logic to do it. While that's compiling, feel free to shoot
another question, Victor. >>Victor: Sure,
we can grab another one. Someone was curious about, can Texture objects be used
as Material parameters? >>Chris: As Material parameters?
Yeah, you just get the -- Ta da! >>Victor: If you
didn't catch that, that was a Texture
object parameter. I believe you can right-click
the Texture sample as well and convert it into a parameter? >>Chris: No,
you've got to convert to object and then you convert the objects
convert to parameter, either way,
Bob's your mother's brother. If I go back and I look at this,
my contrast here wasn't quite set up nicely. We're still getting
a little bit of grass occasionally spawning
on the cliffs, but that's literally
just going to be a matter of me getting this cliff blend
contrast -- sorry, this grass contrast and tweaking that
until I get something. Unfortunately for this one,
I do need to hit apply each time to get it to work out when is acceptable to bring in
those cliffs and when it's not. >>Victor: What would be
the workaround around that if you sort of want to iterate
a little bit faster without compiling shaders? >>Chris: If you're willing to -- In situations where
I've had to really debug that and really preview it nicely,
I literally just run this value out to the diffuse color
for a little while. I just override
the diffuse color and I literally just preview
the black and whites because I have
a pretty good idea of how that's going to look
when it all comes out, but that's not necessarily
at stake. >>Victor: Do you usually use
master Material for your landscape rather than making it
an instance? >>Chris: Yeah. I try to for as long as I can until the point
that I'm forced to change. There's always a certain given
point where you'll be like, “I guess I've just got
to bite the bullet and do that thing now,”
but as much as possible, I typically try
and run it this way. The problem is,
if you build with instances, it's going to be
recompiling your master and every instance
the second you hit apply, and that just slows down
your work flow massively. >>Victor: Mmm, okay.
That's a good tip. >>Chris: It's good
in the long run, but when you're doing
that initial setup and you're just
getting things working, it’s kind of just wastes
your time a little bit. As I said, this is just kind of
like a little bit, and obviously I'm doing it on
like a tiny hill, so it's not really
the best example. I really should have made
a much larger thing. But hopefully
that still gives you an idea. I'll just pull this up
for a second so we have a better idea
of what's up. Again, I’m going to get
this flat node and I'm just going to rip out
some chunks of this. It's going to look a bit strange
for a second until we erode it. So just bear with me. >>Victor: What are the keyboard
shortcuts you're doing to set
the size of the brush? >>Chris: The square brackets
control the size of it. It's just the same as Photoshop. I've got this and I'd just like
to kind of rip this out. Now if I set this up properly and I kind of tweaked
these values and I had a rock that wasn't
tiling quite as much, I'm sure you can kind of start
to see where I'm going from. A lot of people
kind of look at this like it's some kind of like
weird magic that's going on, but honestly, just getting this
kind of setup for your Material starts to just create automatic
diversity in what you're saying which really just helps
kind of get everything going. Sorry, I'm just
bringing this together. That was the first quick thing that I kind of wanted
to cover with folks just to let them know
that that's entirely achievable. The next thing that I wanted
to cover was, quite often people
want to do a technique known as interior mapping. I'm getting ahead of myself.
Let's take a quick look at this. I'm going to create
a new Material and I'm going to
call this M_InteriorMap. This interior map Material,
imagine for a second that you're looking
at a building. When you typically look
at a building in a video game, it's just going to be
the picture of the buildings, facade or windows. I bring this up because
it's very easy to do something that's a lot more than that, but involves us using
a lot of Texture and involves us
using a Cube Map. A Cube Map is -- I'll open this
Asset up -- A Cube Map looks like this.
To a create this, I literally opened up
the Realistic Rendering Demo, so that it's available
from the Learn Tab. So, you can just
pull this down yourself. All I did was literally
just grab a camera cube, sorry, scene capture cube,
which takes a 3D photo for you. I saved that out as an Asset
and that generates this for me. If I was to get this Texture
and pop this into here, unlike normal where you can just
right-click and hit preview, you'll notice that
when I preview it, we get an error.
The reason that we get an error is because
Unreal actually needs to know which direction
inside of the cube it's currently looking at. What happens if I was to get
the reflection vector and feed that in? You'll notice that now
when I look at this sphere, we're kind of peering
through the globe. Do you see that? >>Victor: Yeah. >>Chris: If I was to get
the camera for instance, you can see that it's now
syncing out correctly. It's upside down,
but it's actually syncing up as if I'm kind of looking through
a little mini port. If I was to use this,
that logic in mind, all it's actually doing is it's
feeding in the direction to say, “Hey, you're looking at this
3D thing from this perspective.” I bring this up because,
check this out. If I was to go Interior
Cube Map, this is just a node. If I was to get that
and I was to plug that into my UVs instead,
it gets the Cube Map and it breaks it into
nice little rooms. See that? >>Victor: Yeah. >>Chris: Now,
by getting something like this, this is where we actually
really start to see something that's kind
of powerful because what happens
if I was to get this and I'm just going to apply this
to a cube in my world. Do I have a solid
cube to use here? Sorry, I assumed I would have
a cube because that's -- just for a second,
I'm going to apply this, kind of drop that into the world and I'm going to apply
my interior map to it and I'm just going
to embiggen it. Alright. There we go. You'll see here that now what
we've essentially started to get is something that actually
lets us look into these rooms. The next little thing
that I'd like to do here that I find to be quite useful is I'm going
to get this interior map and you see this little
randomized rotation Boolean that's been set up. If I get that, I'm just going
to read in a value of True, and what that does, you can see
it already happening there, but if I hit apply
it's actually going to randomize which direction you see.
Now we're using one Cube Map, but each room looks like
a distinctive room. This is really great because
if we were to get something like this,
alright, so that's really obviously
just squares, but nothing stops us
from setting the roughness value of what we're seeing here to be
very low, like a value of 0.1. I'm going to multiply this
against a very dark value, like zero in here
and plug that into my emissive, because by
adding that reflection -- so by setting the roughness
to a really low value and setting the emissive color
to a very dark value, it now kind of looks like
we're looking through glass, but behind the glass
it still looks like there's a room present.
Do you see that? >>Victor: Yeah. >>Chris: This is the first step
of doing something that's a little bit
more interesting. The next thing that
I'm going to dive into is definitely going to be
jumping up in terms of advanced workflow but hopefully it gets you
a good idea of how to use this. Give me just one second. I'm quickly going to import
an Asset that I had elsewhere, there is a cube somewhere. Sorry about this. I'm just going
to quickly pull in, one of these is a cube Asset
and I'm currently not seeing it. Sorry everybody.
I just want to get a cube, that's actually set up from the
bottom rather than set up from -- There it is.
Great. Cool. Done. I got what I was after. I want to get a cube
that's actually been set up so the pivot point
is on the bottom and I know I could just
duplicate an existing cube, but I just want one
that I trust. Alright, great. The reason I've done this, so I brought this cube
in instead, sorry, every now and then the Unreal launcher
can do this. Can we shake out of it or not? Every now
and then the Unreal Launcher can to be
a little bit finicky on that. >>Victor: Looks good now.
