Let's Build the RPG! - 13 - How to Build a Scene from Scratch in Unreal Engine 5

hey guys today i'm going to show you my entire process for building a scene from scratch let's get to it [Music] hey guys we got a little bit of a monologue to start this episode so i ask you to hear me out so first of all i am not an artist i have zero formal art training past elementary school so what that means is if i can do the stuff that i'm going to show you today then you can do it too but what i lack an artistic ability i can make up a lot for in preparation with every scene that i'm looking to build i try to do as much preparation ahead of time as i possibly can and then with that scene i start at the lowest level of abstraction and i work my way up so i mean that which requires the least attention to detail and so i start with landscape then go on to things like static meshes then go on to foliage then go on to effects which we haven't even done yet and then dynamic content like the stuff that's really going to come alive in the scene and finally the look and feel the final polish camera effects whatever they may be on top of the scene and some of this we're not even going to get into this episode but that's okay i think you'll have a general sense in my process as we go through it and so why do i consider the landscape the lowest level of abstraction so take a look at this scene right here can you even tell what the landscape is do you focus on the landscape at all so basically the whole point is as you work your way up these levels it becomes more and more critical to get design choices correct but on those early stages like the landscape doesn't have to be perfect and you can learn from the process as you go along and in general i always try to get to a b-plus level of quality before i go up to the next level and the reason for that is i can usually get to a b plus level of quality in one to two hours and i found this out during high school about to lay some experience on you that i'm sure every high school teacher would profoundly disagree with me on and that is when i got an essay in high school if i worked my ass off for like two weeks straight i could probably get an a minus or an a but i found that if i just got it done in the 45 minute study hall period right before the english class i could always consistently nail the b plus and this taught me a lesson that has really served me well throughout my life it's really hard to get to an a quality but if you get fast at doing something pretty well pretty well could be better than most people's best especially if you do it a lot so that's what i recommend for building a scene build a lot of scenes and work towards a b plus level of quality and that b plus that standard for you is just going to keep rising up as you get better and better and so my process is i always have three environments that i'm doing things in i have an experimentation environment i have a rehearsal environment and i have our actual real game environment in the experimentation environment that's what i'm going to show you first is i'm just playing i'm just trying things out that i think might work and most of it straight up doesn't and then when i get to a point where i'm confident enough at every level on the scene that i'm like i think i can actually build this then i build it in the rehearsal environment but i do as much as i can in the experimental environment to the point where i'm supremely confident i can get to that b plus level of quality in the real deal and for the main environment you know typically i'm not even building in that if the rehearsal goes well then i could just copy the environment and that becomes my new main environment so it's always this pristine copy that just has the latest and the best of whatever i've done so before i really get into the example let's talk about my preparation steps that i take to build out a scene and for this i really wanted to emphasize the before in before you build there's a line in the movie spy game that i really love where robert redford says to his secretary when did noah build the ark before the flood so you could do as much as you possibly can before you even start building the scene just having a really strong sense of how it's all going to come together and the more you do that the more you're going to know going into it will this scene actually work so step one planning your scene here's the question what purpose does your scene actually serve like what is your scene going to do and this is hard right so you have to have a purpose for a game which is your highest purpose for that game and then every scene that you build into that game has to serve that primary purpose up to your game and every aspect of that scene has to serve that purpose of the scene which serves the purpose of the game and all of it should be pointing to whatever your point is for the entire experience of the player so once you have your purpose then you can begin identifying assets really those parts of your scene that are going to serve that purpose and so these are things like landscape materials meshes foliage background things like the weather of the scene even stuff like that and in general at this stage i try to go out of my way to say like i don't think this will work but i'm still going to include it in the bucket of things i'll try just to be sure it's not going to work and as you get better at this you're going to be able to narrow down that bucket more and more and that's going to speed up the process but to start you want to throw everything you could possibly think might work in the scene into that scene one lesson that anybody will tell you is it's way easier to cut down stuff than it is to expand on something that's not all that good to start with and so how do you identify what to cut it really it comes down to that b-plus standard that i was talking