RealityCapture to UE5 - Workflow Tutorial

[Music] hey everyone welcome soj here it's great to see you lovely faces again today we're talking about photogrammetry if you've ever wondered how to talented people like quixel make their mega scans assets that is how for those of you who don't know what this is photogrammetry is the process of taking a series of photos of an object essentially scanning it and using these photo to generate a 3d model with the help of an app such as reality capture this process can create extremely high resolution models and get the texture for free in the process albedo normals displacement maps that is what makes photogrammetry so powerful for artists like us now full disclosure this video is sponsored by capturing reality which is now part of epic games they reached out to me a few months ago for a collaboration and i was super stoked to make content for them because i've been using reality capture in production for years in the past i'm happy to make this video because this is a product i genuinely believe in they don't need to pay me to say that now when come to time to export you're going to get a prompt requiring you to pay for your model with ppi credits this is reality capture the payment model and the cost of your model is entirely dependent on the amount of megapixels used you can find a calculator to estimate how much a given model will cost that being said if you use the discount code william 100 you will get 3 500 credits for free but only for the next week so be quick and use the link below to get your free credits before time goes out we're going to take a look at everything you need to know to hit the ground running with reality capture so let's get started with the basics and it all begins with the camera the best camera is the one you have with you so for my own sake i use the nikon d800 with the nikon c62 right here resolution is important here the higher the resolution of your camera the better it's going to be anywhere between 24 megapixels and 48 megapixels is going to be absolutely perfect you can use your phone but a dedicated camera is going to be a lot better so let's talk about the recommended settings and lenses for photogrammetry starting off with the lens itself you're going to want to use the 24 to 35 millimeter lens full frame equivalent for scanning objects this is going to give you a wider field of view you're going to be able to squeeze in a whole lot more reference points in a single image you can use other focal lengths of course use what you have but you may need to shoot more photos to compensate secondly the shutter speed this is what determines the length of time the shutter is open so i recommend using a shutter speed that is at least double your focal length so for example if you're using a 35 millimeter lens i recommend that you shoot at 1 70th of a second or faster this will make sure you avoid any motion blur or camera shake that will give you a blurry image third you want to use an iso value that is no higher than 800 or so of course this entirely depends on your camera if you're using a high-end full-frame camera like this you can get away with higher isos but for the most part 100 to 800 is usually a safe place to start next up aperture the aperture is the opening of your lens here this is measured in f-stops and you should use an aperture of f8 at the very least even up to f12 this will make sure you get a deep enough depth of field with as much of the frame in sharp focus as possible then we have white balance and the exact value here is not important because that's going to depend on your lighting but what is important is to not have auto white balance turned on because that means that each photo that you take could have a different color temperature and with reality capture using the photo to generate textures for us we want the consistency lastly file format shoot in raw if you can and convert to jpeg after the shoot shooting raw gives you the most amount of latitude and pose to correct the image highlights shadows etc so to recap use the focal length of 24 to 35 millimeters use a shutter speed that's at least double your focal length keep your iso as low as you can light permitting lock that white balance and shoot raw so now that we've covered the basics let's take a look at the photos i shot for this project and learn how to tweak them to get the best possible scan quality now you may be asking yourself how am i supposed to take these pictures where am i supposed to be when i shoot the photos how many photos do i need let's start with the basics imagine i want to take a scan of this box right here what you want to do is you want to take photos from every possible angle all around your subject you want to take photos from underside you want to take photos from on top you want to essentially move in a circle around your subject and get photos from every conceivable angle that is what photogrammetry is about you don't want to leave a single side of your subject not photographed get to those awkward angles do what you can to get all the shots you need and a key takeaway here is you want to shoot and layer them the easiest way to think about how to shoot your subject is to think of it as an onion imagine the outer layers are when you are a little bit further away from your subject you have an establishing shot you take photos where the entire subject is in the frame and then once you've gone all around your subject you can move a little bit closer you remove one layer at a time and eventually you get really close to your subject the core of the onion that is the best way to think about how you should scan your model so let's take a look at what i did to scan this cabin right here to give you a better idea of how to shoot these and where you need to be in relation to your subject and how many photos you need so you'll see here i have all of my raw files ingested on my computer and what i mean by shooting in layers is the following pay attention to my first photograph the entire subject here the entire building is in