>>Chris: All fixed. I'm just going to quickly
swap this over to the cube that I just imported and the reason I've done this
is my cube that I just imported, the pivot point is down
at the bottom because this way,
if I scale it up, it just makes it bigger
and if I scale it along the X or the Y,
it's going to, do it accordingly.
What I would really like to do is I would like to get
this building so that -- and I'm going to revert it
to being really visible, what I'd like to do
is get this building so that as we made
the building bigger, we actually added more groups. You can see here that
as this gets bigger right now it's just going to make
the rooms themselves bigger, which isn't really what we want. We want the rooms to stay
reasonably consistent. We just want to make sure that as it gets bigger
it does its thing. There's a good way for us
to handle this. What we're going to do
is we're going to go into this and first thing I'm going to do
is I'm going to kill the tiling. I'm going to set this tiling,
which is currently as you can see, set
to a value of 4;4, which means four rooms
by four rooms and I'm just going to set that
to a value of one. By setting this
to a value of one, it's just going to give me one
big old room as we look through this. Now, the reason I've done this
is I want to manually increase or decrease the size of the UV,
so the tiling through the UVS and I'm getting through
that according to scale. So, here's a really cool thing that a lot of people
aren't aware of. Let's begin by thinking about
the problem that we have. What we really want to do
is as the object gets bigger, we want to make it tile more along
the directions that we see. So, if you imagine that
we were looking at the X axis, which is this one,
we would want the Y axis and the Z axis
to increase how much tiling is going on based on the size
of this surface. Does that make sense,
how that would work? What I'm going to do is
I'm going to go back to the node that we used just before,
which is, I'm going to get the normal,
I'm going to preview it. And I want you to think about
what we actually have here. So, this, when we looked at it
as a sphere, it gave us this. But when we look at it
as a cube, we're actually given colors that correspond to the direction
of what we're looking at. Now, these are negative values
on their opposite sides. So red becomes negative red,
but if I put an Abs function, what an Abs function does is, any value that would typically
go into a negative now becomes positive. It's like
the most optimistic node. So, what this is doing
is a value of negative one, now it just becomes
a value of one, but a value of one
stays a value of one. By getting this, what we're
actually able to do is use this as the way of blending
between the sizes. And then you're probably like,
alright, but how do we get the sizes?
But that's actually super easy, because you can literally
call Object Scale as a node,
which gets you the X scale, Y scale,
and Z scale of the Asset itself. If you think about
the problem now, I mean really it's just a matter
of getting this and saying, I want to get this value and I'm going to say to it,
“Hey, whenever you are red,” okay, so whenever it's red,
that's not going to give us a black and white value
between here and here, Whenever it's red, we want
to get the scale of the Y axis and the scale of the Z axis and we're going to
multiply it against that. And whenever this value
is green -- Oops,
I should really plug it in. Whenever it's green, we want to get the X axis
and the Z axis. Do you see that? So, we're just going
to get the X axis and the Z axis and I'm going to plug it in
to that. Now fortunately -- Fortunately, this means
if we look at the logic that we have going on here, now this one says
whenever you're red, do this,
and when we connect it to this, which says that
whenever you are Y, do this. Which means that
the only thing left is that if you're not X or Y,
you must be Z and if you're Z, we're just going to get the,
the X and Y. Okay? I'm just going to append.
I'm going to plug that in. That little bit of logic there,
just doing that, is now going to get me the scale
according to any UV panel that we're currently looking at.
If I was to multiply that against a regular Texture
coordinate, which is just a standard UV map,
what I'm able to do when we plug that in,
by default, that should look
completely normal. I'm just going to apply. You’ll notice
that now as I scale it up, it's just adding more rooms in.
Do you see that? >>Victor: Yep. That's a
large apartment building. >>Chris: My default size
isn't great here, but we can fix that ourselves. This value is now
just tied into this. If I was to set the scale
of manually to be like 0.25, now it's kind of
just adding those in as I scale up, see that? >>Victor: Yes, you’re a wizard. >>Chris: The whole point of this is to establish
that I'm not a wizard. >>Victor: You're
generating rooms there. >>Chris: Having this going on
is a really good way to start processing how -- Having this going on is a really
good way to start processing how are we can approach
some of the development of our Assets
in a much more interesting way, because all of a sudden
we can start developing Assets that are kind of
intelligent enough to do what we clearly
want them to do. It's worth noting
for something like this that we don't have
any Blueprint involved. This entire thing
has zero Blueprint. It's literally just a Material
that knows what it's doing. So, having this kind of setup
means that can be like, “Hey, I need a building here
and I need the building here, but this one
is going to be this.” It's worth noting that
it's not that much extra logic as well to automatically clamp
the room sizes down. If you wanted to, you could
comfortably get these values that are kind of coming in and we could be
rounding them out and if we round them out, it's actually going to stretch
to the nearest room size. I'm saying this -- I'm
pretty sure this will work. I'm just going to round that
for the time being. Watch it.
See that? >>Victor: Yeah. >>Chris: So, it's doing it
in 0.25 increments right now because I'm manually tiling
and this Material is doing it as 0.25 but the round
is happening before that. So really what I would need
to do is have some sort of value called tiling that was
multiplying this against 0.25. We're rounding that value and I'm going to set this back
to just default. This wasn't actually meant to be
in the presentation, but— >>Victor: Doing it live. >>Chris: We're there. We're
there now. Let's just do it. If I get something like this, see it's now squishing
the rooms right up until the point
that I can fit in new one? Which stops us getting like
half rooms popping in. Honestly if you're playing
a video game, no one's going to notice
that that's what's going on. Do you know what I mean? If they're playing,
they're not going to notice that this room is actually
that squished and if they do, you probably should do a
better job of distracting them. >>Victor: Could you show us the shader complexity
of the current Material? >>Chris: Yeah, sure. That was actually the next thing
I was going to be doing. It's worth noting, you can get
all of this that I've done and you could comfortably -- >>Victor: What does that node
do? >>Chris: Vertex Interpolator -- So, for anyone that's unfamiliar
with Vertex Interpolator, usually developers
used to use custom UVs to pass things between
the vertex and the pixel shader. If you do
a Vertex Interpolator node it literally just does this side
on the vertex and converts it to pixels.
So, it just makes sure that these instructions
are being done. >>Victor: Nice.
That's super handy. >>Chris: This single node used
in certain situations can save massive amounts
of performance, so please use it,
it's good for you. Anytime I make something
like this I always get someone who's just like, “Yeah,
but how complex it is?" And then I'm like,
“It's still green." It's not as green as our
sky dome over here, but we're still pretty good
with how that's looking and it's not that
much extra work. It's funny cause
like these buildings are now going
to correctly instance, whereas if that were Blueprints with a whole bunch of
different parameters set up, unless I was using
custom [inaudible] or something, these wouldn't batch together,
but now they will. Being able to have
all of these buildings comfortably kind of batch,
especially if I just created a couple of instances
that are the exteriors. I've got a more complete thread
on Twitter for this, that you're more than
welcome to check out. My Twitter feed on that,
it’s like a 20-step process where I've gone
through this bit by bit and really carefully
explained everything and then I've added some
outer windows and stuff to it. The files are available
to download from that
Twitter thread as well. Feel free to get these Assets
from there to kind of get a better idea
of what's going on if you don't feel
like going through this video and like recreating it manually. How are we going for time
at the moment? >>Victor: We're
still pretty good. >>Chris: Good is a
relative term, but yeah. >>Victor: Do you want to take
a couple of questions in regards
to what you covered so far or do you want to wait
until the end? >>Chris: Yeah,
that's probably a good idea. Alright, let's do it.