about and you'll know it when you see it like if a certain asset doesn't fit in with the rest even if it looks good in and of itself it's like that's going to mess up the scene cut so initially you error on the side of including it in for evaluation but then as you're cutting you're ruthlessly cutting and there's some things that occasionally i'll see something and i'm like you know that'll work somewhere else really well but it's not going to work here so i'll cut it but i'll keep track of it elsewhere and then when that situation comes up then i can reuse it and the stuff i look for is like does the color work with the rest of the scene is the style right is the feel of it right once you contrast two things you can easily see like oh i like this more than i like this and then you can cut one so then once i've got what i consider a good starting point for that scene then i'm organizing that it's like what do i want serving as like the main center pieces of the scene what's supplemental and all that organization it really goes with the theme with the purpose of the scene and so then once you've organized it a bit then you're ready to experiment and this is where you try out all sorts of designs and this is going to be the majority of what i show you today all the experimentation all the things that ultimately i decided for this scene for our garden fire quadrant nope we're not going to do that and this really is what play is it's like trying things out what's fun what works that's what we're doing here and for that experimentation phase i take that really seriously i try to do it when i'm most fresh like early in the morning got my morning coffee i'm ready to go if you're a night person might be that time for you and then at every step of this process so initially when you're comparing stuff you're cutting down to a b plus then when you're experimenting you're going to cut more so you're basically winnowing from like 10 things down to eight things down to five things and once you get to that solid sense of what's going to work in that scene then you're ready to go when you hit that b plus level don't even worry about getting it to an a it's like let's go let's build because as you're building you're going to have more ideas along the way and typically it takes days to have new ideas like sometimes i'll have a dream or a thought will come to me in the shower or on the walk and it's like oh obviously that's a better way of doing it and that's exactly something that you're going to see with the scene today so let's get into it so with our last quadrant of our garden this is the fire quadrant so the first thing i'm thinking about are what are the colors what are the themes of the quadrant really it's going to be a lot of reds oranges stuff like that so my mind's initially going to the foliage that we need for that so i would go into 3d plants and the flowering plants look like the best candidates here and so i'm basically downloading a whole bunch of these just copying the asset ids downloading them into the experimental environment because the goal then is to then put them all up against each other and see which i like the best the nice thing about qixel bridge is i have a limited set to choose from so i've got our library and now we just got to narrow it down and so thinking for the ideas of the outside of the fire quadrant this was a little bit more challenging so initially i was thinking of some kind of wheat but it is a garden i don't want it to look like dead leaves so then once i get all the assets we're dragging them all in i basically sort them from tallest to shortest and before i even experiment with like different textures and different settings for each of the assets if there's any that's straight up i just don't like well we delete them and then for some if i'm not convinced if it's going to make the b plus grade but i think maybe changing the color or changing the size might do it then i go into the texture for example then i could change the color and see what that looks like so my initial idea for the design of the fire quadrant was this cake-like structure of rings and then i'd have paths sort of in between the cakes and maybe some of the cakes are oblong and different shapes and stuff like that so i was experimenting with like how many layers of flowers do i want and to fix the sunflowers facing in the right direction then i just cranked up the volume but the whole problem with that is they still don't face in the same direction and then you've got really high density which is going to lower your frame rate so i tried other flowers but same kind of problem so then i was playing with how wide each ring should be what the ring should actually be the nice thing about that is i could easily erase the center and then i would just populate with a different kind of flower but the other thing i was finding with this is that the foliage is super dense and i couldn't use this trick for erasing the center if i did a different shape other than a circle so the problem with these flowers is the sheer number of triangles to some of them and that's what ultimately pushed me in a different direction and i didn't come up with this idea right away but in browsing other assets i realized oh geez there's all these awesome sandstone rocks i could totally use these because they basically fit in with the motif they're like this deep brown and i could change that into a more orangey red pretty easily so i downloaded a whole bunch of these different sandstone assets just wanted to drag them in change the size change the scale get a sense of could they all work together if i position them in different places across the fire quadrant and also i was thinking because the rocks are nanite they're going to be far less performance intensive than a ton of foliage in a cake-like structure so then i could just line the rocks with foliage