the frame we want to start off by taking a series of photos of establishing shots we want to make sure that we go all around the building and try to make sure that the entire building is in the frame in all those shots so let's take a look here i've got my first picture i moved a little bit to the left took another picture moved to the left took another picture so on and so forth now you'll notice not the entire building is in the frame here like the top of the bell tower here is not in the frame but that's okay i had a drone with me that was specifically there to take more pictures of the bell tower so i wasn't too concerned about that so moving forward you'll see i'm going around the building bit by bit picture after picture and moving just one or two steps at a time until eventually i get right back to my starting point where i was when i took the very first picture right here and now that i've taken the establishing shot i've gotten the outer layer of the onion shot now we can move a little bit closer so you'll see here and now i've gotten much closer and i'm starting to take pictures of those juicy details on the building itself so you'll see i'm getting really low i'm getting on my knees at some point i was getting down on my stomach to get to really really awkward angles and really this is the entire part of the process you just need to take so many pictures of every conceivable angle of your building now one thing that's extraordinarily important is when you're taking photos going side by side like that it is unbelievably important that you get enough overlap between photos and what i mean by that is let's take a look at this sign here okay as i scroll through the other pictures i've taken this photo moved to the left took another photo moved to the left took another photo notice how this sign is still in the frame there is a lot of overlap between photos you want reference point to be visible in two or three or four photos at the same time because this is going to help reality capture with the alignment part of the process the more reference point that reality capture has between photos the easier it is to align the cleaner your scan is going to be so i've got one two three four five six photos where this reference point is in that same image same thing was done for this sharpening stone over here i've got one two three four five six seven eight nine ten nine or ten images where this sharpening stone is in the frame this is going to make the alignment process and reality capture way easier so having at least a forty to sixty percent overlap between photos is very important here okay so i'm just going to scroll through these photos real quick just to give you an idea of what i did and how i shot the images and how i shot the subject you'll see i'm just getting a better field of view here i was using a 35 millimeter lens and i wish i had a 24 millimeter but i didn't having a wider lens would have helped me a little bit in this situation now one thing i want you to notice here is all of my images are in tact sharp focus there's no motion blur there's no camera shake we got all these crisp sharp details that is what you want if you go through your photos and you notice that one or two of them are blurry delete the blurry ones there's no point in keeping the blurry ones it's only going to confuse reality capture you're not going to get any useful detail out of the blurry photos so delete them get them off your hard drive there's no point in having them that is why you want a fast enough shutter speed to freeze that action even though your hands might be shaking a little bit now for this project i shot 1600 photos for this log cabin here so it's a lot but i honestly think i probably could have shot even more if i wanted an even cleaner scan but for my own purposes this works just fine so now that we've taken a look at how i shot my images let's jump into adobe lightroom and show you how to process these raw files to give us the best possible result in reality capture so here we are in adobe lightroom i've brought in all of my raw images into lightroom and the reason we're doing this is because since we shot in raw we have the ability to tweak and change our image a little bit to help us get not only better scans but better texture details so you may notice right here that my images are a little bit dark especially if you look at the histogram on the top right hand corner here you'll notice here that the kind of gray spike here on my histogram is a little bit too far on the left this means that my image is a little bit underexposed and the reason for that is because this was an extraordinarily dark day this was shot in november in norway which means the days are short and it was overcast meaning there was very little light to work with so i had to compensate a little bit to get a fast enough shutter speed and to use a small enough aperture so i intentionally shot a little bit underexposed knowing that i can lift up the shadows in post and that's what we're going to do here we're going to remove the shadows and the highlights from our images and the reason for that is because reality capture is going to give us textures for free based on these photos you want to remove as much shadow and lighting detail from your photos before generating these textures it's just going to make your life a little bit easier and the way i do that is in lightroom so what we're going to do now is you'll see i have the shadow slider over here i'm going to slide this up it just effectively lifts all those shadows and i'm going to do the same for the highlights bar right here and i'm going to turn down the highlights giving you a much flatter image now you'll see we still have a little bit of shadow detail ultimately that's not the end of the world we can always remove that a little bit later but you know if you want to adjust the exposure a little bit lift it up even more or adjust the contrast make it a little bit flatter you can do that too but in general lifting the shadows and lowering the highlights worked very