Question time. >>Victor: They were curious,
can you blend entire Material attributes values
in a single Lerp node? >>Chris: Yeah.
>>Victor: That’s it. Yes. You can do it.
I think you were doing that. >>Chris: I mean that's
literally all it's doing. It blends everything together. The other thing is,
that's nice about it is if I look at that node
and I click it, you can say, “Hey, when it comes to vertex
attributes, just use A or B, but when it comes
to pixel attributes, I want them to blend.”
This can get around some issues that you get where you're using
a Texture to blend stuff that isn't quite playing nice
for vertex attributes. Even if you didn't want
to do it this way, you can still manually
blend stuff if you need to. At any point, I can get this
and call Get Material Attribute and then I can be like, “Hey, I want the base color
out of this,” and then I can call
Set Material Attribute and I want to set
the base color. Now the cool thing is
that if I feed this into this, I could do some
sort of instruction and then that will now
have all of these properties, but the base color will be
at whatever's plugged in here, so you can override
just that one channel. >>Victor: Do those add more
instruction counts in the Material? The Get and the Set Material
Attributes? >>Chris: Get and Set are actually
less expensive than using this, which a lot of people still use.
Please don't use that. But to my knowledge, I don't
think this adds any instructions because I think
they all get compiled out. Don't hold me to that, but I'm reasonably sure
they are compiled out. >>Victor: It would
seem like they were fetching and
setting the data rather than doing
an operation on it. >>Chris: Yeah,
I think it's just in the graph. I'm sure someone from Epic
that has a better understanding on that side
will be able to clarify. I've never gotten to a point
where I've been like, “Oh, woe is me,
why have I used so many Gets and Sets that my performance
has been terrible?” I'm assuming that it's fine and I've done
some pretty complex stuff on some pretty low-end machines.
I believe it's okay, but hopefully someone else
at Epic can clarify that point. >>Victor: Can you separate the
cubes slightly like for doing an apartment cutaway, walls,
floors, and ceilings? >>Chris: Yeah. You can do that. It requires a little bit
more math to really kind of figure out
how to do it. Essentially what you want to do, if you think about
the problem is, if you consider every value
that's coming in -- I can probably think
[inaudible]. >>Victor: We can
follow up afterwards. >>Chris: No, let's just do it. If I use a Frac node,
one of the things that a lot of people underestimate
is the power of a Frac node. What we're doing right now is
everything that's coming in here is telling us how much this
value is being multiplied by to end up at its new position.
But if I was to get a Frac node and plug that
into my emissive color -- sorry, I just realized
I forgot one thing. I didn't want the Frac node
from there. I wanted the Frac node
from here. What this Frac node
is essentially doing is the Frac node is going to make sure all of
my values between zero and one. So, if I have a value of 4.5,
it's now just equal to 0.5. That same math of a frac node
actually lets us kind of like follow that logic through
and say, well if we know that this is
how big each tile is, nothing stops you
from getting these, and kind of increasing
and decreasing from either side, like squishing these UVs in
and by squishing these UVs in so that the value of zero
is here and the value of one is here,
so that further inside; that's how you would kind of
take that approach. That would kind of push those
floors a little bit together and kind of extend the walls. The advantage
of being able to do that is you can obviously
create walls going on in between the things, but it also means that you can
run that as an external mask that would act
as like exterior concrete. I believe the version
of the Material that I put up on Twitter automatically adds, instead of squishing the rooms
that adds concrete to the top and the left and right sides. So, it actually adds overall
[inaudible] size to like kind of add that extra
concrete on the other side instead of squishing
the rooms themselves, which is a very similar piece
of logic to what we're describing now. Any other questions
before I move on to the next? >>Victor: I think I
had one more here. Would there be a way
to combine different Cube Maps as to get even
more variations of interiors behind the window glass
in one master Material? >>Chris: Yeah. We kind of
started taking on the whole Ian Malcom thing of being so
preoccupied with whether or not you could, that
you don't know if you should. But if you're wanting
to do that, you could do it
in a few different ways. It depends on how random
you want it to genuinely be, but you could comfortably
just run a Lerp that switches because these here, this is a UVW
being fed into a Texture. There's nothing that stops you
from having another one of these and then lerping between them. You could do your cell selection
as to which face you're actually selecting
in a few different ways. Obviously interior Cube Map
is doing that itself by getting
the randomization here. So, if you want to follow
the logic that they are doing along
this true path, by the way,
if you click Hide Unrelated, I can grab just this true logic. If I was to get this -- I want to get that. I want to
get you and your true -- You're pulling in this data. So, how they're handling
that randomization is I think done over here, you could use that same logic
to kind of handle your own exterior,
secondary randomization that's randomizing between
these Textures themselves. That would be one way to do it. Another way, if you wanted
to really kind of make sure that edges of rooms were kind
of matching with one another, you could use
a Sobol randomization which gets you to the integers. It creates kind of a grid
of random cells and the cells themselves
would be the value that's being read into here,
sealed or floored. If this is an integer
being read in -- If I was to get
this and I was to -- That should -- Cool, let's hit Apply. It would help if I -- It’s giving me a random number. And I did the same thing again. I want to get you
and I want to seal you. I think that should work.
In my head, that works. Yeah, there you go.
So that's now actually getting a random number
for each single cell. We could use those values
that way lifting out of it as a way of interpolating
between different things. But because that's reading
in a grid position, it shouldn't be too hard
for you to read in corresponding
grid positions for each cell so that this random and this random have
the same seeded value, therefore they output the same
which would make the lights in each cell kind of match. I can't remember if I'm doing that
in my Twitter version or not. I have done that before though. >>Victor: If these cubes are not
aligned to the world axis, do the Materials still work? >>Chris: As in, like,
If I got this and I turned it sideways,
and did this? >>Victor: Yep, that was
one of the questions. >>Chris: Yes. >>Victor: There we go,
future of architecture. >>Chris: Anything else? >>Victor: I think
that was it for now. I might follow up
with some later on, but please go ahead
with your next example. >>Chris: Rad. Sorry,
I'm not quite running at 100%. The next thing that
we're going to do here is we're going to take a stab
at Post Processing in a slightly different way. I'm just going to just kind of
shrink this down a little bit because it's really
just bit of an eyesore. I will put another one in
because I need something to demonstrate the thing
that I'm about to do. For anyone
that's unfamiliar with it, typically when we create
a Post Process, this is -- actually, I'm going to roll back and explain
what a Post Process is. For anyone unfamiliar
with Post Processing, essentially how it works
is as the camera, it renders all of the stuff
in the scene it then runs an extra pass
at the end that allows you to change
like the exposure, it allows you to change
the color balance and all of that stuff, right? But one thing that's really
useful is we have this concept called Post Process Materials. A Post Process Material
is really useful because it allows us
to actually create a Material that is doing a series of
instructions on the final image. I'm just going to create
one of them really quickly now. I'm just going to create one, I’m going to call this
M_PostProcess. Let's just build like
an outline shader or something. We're going to build
a really basic outline Material. The first thing to do
is really straightforward. You literally get this
and you say, “Hey, you're Post Process.”