and have a similar kind of cake-like effect but it's going to be different for every single rock that's the thing i'm often stealing ideas from ideas i've already seen whether it was an art whether it wasn't well cooking in the case of cakes basically anywhere you can get inspiration so this one here i ultimately realized just too skinny it's not gonna work i tried making a copy extending it out nope that didn't work i'm gonna delete that one and so once i narrowed down the rock formations to those that i wanted to actually experiment with then i could drag one in and this is where i'm going to use the different foliage types to just experiment the cake effect but before that i just disable the blending with the landscape because if it's surrounded by foliage we're not going to see that anyway and i also reduced the number of flowers i was going to use and i really like these amaryllis flowers so i started with those then we moved on to this crouton asset i don't know how to pronounce these things but that's okay luckily i'm not a botanist but at this point i'm basically playing with the density you know trying to get the right balance because i want it the least dense possible for performance reasons but still dense enough to be a good covering and it's also important to get the scale right here right to make sure that each layer of the flowers is slightly taller than the last one so it's going to be clearly be visible and for this last layer of flowers just not colorful enough so i went into the texture just need to adjust that to be a lot darker and so as i'm doing this i'm just writing down in a document all the settings for each of these assets that i need to actually change and then that's what i'm doing in the next environment and then i was thinking about this a lot it's like what's going to be in the center of the quadrant the fire quadrant so i needed something striking something that worked with the other motifs and so what i was thinking about was some sort of unique or special rock formation like this spire here but that spire wasn't going to work because it's a two-dimensional asset so i started browsing for other possible rock formations things that were unique that looked special i could scale them up in the center of the quadrant and would serve kind of a symbolic artistic effect but in short each of them turned out to be a bust i realized right away that it just wouldn't match the rest of the motif of the scene especially if the color scheme is so off like this it would be such a contrast to the other rocks in the scene like sometimes you want that if you want to really highlight a particular asset but if it doesn't work with the style then the highlighting is a negative highlighting and not something that's actually adding to the value of your scene so what i ultimately decided to do was to stick with the sandstone structure and i realized i could take one of these larger formations one of these larger hoodoos and i could duplicate it a few times change the rotation around just give it a particularly unique and pronounced effect so it'd blend in with the rest of the scene but it would still kind of stand out in the middle relative to the other formations then i'm just examining it from all sides making sure it looks pretty good it's important to get those center pieces you know those parts you're seeing that are really going to be focused on by the player it's important to get those right then we were on to our fire quadrant edge i really love these cana lilies i loved everything about them except the dead leaves at the bottom but the first thing i was going to do is just change the color see how that looked to be a little bit darker and now i was going to attempt to try to cut off part of the mesh in the model editor mode this is also helpful for performance reasons like if you can get rid of part of the mesh you're going to have less vertices and less triangles to worry about the problem was i tried using this poly cut and that just wasn't working it was cutting off part of the leaves that were alive too then i played with this tri-select feature and that worked great except for this one tiny dot no matter how close i zoomed in i couldn't delete it and then i was like is the player ever going to actually see that tiny little dot of the cana lily but this tri-select feature in the modeling mode yeah this is a really useful feature to be able to remove part of a mesh quickly but when i saved it i noticed that the parts i removed were still there i realized that's because at lower lods it was still there so i just needed to rebuild the lods and once i did that then it was saving appropriately so at this point i was confident enough that i could build out a pretty good looking b-plus level fire quadrant so now i jump over to my rehearsal environment and the first thing we're doing is flattening out the landscape because if we're painting with the same stones we're going to need a flat landscape to avoid the pancake effect you've seen in the previous episodes and now i'm dragging in each of the sandstone rock formations because we got to figure out what size and scale each of these are going to be where they're going to be placed in the quadrant am i even going to use all of them because we narrowed down the list of these already but we're still going to narrow it down even more and again i'm starting with the lowest level of abstraction after the landscape then the rock formations and only then will we have the foliage on top of that so now i'm trying to get each of the formations roughly the same size and shape and this is where i'll then have my final cut to determine which of these i'm actually going to use and for the small ones i just dragged them out of the scene because they might supplement the scene but we're going to start with the larger ones and go from there and then i have to place