well for my use case but now in lightroom you can also do a subtle sharpening and a little bit of a noise reduction just to help get a slightly cleaner image but don't go overboard because i've said this before noise is detail detail is a good thing you don't want to remove too much of it one thing you don't want to do something you need to avoid if you use the lightroom process is the lens corrections side here you don't want to enable your profile corrections on your lens that's going to remove the distortion in your images and that could affect alignment in reality capture so just don't do that but now i have 1600 raw files to process this way unfortunately lightroom has what we call batch editing so now i can go back to the library tab right here and select all my raw files like so you want to make sure that the one you did edit is highlighted here and i'm going to click on sync settings and now i can leave it at the default hit synchronize and now you'll see adobe lightroom processed all of my raw files and now each and every one is a little bit flat a little bit washed out we're not trying to create a pretty photo here we're just trying to create some usable data to work with it's okay that it looks very flat and washed out and not very pretty at all so now the last thing we need to do is we're going to go to file i'm going to export and i'm going to export all of these as jpeg and the reason for this is because reality captures themselves have told me to use jpegs instead of tiffs anyway so it's going to save you a lot of space in your hard drive jpegs are going to be more than good enough especially if you tweaked them the way i did so we're going to export a jpeg color space srgb quality mac to 100 hit export and lightroom is going to export all 1600 plus of my raw files as jpegs automatically we're now ready to jump into reality capture and get started on our scan now it is worth mentioning that you don't need to use lightroom any photo editing app will work just fine there are lots of options out there such as photoshop capture one affinity photo the list goes on i just use lightroom because it's what i have and it works very well i would also like to shout out clint jones aka punisher because he too has made a tutorial on photo scanning so if you want to learn additional tips and techniques for getting better scans check out hit video in the link below and now comes the moment to download reality capture and get our images in there so if you don't already have reality capture installed click on the big download now button and get it installed once you have reality capture installed you open it up for the first time this is likely what you're going to see so before we do anything at all let's talk a little bit about navigation and the windows because the layout can change and it's a little bit weird and confusing at first so let's start off with the very basics you'll see here my viewport is split into three different windows i've got the 1ds right here here i've got the 3ds and right here on the right i've got the help window you can change the layout on the top left hand corner here by clicking on either single view or one plus one plus one or any other layout that you want to have personally i like the oneplus one plus one split it up in thirds now as for the terminology the 1ds here is where you're going to find all of your image information your texture information your model information your model data your vert counts all of the settings all of that happens in your 1ds so you're going to be doing a lot of work in the 1ds menu so be sure you always have this visible on the side it is very important here in the middle we have the 3ds it's a 3d viewport you'll see we have a grid and the control for the navigation here are very straightforward it's right click to rotate you left click and hold to pan like this and you use the scroll wheel to zoom in or out now you can change the pivot point of where you rotate around by double clicking anywhere on your grid you'll see a new gizmo showed up here and now when i rotate my camera will rotate around that point if i put my pivot point over here by double clicking you'll see my camera rotates around this point right here so it's pretty useful when you're navigating around a large object and your camera is rotating in a weird angle and lastly our third viewport here is called the help window and the help window is probably the single greatest thing in reality capture reality catcher is a perfect blend of extremely complex and advanced tools for advanced users but you'll see here we have the quick start guide and that is essentially a guide that will hold your hand through the entire process of starting your scan from scratch if you've never done any photogrammetry before the quick start is the way to go we're going to be using that for this tutorial just to show you how easy it is to get into photogrametry so let's do that right now let's hit the big quick start button right here and get started you'll see step one is to add our images what we're going to do is we're going to go up top here click on plus folder a new window will show up and you're going to find the folder where all of your jpeg that we got from lightroom are or wherever your raw files are wherever your tiff files are wherever the images you want to use are located i'm going to hit ok and now once you've loaded your images you're going to see here in the 1ds on the left hand side images 1758 if you click the little plus you'll see each and every one of your images have been loaded now just be careful if you do click on an image sometimes it can change your layout right so you'll see now my quick starts at the bottom one ds on the side we've got 3ds and help and the 2d don't panic it's fine you can go back up top and set it back to one plus one plus one once your images have been loaded we're now ready to click on the continue button in the help window it's going to tell you about the window layouts and the navigation which we already talked about so we're going to click