Okay, now you'll notice that that just ditches
all of your channels except for emissive color.
That's good. That's fine. Now as a quick thing,
if I was to get this, I'm going to go to Scene Texture and I'm going to select
Post Process Input Zero. Just for the time being,
I'm going to multiply that against --
Give me a color Victor. >>Victor: How about green?
>>Chris: Green. Okay.
I’m not going to multiply it because that's going
to look terrible. Let's do a Photoshop-like
overlay. There we go. That's now putting this image
over the top of it. See how that's going on there? >>Victor: Yep, we've got night
vision now. >>Chris: We have night vision
now. Oh that's why he chose green. Good job. We're now kind of like getting
that kind of blend going on and if I wanted to, I could hit apply and I could
push this out to the scene. I just click here and I say,
“Hey, I want to use an Asset and the one I want to use this
called Post Process Outlines,” which is now doing this.
What's really important here is if we want to, we could
create an outline shader. There are a few different ways
to do an outline shader. I'll create a reasonably
basic one, I guess. Let's just follow
through that logic. I'm going to us a node
called Sample Scene Depth. You could do this manually,
but I find this is an easy way to quickly set up
what I'm about to do. Imagine this node
given an offset, it will go ahead and it
will look at how far away the current thing in the scene
actually is. So, what is the depth
of the thing in the scene? What we can do that's quite
useful was if I wanted to, I could get this
and I could duplicate it. I'm trying to think of how
to demonstrate this next bit. Pretend we're sampling
this pixel here. If I wanted to look
at this pixel and this pixel and this pixel
and this pixel. So, the ones that are above, below to the left
instead of right, I could actually look
at all of those and see how different
they are to this middle pixel and by doing that I can see
whether it's an edge or not. Because if there's not much
difference between these things, then it means
it's probably a flat surface. But if there's a big change,
it means that -- if this is coming up
with a really low number and this one over here is coming up
with a really high number, I know that this is
really far away, but this one isn't,
therefore, there's a line here. I hope that makes sense. We're going to do that.
Reasonably simple. All you need to do is, there's
a nice cheap way to do it, use this
Blur Sample Offsets node, which already reads in X,
Y offsets of 0:1, 1:0, 0:-1, -1:0.
It's quick way to do it. So, I'm literally
just going to get this and be, plug in all of you,
and you're probably like, “Alright, but that sounded
like a reasonably complex task of evaluating these things,”
But actually it's kind of not. If I got this, all I need to do
is I add all of these together and then I see how different
they are to this one. To do that really just means, if I was to multiply this
by negative four because there's going to be
four of these things and then I add
everything together. >>Victor: I think we lost
your screen share there for just a little bit.
Okay. You're back. >>Chris: Did you get my audio
or just the screen share? >>Victor: I think
it was full hedge. It was only a couple seconds. >>Chris: Okay.
What was the last thing you heard? >>Victor: Negative four.
>>Chris: Okay. So, by multiplying this
by negative four and then adding
everything together, this negative four is now -- it's equivalent
of kind of adding this and then subtracting this,
adding this, and then subtracting this,
adding this, and then subtracting this
and adding this and that, subtracting this. By multiplying it by negative
four ahead of time before I just add
everything together, it means that I only would
have to add it one time, if that makes sense. By having
this kind of set up here, I'm getting all of those things and what that's kind of
doing for me -- I actually, you know what,
I'll just show you. If I get all of those values, and I was to saturate them,
look at that. >>Victor: Yeah, it looks pretty
cool. >>Chris: I look back in my world
for a second. This is what we see and that's
because that's creating a reasonably basic out -- You'll notice,
it might be hard to see because
it's getting pretty framey, but if I was to get this and it's getting a bit of jitter
going on, but you can fix that jitter by just getting
this Post Process Material and saying, “Hey, I want you
to happen before tonemapping,” And what that means is when the
anti-aliasing pause kicks in, it will actually smooth out
all of the stuff that you're seeing there. Is that visible? Can you see
what's going on there? Can you see a difference
in this? >>Victor: I saw the thought
a difference there, It was especially
clear on all the -- all the white ones. >>Chris: Cool. That's a reasonably
basic Post Process. Obviously, we could have
done this in a cool way, but this is a reasonably
basic digitized looking Post Process thing. Let's do something people
don't often do. Let's go one step further. I'm going to delete this from my
Post Process for just a moment and I'm going to go back to here
and I'm going to grab this. Actually, just before I do that,
now that I think about it, I’m going to apply
this again, sorry. One thing actually I forgot
to mention for any folks that are following along at home is if I add that image
that you just saw and I multiplied it,
one minus just flip the zeros
and the ones, if I multiply that against
the Post Process color, we're going to get that. Normally I would need to do
some masking that removes the background stuff as well so that we don't get
that happening with the sky box. But see how we now
kind of getting that applied to the standard imagery
that we would otherwise see. I figure I should probably show
people how to finish the out. So, we've got that applied. What I'm actually going
to do this time is I'm going to get all of this,
I'm going to copy pasta, I'm going to create
a new Material, his is going to be called
M_ -- I'm going to call
this Box Process Now, this is -- Box
process sounds -- >>Victor: Names are important. >>Chris: Yeah, I'm going to call
it Moveable Process, just as a reasonably
straightforward way to do this. I've got this M_MoveableProcess, I'm going to delete
the existing Post Processes, so we're back
to where we were before. What I'm going to do
is I'm going to get a cube and just drop that into my world and I'm going
to make this a big cube. One of the really cool things that a lot of people
don't realize is that when something
is running translucency, it can actually run
almost all the same logic that your Post Processes can.
So, check this out. If I go to this and I say to it,
“Hey, you are translucent,” and I'm just going to plug this
into the opacity and then I'm going to hit apply and I'm going to set this box up
to have that Material, I called it movable. Now when we look through
this box, we're actually seeing
that Post Process being applied to it.
Do you see that? >>Victor: Yeah. It's coming
through. >>Chris: One thing that actually
could be rad here is if I ditched that one minus, so we just have
what we originally had. I could run it like this. Actually, what would look
even cooler than what I'm doing right now, now that I think about it is,
I'm going to multiply that against like a rad looking color
of some sort, let's do orange this time.
That's right guys. Colour with a U,
livestreaming from Australia. This might look a bit
psychedelic. That's pretty cool. So, we've got that
going on there. With this, what we can see
at the moment, if I was to go into this queue, it's going to bleep
out of existence. Do you see that?