each of these exactly where i want them in the quadrant the first thing is aligning it with the angle of the quadrant for example against the river here and then subsequently i need to make sure that they're evenly spaced out because there's going to be rows of flowers around each and i want to make sure there's still room for paths so even at this basic stage i'm still trying to visualize what that final scene is going to look like now there's still a couple of places on the landscape that are empty and so i figured i could use a smaller rock there the other thing i realized is one of these assets is particularly narrow and if i could combine it with another one of the smaller assets it might make it look a little bit more robust now on to the centerpiece so i had already experimented with this so i knew the concept would work but it's still a matter of finagling with the rotation with the scale with the sizing and just make sure that it all comes together as one piece and as i'm doing this i'm checking to make sure i can't clearly see where one rock ends and another begins then once i got all the pieces in place then i just need to center it last thing i did with the sandstone is make it a little bit darker i experimented with this earlier and i didn't want it to be too orange but definitely a little more orange than it was to go along with a fire motif now on to foliage painting so i started with the red amaryllis flowers which i really loved and as soon as i start painting that initial layer of foliage then i have a sense of okay are my rocks spaced out enough to provide room for paths because ultimately i'm going to have three layers of foliage surrounding each rock and now on to the varia gated variegated crouton cretan and remember so the rock is the interior it's the tallest and then the first layer of foliage the interior foliage is the tallest foliage and then the middle layer is a little shorter and then the last layer is the shortest so once the painting is complete then we're focused on aesthetics so the first thing is are the textures correct so are the brightness settings good and for these yellow bermuda buttercup flowers i really wanted to make them pop and then just playing with the lods making sure that each of the flowers isn't suddenly changing in clear view this is all stuff we've done already in previous episodes you already know how to do this stuff if you've been following along and now on to painting our nordic coastal ground cover our pebbles and for this i kind of went overboard to start and once i painted all that then it was time for the fps test and yeah i knew it wasn't the foliage because i had experimented with that earlier it was definitely all the pebbles i think i have this inflated expectation of nanite because it's just so darn cool but the problem with these meshes is i'm literally overlapping thousands of meshes in the scene in the one area where nanite struggles is where you have multiple meshes all overlapping in the same location so that nanite has to process them all simultaneously but i realized it wouldn't be a net negative to just take out some of these pebbles because it was kind of overwhelming as is in the scene i wanted more of the green to show through some of the garden so instead i reduced it down to just paths having the pebbles but it's often the case that being limited by performance it actually forces me to rethink the scene and it comes out better than it would have otherwise and with each quadrant of the garden i'm going to have actually three doors or three paths leading up to the walls i should say and you haven't seen this yet but we're actually going to complete the walls next episode but with each quadrant of the garden you see it actually borders on three different facets of the walls so each one will have an individual path leading to it and once i drastically cut back on the pebbles then i could test out our fps and sure enough yep that raised it by about five to eight fps so now it was on to painting rocks on the side of our paths just like other quadrants and you know for all these techniques i've already mentioned this but it's just stuff i've seen work elsewhere there's a line in the movie pirates of silicon valley that they quote from either picasso or some other source but it's both steve jobs and bill gates quoting it and the quote is good artist copy great artist steele i won't ascribe to that but genuinely speaking all art is inspired from something else so make use of the inspirations in your life so then we're just testing this out once again making sure the addition of those path rocks doesn't affect our fps and yep it's totally fine so now on to painting our river edge so first we just needed to get our cana lilies up and running and that was the standard stuff so i'm not going to show you that here stuff like making them a little bit darker changing the lod distances and the last thing was using the model editor mode to actually cut off the dead parts of them and once i did that it was just a matter of setting up the foliage setting so that they painted correctly and so that basically got the quadrant up to a b plus we got the basics down now we still had some billboard switch issues with the lilies so i just went back into the meshes and switched off the billboard material but in general the quadrant was at that b-plus level at this point and now it's at the point where it just needs refinements whether it's things that occur to me immediately or things that come to me in future episodes the one thing that came to me immediately was wouldn't it be neat if i raised up each of the rock formations a little bit so that it appeared raised relative to our path and that was a nice effect so the next step is the next level of abstraction and in our case here it's what are the truly meaningful the symbolic elements of our scene and again