continue one more time and now the next step involves aligning the images unfortunately this sounds a lot more complicated than it really is this is an automated process reality capture is going to do the aligning for you you just need to hit the align button and let reality capture work its magic so you're going to find the align tool up here in the workflow tab and we're gonna click on align images make sure you only do this after you've imported your images and assuming you've taken your photos correctly you've deleted any blurry images and all your photos turned out well hit the align button and you'll be ready to go now just a bit of a warning when you're using as many photos as i have aligning can take a while and i'm talking hours depending on how beefy your computer is so if you're using a low end pc it can take a while just so you know two hours later so once the weight is over and your images have been finally aligned you should be presented with the following you should find a nice point cloud of your scene right here along with positions of every single photo that you took so all of these white dots here these are all essentially icons indicating where a photo was taken now before we continue i just want to take the time to show you exactly what happened here how i took my photos where i was and what exactly you're seeing here all these white dots here are all positions of where i was in relation to the building when i took the photo so you can absolutely see the pattern i took when going around and around and around the building you can see what i mean by shooting in layers i started off way up here went all the way around like this all the way around and then got closer and finally up so close that i was getting some almost macro photos of the building you can see here exactly where the camera was pointing at you'll see i had a ton of close-up photos here same thing took a series of photos like that backed up this is what i mean shooting in layers like an onion starting from the outer end and working your way in a little bit closer now you will notice however i do have some dots up here as well like how the heck did i manage to get all the way up here and that was done with the drone so i had a very small mavic air tiny little drone i didn't use it much for the rest of the building because like i said it was an extremely dark day and there was just not enough light for that little drone to work well the images were very noisy the iso was way too high and it wasn't giving me good results but it was bright enough on the top part of the roof so i took some photos up here around the bell tower just to help the alignment a little bit and get a little bit more detail on the top bell tower because since the shots were all taken at ground level it was very difficult to get any kind of photos on the top of the building here so a drone would very useful in this regard in general if you're shooting building it does help a lot to have a drone but you don't absolutely need it not strictly necessary you can get some pretty decent results from ground level but again the more photos you have the more angle you have of a given subject the better your result will be there's no way around that so if you have a drone make good use of it if you don't it's not the end of the world this should give you a very good idea of how many pictures you need to take and where you should be when you're taking pictures now in the event that you don't see something and if for whatever reason you don't see anything on screen nothing is there take a look in your 1ds and make sure you don't have several components it is possible that you end up with something like this where you only have like you know a tiny little point cloud something irrelevant this can happen sometimes so if you don't see what you expect to see take a look at the components maybe you'll find what you're looking for so here if i zoom out you'll see i have point cloud information of my entire scene we've got a whole bunch of trees basically the whole area around where i took pictures now we don't need all of this information i don't want to generate data based on these trees over here i really just want the building right here nothing more so we're going to be able to tell reality capture what to do with this orient it properly and basically ignore anything outside a determined volume so once you've found your component here with your point cloud data now we can go ahead and click on continue right here and so here in the help panel it's going to explain to you what the scene component is all your camera poses and your point cloud and also once your alignment is done be sure to save so now comes a time to set our ground plane and this is essentially the process of you know aligning our building correctly because you'll see that the building is completely crooked compared to the grid we don't want that we want our geo to be nice and straight and aligned properly on the grid we're going to go to scene 3d and tools and where it says set ground plane click on the little arrow and click on define ground plane now you should get a gizmo that shows up in the middle and we're going to use this gizmo to align our scan to the grid it doesn't need to be absolutely perfect but it does help to get it as close as you can this seems not too bad to me and good enough so now you'll see our building is much straighter along the grid than it was previously once you're happy with the orientation of your point cloud we can click on set ground plane one more time and it should be locked into place scrolling down on the help menu we're going to hit continue one more time and now we're going to define the reconstruction region and the reconstruction region is the bounding box that you see around my building like this what that means is everything that is outside that bounding box is going to be ignored so all these trees that we see in the background over here and over here those will effectively be ignored since they are outside the bounds of this bounding box so in