That's not good for anybody. So, if I get this, I'm just going to literally
make this two sided. By making this two sided -- >>Victor: If you didn't
catch that, Chris, do you just want
to show the checkbox again? >>Chris: Sorry.
It says two sided. >>Victor: There we go.
Under Details. Cool. >>Chris: And now if I go into it,
we’re now able to call a Get. We're not out of the woods yet.
We have this issue going on. Does anyone want to guess
why this is occurring? Victor, you got an idea? >>Victor: Well, the box
is below the landscape. >>Chris: Exactly.
It's only going to be two sided when we can actually see
what's there. Okay? So, alright,
that's a bit of a problem. So how do we solve this? Well, there's a really cheeky
way to go about solving this so please don't judge me for everything
you're about to see. Now that I've put that forward,
let's check this out. If I disable the depth test
on this box, there's a checkbox
called Disable Depth Test. This will now render
through everything. Give it a sec, if I go into it, I can now see
that that's occurring as well. Do you see that? >>Victor: It's coming through
a bit choppy, but I think we can see it now.
I’ve used that checkbox to do player names
on your team and such. You’re able to see them. >>Chris: Here’s the issue
we're going to get, see that? Not ideal. Fortunately, there is a nice
little thing that we can do, which is,
there is a thing called World Positioned Behind
Translucency. World Positioned Behind Translucency
literally just gets you the position
of anything you're looking at that's behind the object
that you're seeing. So, if it's a translucent object
instead of getting you the world position
of the cube surface, it would be getting you
the position of the grass or the building behind it.
Do you get what I mean? >>Victor: Yeah.
Just a quick request too, when you're doing
these quick motions and just showing
the little things, just go a little slower. Since we're a couple frames
behind due to the output rate. >>Chris: Sorry about that
everybody. >>Victor: Or you're just
too fast for us. >>Chris: No, I'm just trying not
to get into a political rant. I've got this world position, so if we get
that world position, the first problem
that a lot of people have when they approach
these kinds of issues is, they don't approach
these problems from the perspective
of localizing it. But if we assume that
the inside of that box is 0:0 and then it's a local position, all of the math that I'm about
to do becomes much easier. So, there's a thing called
a Box Mask 3D and what is,
it looks at to two things and it just works out if they're
inside of the box or not. I’m getting the world positioned
behind the translucency, and I’m going to tell it
that the location is 0:0. Now this isn't going to work yet
because I am missing something. This world position
is in world coordinates, but if I transform its position
so that it's in local space, I'm going to tell it that the information
coming into it is absolute and I want it to come out
relative to the box that I just added, I then just need
to get these bounds and set them to the objects
local bounds and I think if I plug this
into opacity, we should see what we want. The bounds need to be the size,
not the minimum. Let’s do this again. What we have here,
I can now move, I'm going to put this box,
wherever it goes, provides a Post Process
just to that region that you can physically see into and then when I get in there
with it, I can be like, oh, Post
Processing and then I come out and it's doing its thing. For that kind of effect, we could use this
for all sorts of stuff. I actually gave you a
reel earlier, do you mind playing a thing for me
that I sent you earlier, can you play the one called Movable
Post Process through your feed? >>Victor: Yup. >>Chris: Thank you. So,
I'll talk over the top of that. This is something that I did
as a quick demo when I was using
this feature elsewhere. I whipped up a procedural
level creation tool so you could just walk through
and hit buttons and it would play intelligently, place bits and pieces
in the world. I'm actually using that
Post Process concept of just like it's not actually
a Post Process, like it is just translucency
doing some fancy stuff, but using that
as a spawn technique that grabbed all of
the regular Assets in the world and quickly just had them. The box is actually
just shrinking down. If you see what's going on
in that scene, can you see how the box
getting smaller and smaller? That’s actually all it's doing
and it creates a really nice little spawn effect
that you could use without having to do like a ton
of extra pixel instructions that are actually on
every single unique Asset within the thing
that you're spawning. So, it's kind of nice to be able
to just spawn stuff and be like, “You're what I care about.” If you wanted to, you could
probably use like custom depth or something like that
as a nice way of masking out specific Assets
too with less overhead. This can be approached
in a few different ways, but I technique like this
is very rarely actually used, but as far as I'm concerned,
this is really quite valuable. If you don't mind jumping back
to my share, the last little thing I'm just
going to show you on this front is if I go to this Asset,
if we went to -- there's this value
called Edge Falloff. Now this isn't going to work
out of the box. I’m going to create a scaler
called Edge Falloff. You know what, I might actually
show you the problem first, so you can see what occurs. Watch what happens if I plug
in an Edge Falloff of 200 straight into this. If I was to just read in
an Edge Falloff of something like that and just plug in a value
of like 200 or whatever it was
that I just did, it's not going work. It's not going to properly
give me the size of what I actually
need it to be. What I actually need to do here
is I need to get the bounds and I need to subtract
the Edge Falloff size from that so that it actually
brings it back in. So hopefully this will make
sense to you in just a sec. I'm going to get that value. I'm going to subtract.
The Edge Falloff is a radius. So, I need to subtract it
from both sides. I'm going to multiply by two,
I think this will work. That did not work. It actually did a little bit,
let me just bring the Edge Falloff
to be a much smaller number. So that's now kind of creating
that nice fade in and fade out of the Post Process.
Do you see that? >>Victor: Yeah, I see it. >>Chris: By kind of subtracting
that first, it actually gets you
like a nice blend. I found some pretty cool stuff
with this. I put together a tune shader
a little while ago. It was really cool
to like walk into a tuned world and then walk out
of the tuned world. You know what I mean? That kind of stuff is really fun
and you can kind of create some really unique
and interesting effects. Being able to kind of like walk
into something that suddenly makes
the whole game experience, just feel like something
you haven't really seen before instead of just changing
the camera effect, it's kind of a nice way
to handle all of this. I might do one last little thing
just because, I'll be honest with you, I'm kind of running
out of steam. >>Victor: We're getting
short on time here as well. Did you want to show
the mannequin there and the World Position Offset? >>Chris: Yeah, that's what I
was going to do really quick. >>Victor: You're doing
well, Chris, considering
how late it is there. We can all see
what time it is for you. >>Chris: What I'm going to do
this time around is, I'm just going to mess
with the existing mannequin. If I was to look
at the existing mannequin, which is opened up
on a different screen somewhere, if I was looking
at this existing mannequin, this is what we've got. Now remember earlier I used
the World Aligned Texture node to project onto this Asset? So, if we were to do the same
thing with our mannequin, we would actually have
a bit of a problem. I'm going to ignore all of this. I'm going to get everything
that's coming out of that and I'm going to call
Set Material Attributes. I'm going to get the base color and I'm going to set the base
color on the mannequin to be -- sorry, I'm just going to
get a noise Texture. I guess that'll do. By default, we're going to have
a bit of a problem. If I was to multiply this value
by dark red -- Nah, I'll go for red,
just so it's visible. If I was to overlay this, Victor, you there?
I just feel like -- >>Victor: Yeah, we're here.