we're going back to that theme is everything we do in a given scene in a given part of our garden in this case needs to point to a higher purpose now i'm not going to do that in every single quadrant this episode but i want to highlight two specific quadrants and that's our air quadrant here and our water quadrant in the background so last episode i was talking about the water quadrant and our paths were coming together great but the problem is it's looking pretty sparse right feels like it's really missing something and that key thing that it's missing are trees and the same thing goes for our water quadrant because in the center of every quadrant i want a center piece something that's truly meaningful that brings the whole theme of the quadrant together and my idea for the water quadrant was to have an absolutely gigantic tree in the middle of it in the theme of what's called the world tree which is a really common archetypal idea across many cultures many religions but here's the problem our tried and true method of getting free assets to bring into our level it's not going to work in this case there are no tree assets in quicksilver bridge so where do we need to go so the next best place is to go to the epic games launcher and start looking at free content packs by looking through the free content packs compatible with ue5 i narrowed it down to two content packs the open world demo collection and the city park environment collection i downloaded both of them to try them out the open world demo collection gives you the ability to simply add it to an existing project so i added it to my experimentation environment but the city park environment collection has to be installed as its own project the open world demo collection has been a showcase of unreal engine for almost seven years now but it's beginning to show its age the assets are not quite as high resolution as those of newer content packs plus it only has tree assets that you see here in trying out the city park collection which i was less familiar with i knew right away that this was what i was going to use the assets are incredibly high quality but not only that it comes with almost a dozen different trees to choose from along with bushes grass and other high quality assets like a realistic water material the icing on the cake for me was this weeping willow tree these were going to go perfectly with the overall motif of the air quadrant even though the city park collection cannot be added directly to an existing project it's pretty easy to add the assets to whatever project you need just go into the project folders content city park copy everything from materials down to textures and then navigate to the project content folder for which you're going to use these assets create a new folder for city park or really whatever you want to call it and then paste in those files that you copied when you launch your existing project they will then be accessible and the entire thing is only about three gigabytes now it was time to test out these willow trees and decorate the air quadrant and dragging in one of these assets wow the city park collection is truly an amazing content pack and this is exactly why you don't need to be a 3d artist in order to create a game now because there's so much amazing free content available out there that you can create virtually anything simply with the palette of assets you already have at your fingertips and there were only two things i wanted to try to improve about the out of the box settings for the willow tree the first was the color i was hoping to get a much brighter more intense green to match the intensity of the foliage in the air quadrant so i reversed this desaturation setting so that the tree color would be even more saturated and the last thing was how much the trunk of the tree was moving from the wind settings but that was an easy fix for these assets there are surprising number of wind settings but just by playing around with them i was able to figure out which settings control the leaves versus the branches the tree actually has two materials one for the main trunk and one for the leaves and so i could adjust the wind settings of the trunk and branches to be different than the leaves i basically cut the trunk wind settings down to a quarter of what they were and the rest of the wind settings are reduced to half to give them a more realistic look i then went into the mesh itself to adjust the lod distances what i knew was going to be a challenge was the performance of these assets as always but these especially because they were such high quality so if i have 20 trees in the quadrant i wanted to make sure that the lods appropriately reduce the triangles and vertices as much as possible for those trees further away from the camera and then it was time to paint these as of course foliage assets even though i wasn't going to have many of these trees in the scene it's still less performance intensive to paint them as foliage instead of placing them as individual meshes in the scene so i just created a new foliage asset in that same folder adjusted a few of the standard settings and i was ready to paint now in the case of things like trees i recommend choosing the single option from the foliage editor which allows you to place individual foliage meshes in specific locations where you need them i wanted two of these trees right at the entrance of the air quadrant almost like mythological psychopomps air spirits that are guiding the way but then right away i realized that i forgot a setting with the foliage obviously we don't want trees growing in sideways so i needed to undo uncheck the aligned to normal checkbox and now i'm ready to go and at every step of this process placing these i was assessing the balance between two things performance and aesthetics from a performance standpoint the more of these trees i had close