order to set the reconstruction region we're going to go to scene 3d click on the tool tab again and where it says set reconstruction region click on a little arrow and i'm going to click on set region on grid now i can just click like this and you'll see a rectangle shows up around our building click again and lift it in height and now you'll have your new reconstruction region now we can fine tune this don't worry so you'll see we have these colored dots here this is how we can kind of set our reconstruction region the way we want so make sure that the bounding box encompasses the entire volume here and make sure you do it for all sides you want to kind of make sure that the reconstruction region kind of snuggles up as closely as possible to the bounds of your model this should be good enough now once you're happy with the shape and size of your reconstruction region we're going to scroll down and hit continue one more time now it's time to do what is quite possibly the most time consuming part of this entire process and that's computing the model this is the part where reality capture is going to generate the 3d model for you and in order to do that we're going to go back to the workflow tab up top here and we're going to click on calculate the model now there's a little drop down menu here i like setting it to normal quality you can do high quality if you so desire if you're looking for the absolute best possible scan but in general in my experience going to generate a higher resolution model that doesn't necessarily have more detail so if you know your photos are absolutely top-notch you can consider the high quality setting but for the most part normal quality is going to be good enough and take way less time so once you've saved we're going to hit the normal quality button here and let reality capture work its magic many hours later so after the painstakingly long wait of generating your 3d model you finally finished it's here and your model might not look the way you expect especially depending on the size and the polygon count of your model you may notice that it still seems to be a point cloud you know it's not polygons what are we looking at here the reason for this is because reality capture is only able to display 40 million polygons on screen at one so let's click on the model right here and in our 1ds you'll see here that this is where the 1ds comes in real handy this is where we're going to find a whole bunch of information on our model and right here we've got triangle count 264.2 million it's a lot it's a very dense model and down here in the report we can see a whole bunch of other information we can see how long it took to process so it took me two hours and five minutes to process this 3d model it took a while but now how do we see what we did how do we see our final result how do we see the actual polygons themselves now there's two ways we can do this the easiest way is to simplify this to down to a number below 40 million or we can use what we call the clipping box and the clipping box is essentially going to only display the polygon within its own volume a bounding box if you will so in order to access that we're going to go to the view tab up top here and you'll see here on the right hand side we've got clipping box i'm going to click on a little arrow here and i'm going to click on create from construction region and then click on this again and hit edit clipping box now you'll see we have the little color dots again we're going to resize this down like this and we have to make it much smaller so that fewer than 40 million triangles are being displayed at once so because our building is very dense we need to make our clipping box very small so you'll see just this right here is still more than 40 million triangles and so now once you've made your clipping box small enough it is now displaying your actual polygons it's no longer a point cloud now when you zoom in sometimes you may feel that the alignment cameras are kind of in the way these little things here they're kind of annoying so you can hide those by going up here in alignment cameras and clicking on enabled uncheck that and now they will be gone they you didn't delete them you're just hiding them that's all so now looking at our model this turned out pretty good we got quite a bit of detail on the wood here it's not a perfect scan like i said i probably could have taken even more photos but this was a massive building and honestly for me this is good enough so that is what the clipping box does due to the density of our building here using the clipping box is not ideal because i want to see the entirety of my building i don't want to have to load small sections of it so that brings me back to what i was talking about when it comes to simplifying the model essentially decimate it down from 264.2 million triangles down to something like 39 million triangles just underneath that 40 million triangle threshold so i'm going to go to clipping box up here again and i'm going to clear the clipping box and now i'm going to go to the workflow tab up top here and you'll see the simplify tool right here we're going to click on this now you may wonder if nothing happened take a look at your 1ds again and you'll see we have the simplify tool settings show up here whenever you use some tools in reality capture their settings will show up in the 1ds on the left hand side here it's very good to know about so you'll see we have a whole bunch of settings here we have the target triangle count i'm going to set this to 39 million that way it's going to simplify or decimate model right under that 40 million limit make sure to type it set to absolute and when you're ready we're going to hit the simplify button this won't take nearly as much time as computing the 3d model just so you know so once the simplification process has been completed you'll see here that when we simplify reality capture also spits out a normal map for you which is absolutely awesome so depending on how much you simplified it will bake the appropriate normal map