>>Chris: Cool. Alright, so I'm just
overlaying this onto it. We would have a little bit
of an issue. I'm just going to hit apply and I'm going to look
at my mannequin. >>Victor: Waiting for the
shaders. >>Chris: That looks okay. Right? It looks okay-ish.
The problem is this, if I was to apply
an animation to him, did everyone see
what's going on there? Because it's a projection
in world space, the Asset is kind of shimmering.
I hope my connection’s able to -- >>Victor: I think coming through.
Yeah, we can see the Textures sort of staying still
in world space, whereas the mannequin is moving. >>Chris: Wonderful. A quick little thing that I just
want you to be aware of is, for something like that,
all you need to do is uh, if I call
a Pre-Skinned Local Position; Pre-Skinned Position actually
gets me the world position when it was in its t-pose and you've got to use the vertex
interpolator for that. Watch what happens
when I plug this in instead. I'm just plugging that in.
See what that's doing now? >>Victor: In just a moment.
There we go. >>Chris: See what's actually
going on there? That Pre-Skinned Position because it's
the T-pose position, I can actually just project
onto the Asset and even though
it's moving around, you're not getting
that shimmer anymore. Victor, it might be a good idea, I gave you some b-roll
on this effect just before the livestream
started just in case, do you mind playing it for me? >>Victor: Yeah,
what was the name of it? >>Chris: It's
called local bounds. >>Victor: Coming right up. >>Chris: Cheers, peers. You can see on
the left-hand side is what you would normally
get and on the right-hand side is the Pre-Skinned
Local Position. The cool part of this
is -- If you drop back to my screen
now, if that's okay? Now the cool part of this
is, at the moment, this is using the world sizing,
but if I wanted to, I can actually get
this Pre-Skinned Local Position and I can divide it
by the Pre-Skinned Local Bounds, which divides it
by the size of the Skeletal Mesh when it was in its t-pose and I'm going to set
this Texture size to be a value of one.
What this actually goes ahead and does is by dividing it
by the Pre-Skinned Local Bounds, it actually aligns the Texture
to be projected according to the size
of the Skeletal Mesh. You don't have to be finicky about
what exact size the Asset was, you can just have generic skins. Now this is really good because think about
if you're making a video game and you want it to have
camouflage on the character, or you wanted to project
like a cool decal onto their t-shirt
or anything like that, being able to just read
that information in, means that what you end up with
is a really elegant and quick way to generate
a ton of content. You can really
quickly generate uniforms based off of camouflage Textures
being read in. You can really quickly generate all sorts of weird
and wonderful patterns that are just being brought in
and that really takes -- I'm not saying to do that
in a game for every skin, but let's say you had a game
that you wanted have hundreds of thousands
of loot drops for the player. When you're able to use
something like that to just really quickly generate
a weapon skin for literally every weapon, by just projecting
that particular camouflage type into that particular region
on the Asset. That's really valuable
because you can then use that for your smaller Assets and then you can still do all of
your custom detailed hero Assets that look really, really nice
that are the bigger features. This is really important
for developers because we want you to not get
bogged down by workflow issues that just shouldn't exist. When you're able to kind of
get this Pre-Skinned stuff and just project things
onto Characters or onto weapons or to items in general,
you can create a lot of value that you can keep
rewarding players and just keep giving them
all sorts of free stuff in your game. Keep giving them drops,
make things look interesting and that's pretty great. Please keep in mind that the
Pre-Skinned Local Position and the Pre-Skinned Bounds
are pretty valuable. You can see the Bounds
here has just made this go between zero and one and it's just kept it
to a nice size, a nice scale for the projection,
which is great. Last little thing.
I'm kind of out of time to really do
the procedural gen stuff that I was going
to quickly cover. What I'm just going to do
is load up a project and just show you
some of the tools that exist so that you can jump into that. My apologies for anyone
that jumped into the stream. The prop gen stuff
I originally intended on doing, but it’s kind of involves
a lot of moving around and would have involved me
moving quite quickly and I wasn't quite comfortable
doing that and not slipping up
at this hour. This world here is just being
procedurally generated. The forest has been brought in using the Procedural
Foliage Volume tool. This is available for everyone,
just go to Edit, and then go to Experimental
under your Editor Preferences. So Experimental and you'll see
that we have Procedural Foliage and the Procedural
Foliage System literally lets you set up
rule sets that tell it -- you can literally set up
this Procedure Foliage System is looking for these
three types of trees and if I open
the each of these trees, each of these trees
just has some rules that I've set up that say
how much shade they cast, whether they can grow
when they're in the shade, how many seeds they spread whenever
are not pollinating -- seeding? What's the word here? You set up all of these rules
for these trees and if I was to go back here
and simulate it, it actually regrows that forest
and it throws them all in. This is a really
useful little tool because it actually syncs up
with the Foliage tool. If I generate the forest
and then I'm like, no, you can just remove those trees,
you know what I mean? So, we can still get like that nice little
custom layer in there. It just saves you having
to paint all of that because,
why would you waste time? If you can just kind of drop
that straight in and have a good time, it just means
you can just like be, give me a forest,
I'm going to carve out the area that I actually need,
let's all move on with our day. That's, that's the first thing that I kind of
wanted to touch on. Now, the next thing that
I quickly wanted to touch on is this landscape itself. I'm just going to hide the trees
for a second. This landscape itself is actually
completely procedural as well. So if I go to
my Content Browser, I'm going to run the widget,
see this thing up here, I got stuck without internet
for like six hours and I was trying to look for
a project to just knock together while I was
waiting for something, that's a long story, and this was kind of
what I cooked up real quick. This tool, if I click here
and hit Generate, that generated, see that ditch
corresponds to here, or if I wanted to have to
pull this across and be like, “Hey, let's generate that,” and you can see here
that that dip is there. By the way, the Foliage should
just stick to your landscape because of how
the Foliage System works. So, it will actually, even after
changes to your landscape, it stays friendly
and as of 4.23, you can also have generic
actives as follows too, which was really great. Anyway, back on track. So being able to have
this kind of system is reasonably straightforward. Now this looks like something
that would be complex, but it's actually really not. I'm going to open up
this Editor Utility Widget, this doesn't involve
a single line of code. I'm going to open up
this Utility Widget, this is all Blueprint and
I've literally just gone ahead and created the widget by just
dropping in a display image, which is the preview
and then down bottom I have generated like
just a bunch of sliders and a big old generate button, but if I look
at that display image, it's using a thing here
called Preview Map. I’m going to open up
the preview Material. So, this is what you see on
that little clickable area. This literally takes in a
Heightmap called landscape RT. I've been a little bit cheeky
with how I've set this up to get contour lines working
so I could actually see what the height
was really doing here, by multiplying it by 10
and then fracking it, we're going to preview that.
By multiplying it by 10 and then fracking it,
it actually gives me a nice gradient
between all of these layer lines and then by kind of just doing a
little bit of extra math there, I’m subtracting 0.5
and multiply that by at two, which kind of gets
the bright areas in the middle and the black areas
along the side and then heavy contrast,
which gets me this. Then I'm just literally lerping
in some green. That actually gets me like
a nice little preview screen. This kind of math
isn't too complicated and it's actually just seeding
off of the render target. Now if I look at
the actual logic involved; I'm going to get back
to my graph, this is the entire thing.