by the more intense of the scene but from an aesthetic standpoint it's a bit more complicated because how shady do i want this part of the garden to be and so i figured a good balance of some parts of the path in the sun while others were in the shade was the right way to go and now judgment day the infamous fps test all right all right not great but not terrible and i knew that the air quadrant would be rough specifically because of these trees and perhaps that's good it means that the different quadrants of a construct level will have different base frames per second which means i can test gameplay in those various situations so for example how will a certain spell or effect perform in an environment that's not too performance intensive versus one like the air quadrant here where it's more intensive and to be frank there's no way i'm giving up these trees just look at the shadows dancing on these flowers how cool is that so now on to our water quadrant the problem as it currently stands with the water quadrant is the sheer density of our foliage it's really lowering our frame rate especially in the interior so what if we could solve two problems at once what if we hollow out the interior creating a space for something that could truly add meaning and value to the scene so my idea was to plant the world tree in the middle of a shallow lake after all it is the water quadrant so the first step was to erase all the height data in the middle and then racing foliage a little bit unnerving but i decided to do it in every place that there would still be a thick band of foliage between the paths and the lake thick enough that the lake couldn't be seen from the paths i then lowered the entire area for this i made the brush sensitivity very low holding shift to lower the height of the landscape while sculpting next was flattening making sure that the underwater ground would be an even level across the entire lake and then onto the water water can be handled very simply in ue5 once you have the right material and luckily the city park collection came with a great realistic water material all you need to do is assign the material to a plane and then scale the plane up to whatever size is needed now for our actual rivers this isn't going to work because the flow of the river needs to match the bend at every point of our river but for a lake this is going to work just fine then i just needed to scale up the plane mesh to the right size centering it properly in the lake and then raising it to an appropriate level and now is the big decision the picking of the tree feels like i'm walking in a home depot's garden section i decided on this big elm here but not only that i decided to stack two of them directly on top of each other but rotated roughly 180 degrees to provide the illusion of an even denser tree this is the world tree after all then it was just a matter of proper placement and then scaling i initially thought it should be twice as large but that didn't feel quite right turns out three times was the charm i decided on a feature where there would be a little beach right around the bend of the lake it just seemed to work out that way geographically and then on to the lake edges i thought about using a bush or some other foliage from the city park collection but ultimately decided to just keep it simple same bamboo we used previously for our river edges and now the moment of truth exactly as hoped for fps much improved the air quadrant is now the low fps quadrant no longer the water quadrant can't wait to get into water in earnest we're going to have water niagara effects sound effects actual flowing rivers footsteps with water splashes tis going to be b-e-a-u-tiful after that just needed to clean stuff up fill in some of the gaps between the edges and the rest of the foliage the very last thing i wanted to experiment with was these lily pads i saw resting on the water in the city park collection so first i drag them in and they look pretty neat blending in with the water so next of course need to create foliage and then attempt to foliage paint them the problem was it wasn't working i realized i needed to check both the static mesh and the translucent filter checkboxes to allow them to be painted directly on this water material and the end effect looked pretty amazing what's the difference between art and science or art and engineering for that matter science engineering mathematics it's all quantifiable it's all measurable all parts of the equation can be taken into account for the outcome art is qualitative it's felt to me it's a knowing but it's annoying deeper than simple rationality because you feel it it's like physical attraction towards someone you can't explain it fully and yet it's more real than any fact floating around in your head in ancient terms or if you prefer to think of it in symbolic terms art comes from the heart while science comes from the mind true art deals in the symbolic things that are tangible yet point to something abstract something that can't quite be defined can't be cleanly categorized doesn't fit neatly inside a box there's a pattern that it points to but that pattern can't be defined mathematically true art like the totality of nature is messy and it's a process but it's a process of bringing the physical more into alignment with the symbolic the world of the ideal that world that we know is there but can never quite touch we can only get closer and closer to the ultimately beautiful we can never define it directly so with all of this stuff go with your instincts how it feels and those instincts will grow and they'll get stronger the more you do this the more you practice your craft you

Channel: NumenBrothers

Views: 4,231

Keywords: Unreal Engine 5, UE5, Building a scene, creating art in Unreal Engine 5

Id: Qa0r8P6moJc

Length: 29min 39sec (1779 seconds)

Published: Sat Jun 04 2022