for you this is something that you get for free this is one of the many things that makes reality capture so great and now you'll see here in our 1ds we've got two models we've got the first one which is 264 million tries and we've got the new one which is 39 million tries and now we can change the view mode we don't always need to see the building in normal map mode we can go to the view button up top here and go to solid we can choose between the three mode we can choose between the kind of the point cloud thing that we had before we can choose solid which is basically going to be our 3d model preview with the clay material on it and the sweet one is going to be your textured information but since we don't have textures yet the only thing it's going to show is the normal map so zooming in a little bit closer here i'm going to hide the alignment cameras and you'll see we've got our scan and it turned out pretty good like i said before it's not a perfect scan and when you don't have enough photos this is what's going to happen you're going to start getting areas that are kind of a little bit splotchy and the reason it looks like this here is because i just didn't take enough photos of these nooks and crannies like that but that's okay because i know that my project my camera is going to be pretty far away we're not really going to be seeing those awkward angles very much and so this is a result that you can expect if you've taken the photos correctly if you've taken enough photos and if the alignment process turned out well now if i wanted to i could export this model directly into unreal engine because with nanite it can support 39 million polygon meshes but i'm going to simplify it even more because i don't want to have that heavy of a mesh because the 39 million triangle mesh is still going to be very heavy and i think my final result is going to be roughly around 5 million 5 million is still a lot but like i said with nanite in unreal engine 5 this is not a problem so the nice thing about generating simplified models that it's a non-destructive process you'll see here i've got my 264 million triangle mesh the 39 million one and my new five million triangle one right here you can visualize them by clicking on the little eyeball next to them right here and it's going to show you the previous model now the last thing we need to do before we bring this into unreal engine is to generate the textures when we simplified our model it generated the uvs for us so let's take a look at our 1ds here you'll see i've got 16 4k textures so it's got 16 muted for an asset this size that is totally normal if you're working with a smaller model you're not going to be getting 16 newtons or 16 texture sets it's going to be a much more reasonable number but because this building with 10 meters by 6 meters it's a very large asset it's going to need a lot of utens or texture sets so how do we generate the texture from this taking a look at our help menu here we have two options we can use the colored version which is essentially just vertex colors and we can use the textured version as well so we're going to want to use a textured version so in order to do that we're going to go to workflow on the top here and we're going to click on the texture button it's really as simple as that make sure you have the right model selected hit that texture button and be prepared to wait a while one more time 10 coffees later now after the long wait we finally have our model textured and ready to go so if we zoom in here you'll see the texture turned out really sharp it turned out really well and if you found that your skin didn't look super clean the texture is often going to make up for that so looking here i mean this thing just turned out so much better than i expected and i'm extraordinarily happy with the way this turned out like just the way that these logs these worn out logs are turned out it's just so sharp so clean super awesome like these rocks turned out spectacularly well this is something that would be extraordinarily difficult to make from scratch by hand so i'm very happy with the way that the model and the texturing turned out so again looking at our 1ds here you'll see i've got 16 udims texture sets at 4k you can absolutely change that if you so desire if you want to fine tune the resolution of your textures the amount of texture sets that gets created by reality capture you can absolutely do that by going to the mesh model tab up top going to unwrap settings and here you can choose the style so i'd show the maximum texture count where i said i don't want to have more than 20 udems or 20 texture sets you can set that maximum there or you can set it by fixed textual size or adaptive textile size so you can choose absolute the resolution you want your texture to be at as well you have full control if you want a single 4k texture you want to have a 2k texture you have full control of it right there now once you're satisfied with the way that your model looks you're satisfied with the textures now comes the time to export this directly into unreal engine 5. we don't need to do any kind of retopology because even with 5 million triangles nanite in ue5 is going to handle this like a champ so in order to export our model we're going to go to the workflow tab up top here and we're going to click on the export button right here this will bring up a new prompt and we're going to export at fbx and i'm going to call this rc to ue5 001 now when you get your export settings here there's a few things that we need to change here so first of which is export vertex normals set that to yes you don't want to save your mesh by parts we only want to have one big model texture settings we want to export the textures the tile type we'll be exporting this as udems if you're not using texture tiles you don't need to worry about this for the normal layer i'm going to export this as jpegs and 24 bit bgr so this is a little bit unconventional but 24-bit bgr basically means an 8-bit rgb image but 8-bit per channel so 8 times 3 is 24. same