It's actually reasonably straightforward and most of it,
like these things down here, if I zoom in, all they’re doing
is saying on value changed, call this Update
Variables function. That’s all,
all of those are doing. Every slider, every time
I change the slider it's just
updating the variables. I hope my stream
is keeping up with me. >>Victor: Yeah,
it looks alright. >>Chris: When it calls Update
Variables, all it's doing is,
when we first create the thing, I'm creating a dynamic Material. I'm getting that dynamic
Material, and it's a really simple
landscape generation Material that I created. This is literally just
a couple of noise functions to create something like this. And I've added like
a scale value to make it bigger and I've added an offset
for different things. But it's actually
reasonably straightforward in terms of what
it's actually doing. All I'm doing in this widget
is literally getting this and I'm calling, “Hey, I want to update
all the variables. Your offset is just equal
to what the slider has. Your scale is equal
to what this scale slider is. Your reduction amount is what
I'm using for that black area. Wherever I click
your reduction size is this.” I'm literally
just running through and just setting up
all of these things and just letting it do its job. The only thing that's unique
here that's worth noting is, see this Draw Material
to Render Target? So, this literally
takes a render target and I'm getting this Draw
Material to Render Target and I'm automatically then just
outputting the landscape render the target from that. So that's creating like
a black and white Texture and after that creates a black
and white Texture for me, anytime anyone hits the create
landscape button., so anytime anyone clicks
this big old generate button down here,
what that's actually doing is it's regenerating
it just in case and then I'm being
a little sneaky, I'm saying Get All Level Actors and just grab
the first landscape you find. Then we have this function here, which is really where the meat
of what we're doing is. A lot of people are unaware
this even exists and it's called Landscape Import
Heightmap from Render Target. By importing the Heightmap
from the render target, literally just takes a black
and white render target and sets that
as the height of the landscape. From here, I've done a couple
of other little things. I actually generate
a grass Texture as well based on the landscape. So, I'm actually
generating grass patterns based on the Heightmap. So, this is a separate Material,
again that's getting rendered. This essentially
it's just looking at the slope and saying,
“Hey, how slanted is it? It's pretty slanted,
maybe make it dirt.” So essentially this is like
a static version of that cliff painting
that you saw earlier and the advantage
of making static instead of making this a world
projection like I had before is it's just
a little bit cheaper and it's a little bit easier
for artists to modify because they can go over
the top of that and be like, “Look, I can see
that it's generated and added some no grass there, but I really wish
it had some grass there.” We can see that like by having
that as a static thing that initially gets pushed out. It's worth noting that
this workflow is good right now, but it's not really
where you want it to be. And honestly it's why after
thinking about it a little bit, I was a bit hesitant to do this
as the core of the livestream because if you go
to Edit, Preferences and you go to Experimental, there is the Landscape
Layer System, it’s something that's
experimental at the moment, which allows you to create
Blueprint brushes. Being able to kind of use
the experimental landscapes whenever you're creating
a new landscape, it's going to mean that render
targets like this I've just done on separate
landscape layers, rather doing them using
the functionality that you're
seeing at the moment. That's in 4.23 at the moment. That's something people
can tinker around. >>Victor: That’s the Editor version
you're working in now, right? >>Chris: Yeah. This is all just vanilla .23,
nothing fancy about it. That general procedural system
is probably much better for a landscape discussion
in a future version of Unreal once those layers are kind of
a bit more stable and everything's working. >>Victor: I'm sure
we'll cover that. >>Chris: I would hope so. Just a quick heads up for anyone
that's fooling around with this stuff, be aware that if you go
to Edit, Plugins, if you go here, you'll see Editor
Scripting Utilities. I think it's disabled
by default, but it adds a ton of extra
functionality that you can do. So, it adds a ton
of extra little calls that you can do
within Editor Utility Widgets and the other one
that's really powerful is, I think its Blueprint Material
and Texture Nodes. So, this gives you the ability
to do Blueprint only nodes for creating static Textures
and all sorts of stuff. This is really powerful when you're doing procedural
landscape widgets and such. >>Victor: We had a Nick Pfister
on a couple of months ago showing off how to use
the Editor Utility Widgets. We also have
good documentation on that. Pretty straightforward,
how to get started. >>Chris: With that said, I think
that's probably the core of what I really
was going to be covering today. I guess it's a good idea
for us to kind of go into to a question. I might stop sharing
and just switch back to video feed and deal with my,
my 5:45 AM mug. So, I’ll switch over
and we’ll go from there. If anyone has questions
from this point, I guess now now's
the time to do it. >>Victor: We actually have
a lot of questions. I don't actually think we'll be
able to go through some of them and I've been paying
attention here. Let's see if we had some earlier
that I’d be able to grab, just one or two,
there's quite a few. Would you be willing
to maybe go through them and pick some out that are
relevant and answer them later? Just on the forum, you can get
some sleep and then we'll take -- How about we do that? >>Chris: No, I can answer.
I can still do 10 good minutes. >>Victor: Let’s go through
a few then. What a champ. Is there such a thing
as vertex normal local space? >>Chris: I believe so, yeah.
Let's find out. Yeah, Pre-Skinned Normal. >>Victor: The answer is yes.
What was it called? >>Chris: Pre-Skinned Normal. By the way, a little thing,
and I know this is weird, but a lot of people
don't know this. Shift D aligns nodes
to the right side. Shift A aligns them
to the left side. This works in Blueprint as well. That's just a nice thing
to just keep your nodes. There you go.
So, align left, align right. And if you ever have nodes
plugged in like this, I'm not sure if you noticed
that I was doing it, but if I was to click these
and hit Q, it will straighten them
and just auto-fix your stuff. Just a lot of people aren't
aware of Q, Shift A, Shift D. I just realized that
you can't see my screen. >>Victor: No, we cannot. >>Chris: Sorry. >>Victor: Next time.
That's also what makes going through the questions
a little hard because I know
that you're just like, “Oh yeah, it's right there.
Let me show you.” >>Chris: I'm going
to ditch my video. Let's go back to the screen
while I answer questions. >>Victor: Here we go.
>>Chris: If you hit Q, they'll just
straighten your notes. If I grabbed things like this
and you Shift D, Shift A. >>Victor: You get dancing nodes. >>Chris: Yeah, dancing nodes. That's just a good way of
just not having to deal with straightening stuff
manually. >>Victor: Or marking them,
right-clicking and selecting the option
in the menu. >>Chris: Oh God, yeah.