thing for the color layer here and the last thing we want to do is in the export transformation settings you want to set the transformation preset to maya arnold unreal right here very important once you've done that we're going to hit ok once that export is done let's jump into ue5 and get our model in the engine so with unreal engine 5 open now is the time to import our model from reality capture i'm bringing in that 5 million triangle model straight into the engine it's not going to be a problem as you're going to see real soon so what we're going to do is we're going to right click in our content browser click on import navigate to wherever you exported your file to and import that fbx file when you get your fbx import dialog box there's a few things that we need to make sure we have checked on so the first of which is build nanite you want to make sure that this is turned on here we're going to set the normal import method to import normals and lastly we don't need to create material we're going to create our own because i'm going to be using udems here so i'm going to make my own material i don't want the fbx to make them for me with that done we're going to hit the import button and now we can simply drag and drop our new asset into the scene right here and just like that we've got our model imported from reality capture into unreal now comes the time to make the material for this now by clicking on the model you'll notice that there's a lot of material ids and basically this is because realitycatcher is generating a material id for each texture set but because i'm going to be using udems here i don't need to have a material id for each texture set so we're going to be going to the model editing tools up top here and i'm going to remove all of these material ids except for one and with this new window open up top here we're going to scroll down until we get to matte edge right here this can be a little bit slow sometimes don't worry so with mat ed selected we're going to scroll down on the left-hand side here all the way to the bottom we're going to select all this is going to select all the faces on our model we're going to scroll up and we're going to delete the material ids now what we can do is we're going to create one new one like that now we can assign a material any material it doesn't matter for now i'm just going to give it this one for now and i'm going to scroll down and assign selected material it's going to take a little while hang in there once that is done we're going to hit the accept button right here and again you may have to wait a little bit depending on the density of your 3d model now once that's done now we can go back to the select mode up top here and obviously we don't want this material we're going to create our own material now but now when we click on our model right here you're going to see we only have one material id and that is all we need so i'm going to create a new material by right clicking here and creating a material like that and i'm going to call this rc to ue5 matte now we need to import our textures now i've made a dedicated tutorial on how to use udims and how to get them set up correctly in your scene right here so if you don't know how to use udems i recommend you go watch that video so you can follow along next we'll import our textures right here and you'll see i have a whole bunch of textures here right i've got 1001 through 1016 for both my normal map and my albedo map so we're only going to import a thousand one for the albedo in a thousand and one for your normal map hit open now if your project settings are set up correctly with virtual texturing enabled you'll notice here in our textures when we mouse over we've got 1001 through 1016. all of our textures have been imported correctly but we only need to have these two texture files brought in we don't need to bring all 32 of them so now opening up our material here i'm going to bring in my albedo and my normal map and i'm going to plug this in like that once your textures are connected we're going to hit the save button right here and we're going to apply this new material to our building and just like that the udems take care of all the texture sets we don't need to worry about having one material id per texture set having udems is a much more elegant solution to work with that is why i recommend using them for large assets where you need a lot of detail here we have our building fully textured in unreal engine 5. now in order to get the roughness texture you can probably tweak the albedo map a little bit desaturate it invert it boot that contrast basically generate a roughness map from your albedo maps themselves and that's it and once you made your material throw your scanned and textured model into a lovely environment like i've done here and the world is your oyster the fact that we have nanite making this entire workflow possible not having to do a re-topo re-uving re-texturing everything baking all those textures again so this is the workflow i used to create the shot that you saw in the introduction of the video i got up close and personal with this 300 year old cabin i made a close-up version of the sign right here i scanned it brought it into ue5 and rendered it the possibilities here are endless so i encourage you to try photogrammetry for yourself because now that we have nanite in ue5 with virtually unlimited poly accounts the reality capture to ua5 workflow is a game changer for artists and i cannot wait to see what people start creating with this and that's it folks if you found this video helpful do consider subscribing and hitting the like button down below thanks so much for watching and as always happy rendering

Channel: William Faucher

Views: 9,404

Keywords: UE4, Unreal Engine 4, Unreal Engine, Cinematics, 4.26, UE4 4.26, UE5, Realtime, realtime rendering, rendering, CGI, 3D, 3D Artist, reality capture, realitycapture, capturing reality, capturingreality, photogrammetry, 3d scanning

Id: WrCOhes1Zgg

Length: 42min 4sec (2524 seconds)

Published: Sat Dec 18 2021