Hotkeys, man. The nodes that I was
mentioning before, which I put on the screen
and then straight up forgot that I had was just a Pre-Skinned
Local Position, Pre-Skinned Local Bounds. >>Victor: Why is it called
vertex interpolator if it's a pixel shader
only node? >>Chris: It’s interpolating the values from vertex
through to pixel, I guess. You'd have to ask the person
that had named it, but I believe that's why. I think it's getting the pixel
values by getting the verts and then interpolating
to get the pixel position. The equivalent
of the pixel side. We got documentation of it, but yeah, I believe
that's why it's called that. >>Victor: They were curious
if any of this content might be available
after this stream? >>Chris: The buildings
that I did are public. I put them as files
on my Twitter feed. What was the other ones
that we had here? What else did I do tonight? Post-process stuff,
I didn't do the Post Process, but that's pretty
straightforward. I can send that through
and you can post that. The Material itself was super
straightforward though. >>Victor: It's nice sometimes to
be able to look and break it down. >>Chris: If you want to,
I can push that through. It's not a big ask. I might do it after I've slept
though if that's all right. The Material generation stuff I can send through to you
if you want. It's kind of hacky because I'm
just like throwing it together, so it was pretty ugly
to look at. I spent zero time making
the widget itself look pretty. I was like, oh, functional,
and then just called it. But if you want,
I could post that. >>Victor: We'll take a
look at it after stream and see what we'll
be able to share. These are all specific to what
you were doing right then. Is any of the operations you've been doing
in the Material graph here SM5 only or do most of them
work on other platforms? >>Chris: That's a good question. I don't believe so, but I don't
want you to hold me to that because I'm not too sure
about these ones. I haven't run them
on a non SM5 device. I think it's all fine though, but you'll need to,
don't hold me to that. You'll need to check it
yourself. >>Victor: Yeah, it's fairly
easy. You can turn on the
ESGL 3.1 preview and just apply the same
Materials in your scene there and you can check it out. >>Chris: I’m not too sure
about the building stuff. Your interior maps might
have issues, but I doubt it. I'm not going to
answer that question just because
there's a bunch of Materials they have that I would have
to go through to verify today. To my knowledge it's fine, but
you know what I expect ES 3.1 to do versus
what it actually does, ever anything that's similar whenever it comes
to advanced effects. >>Victor: Sounds good.
I think that was it in terms of what you covered and
questions regarding them. It's getting really late and I know that you should try
to get a little bit of sleep. Chris actually went to the bed and then he got up for
the stream and I recommended him
to go back to bed. True champ. We’ll see
if we can get you back here. >>Chris: That’s alright.
It’s the job. That was it for the,
through the core stuff, I guess. Are there are no other
big questions that are really kind of popped up that you
can think of that are -- that are worth jumping on? >>Victor: In regards to -- Some of them don't have
enough context for me to know precisely when they when
they were asked, unfortunately. So, it's just can you show
some slight variation? Not entirely sure unfortunately
when that was asked. So that's a good tip
to all of you. When you go ahead
and phrase your question, try to be a little bit specific
about the question maybe you mentioned like,
“Oh, you know, when he was doing the landscape
Material or the Post Process,” and then it'll be a little bit
easier for us to know when that was asked. But yeah, that's really cool.
A lot of cool stuff. If you want to see
some more work on this, Chris's Twitter profile
is another good resource. You can go follow him there
where he frequently tweets about all these cool things. I think its kind of a cool idea
that we get together and we show some
a little bit more, exactly how you did some of
these things on the livestream. I think that's
a pretty cool concept. Sort of grabbed the best
of what you've done recently and then go over it. I think with that said,
thank you so much Chris for coming on
and doing this with me. For all of you
who are still watching out there and stuck with us for this long,
I'd like to let you know that we are doing transcripts
of all of the livestreams now. So, on YouTube after a couple of
days whenever the people are -- it's actually done by a person,
so they go through, they listen to what we've said
and they write up captions and that transcript we provide a link to in the
YouTube description and it's a good way for you
to go and Control F and search for specific keywords if you were interested in
a specific part of the stream. >>Chris: I'm so sorry to whoever
has to translate my rantings when I'm trying to work out
what I'm doing. >>Victor: That’s on the
captioner's job later on. Don't worry, you don't have
to worry about that. As always, Amanda went ahead
and posted a link to the survey. I think we're going to go ahead
and do that again. If you fill it out
and drop your email in there. Let us know how we did or you
would like to see in the future. I generally go there and use it
as a reference to what you guys want to see on the stream
and if you enter your email, you are part of
a little t-shirts sweepstake that we're running.
If you're interested in meeting other developers
around the world, go ahead and go to
unrealengine.com/user-groups and you'll be able to see if there is a Meetup
group near you. If there is none
and you have a group of friends or people there that are excited
about showing off what you do. Go ahead and send us an email
to community@unrealengine.com and we can tell you all about what it means being
one of our meetup organizers. Always make sure you check out
our forums. We have an unofficial, Unreal Slackers
channel on Discord, where people are really helpful. You can ask your questions
or if you have any problems, let them know and try to
get some or even just feedback. Facebook, Reddit,
you know the deal. That's where we're at. Our community spotlights
that we have every week, we usually source them
from our forums, but you can also just go ahead
and ping us either on Discord or like I mentioned
community@unrealengine.com. It's a good email address
just to let us know about what you're working on,
it's always really fun to see. We're still looking for
more countdown on videos, so 30 minutes of content of you
working in the Editor, speed it up to five minutes and send us that with together
with your logo, separately, so that we can do
to cool little countdown timer, then we might go ahead and put you on
one of our countdowns. If you're streaming on Twitch,
like we do, make sure you use
the Unreal Engine category so that we can come
and have a look when we have a couple
of minutes in the day and as always follow us
on social media. Chat, please give Chris your biggest
thanks for sticking with us for two hours now and technically you've been
with us since the prep, which is an hour
before this, sorry, half an hour before to stream. So, two and a half hours,
big up to Chris. Thanks again. Will you be at Unreal Dev Days
next week? >>Chris: No, I won't be. Actually, on that note,
I will be at GCAP, which is Australia,
New Zealand events, for any developers
that are at GCAP, next week. So next week is Melbourne
International Games Week. We have a meetup going
on Monday night, which you can find through
the Unreal Engine Melbourne Meetup group. We'll have a couple
of devs there. >>Victor: That’s meetup.com?
>>Chris: Yeah. So just hit the Melbourne meet
up and you'll see it there. But we're having just
a general mixer, so it’ll be a couple
of couple of Epic folk floating rounds that could
hopefully answer some questions and for any of the Australian
devs that might be attending, feel free to send me
any videos that you've got because we're running it
at an E-Sports bar in Melbourne, which has a bunch
of TVs kind of streaming. We're going to be
running developers’ games on each of those streams. So, if you've got either
a trailer or anything like that and you want to put in there,
or if there's a known stream that might be running,
please let me know. >>Victor: That’s awesome.
>>Chris: Thanks for having me guys. >>Victor: Always, Chris. If you are
at Unreal Dev Days next week, we will be there,
both me and Amanda, so please come say hi. With that, next week
we will be -- Actually, I have -- there’s
this tool called Mod IO. I’m not going to talk too much
about what it is right now, because I want the people
who actually made the tool, but they will be
on the stream next week talking about how you can make
a mod framework for your game. That's going to be really exciting.
That’s Thursday next week. Until then, goodbye.
Thanks everyone for hanging out.
