NAB 2016 Rewind - Athanasios Pozantzis: Re-topology, Modeling & the Perfect Planet Earth

Photo of Cineversity

Instructor Cineversity

Share this video
  • Duration: 97:46
  • Views: 11937
  • Made with Release: 17
  • Works with Release: 17 and greater

Thanassis Pozantsis (aka noseman) demonstrates a plethora of Cinema 4D tips and techniques.

Thanassis Pozantsis (aka noseman) demonstrates a plethora of Cinema 4D tips and techniques in the context of two projects: an advertisement for Ford and a visualization of Earth. In the project for Ford, Thanassis projected final car illustrations onto purchased models, using Cinema 4D’s modeling and sculpting tools to prepare the models for camera mapping. Thanassis also shows how to retopologize a model with the help of the Polygon Pen and Shrink Wrap Deformer, and how to model tire rims using MoGraph, Symmetry and Cinema 4D’s polygon modeling tools. Learn how to create a simplified car suspension rig using Xpresso, and how to layer freely-available earth images from NASA using Cinema 4D’s material system.

00:20Ford Advertisement
05:21Isolating the Physical Set (Camera Mapping)
11:20Preparing Purchased Models
22:27Phong Break Selection
27:04Mesh Checking
32:42Reprojecting onto Model
42:27Tweaking Model with Sculpting Tools
48:03Retopology using Polygon Pen
54:01Shrink Wrap Deformer in Retopology
56:35Modeling Tire Rims
68:00Simplified Car Suspension Rig
83:35Masking Materials based on Light using Lumas Shader
89:16Create Atmosphere using Fresnel

A Download file for Thanassis' Earth is also available in this tutorial. Simply press the "File" button under the Information Tab.

Recorded Live from NAB 2016 in Las Vegas.



- [Athanasios] So what we're going to see today are a couple of projects. I've got a double whammy, two presentations back to back. And for some reason Maxon managed to convince me that that's in my best interest. I just bought into it. So you're going to bear with me for a bit. We are going to see two breakdowns of two different projects. The first one is an advert I did through a very talented studio called Tantrum. I think the webpage is, but it's guys in Toronto where I live and they're right next to Mr. X, for anyone that knows one of the biggest studios on this planet, does a lot of stuff. Anyway, I was lucky enough to work with a group at Tantrum for a very small preroll for Ford. And let me show it to you, it's very simple. So go to Noseman folder. And let's go to full preroll in AB and let's so the deliverables. And let's open them. They were already open and I'm going to maximize and loop. Oh, this is the other Quicktime, okay. There you go. It's as simple as this. And on it's own, it's not extremely exciting. But I'm going to tell you what I had to work with and you'll understand exactly why it could be interesting. I did this for three cars. And the third one somewhere around here. There you go. So you can see wheels turning and spinning and suspension and all that. What I was given where three Photoshop images in CMYK. So let me check them out, supplied images. My eyesight is failing me, anything closer than three feet is terribly blurry. People say it's my eyes. I think my hands are too short but that's another story. So here are the vehicles. I'm going to open these up in Photoshop. There you go. So this is a CMYK image of each and every one of these. And basically I had to work with this because the client wanted this to be the final visual, flipped of course, they wanted this to be the final resting position. Now, what happens in advertising stays in advertising but that's another story. They manipulate the shape of a car in two dimensions to get it to look good. That doesn't mean that it's perspectively correct. And there's never information about it. For example, they can copy and paste the wheel from the front to the back or the back to the front. And of course, when they just scale it down, it's going to have all sorts of problems. So I had to work around these things. So I had to isolate aspects of the car and basically the process is to create a model of the car where I can re-project this image. Then I need to correct the actual mesh to accommodate for the particular image, then model the rims, model the tires and model, sort of, the rearview mirrors, the mirrors. Then do everything else, like the animations and the rigging and so forth. And lastly, I had to use the Take system in order to separate it so I can composite it and post. The final project looks something like this. Noseman, there you go. Ford preroll and let's go to Ford Mondeo, double click on this. Now there's one thing I'm going to turn off. I think this goes fast enough, just need to lower the textures because they were quite high resolution. So rewind. Now, funny enough this is actually what I rendered as the final sequence. And because the actual lateral motion of the car I did in After Effects, there's no reason I should do it here because the timing can change and so forth. As much as you can do in post, when it's simple stuff like this, do it in post. I mean, there's no reason not to. And what I did, I rendered the background separately, the shadow separately and other stuff, and the car separately and then I just composed it in AfterEffects and added motion blur which hides all the issues. So let's take a look at the car for a second. You can see that there are all sorts of tears and so forth but nobody cars because that's part of what's not visible. From the particular point of view, the car looks fantastic. I'm doing my view now until I get back to the original, there you go. So from this point of view, you see no problem whatsoever. So let's begin with the breakdown of this because it contains a little bit of everything, a bit of texturing, projection, a lot of modeling. All right, number one, the first thing I had to do was isolate the set. I'm going to begin with the simple stuff. This was a photo of the set because they actually shot a proper commercial and they were on a set and they wanted me to replicate this but with a continuous background. So all I could take from this is this little backdrop and it's very simple to actually isolate that in Cinema 4D. So what I'm going to do is create a material, go to the "color channel" and add that... Where are we? We're here and we got, not deliverables but the supplied images and get the set. Now, one thing I do is I just go and make my camera exactly the same resolution as the image when I'm doing simple stuff like this. So that's 6016 x 4016. Fantastic. And I'm going to create a background object and drop this on the background object. Fantastic. I'm going to get rid of everything I don't need, things like grids and all sorts of things. Now, mind you maybe I need the grid temporarily. A big part of what we do for these kind of projects is eyeballing. There are tools in Cinema 4D to calibrate the camera based on all sorts of rectangular objects and perspectives and stuff like that. But in this case, you will see how fast eyeballing it is and you can't bother with the tools sometimes. And that's one of the rules, if you can actually quickly do it. And it's acceptable, just go and do it, don't try to be too scientific about it. So, we can see...there you go. I mean, how tall is this? I don't know, two times a human's height which means the horizon will be right there in the middle. Even if I'm wrong, it doesn't make a difference. Number two, let's make a plane. Sorry, I'm testing the thing. I'm trying to irritate Rick on the back. So, +Z and what I'm going to do here is just move it left or right and drag this on this with Command pressed so I can make a copy. So it's still frontal projection. Now you can see it's a bit blurry. In order to clean that blurriness, we go to the material itself and tell the editor that we need something a bit higher resolution and you can see it's cleaned up now. So what I'm going to do in this case, I can even turn off the background, I'm going to eyeball this so I get pretty close perspective. You don't have to be too accurate. And if you have this on, you may be able to see, there you go. There you go, that doesn't look that bad. And I can always make it a bit, there you go. I cropped it. Now, if I want to fine tune my plane, all I have to do is go to the plane, add... I'm not suppose to touch this but there you go, one and one. So I got rid of all the subdivisions and we're going to make it editable. And very simply select these points and bring them up, Eyeballing 101, just move things around, there you go. Now, if I get rid of my background, I have this isolated thing. Now watch this, this is the best way to break it. I'm going to navigate my camera and I broke it. Did you know that Cinema 4D has two undo systems? One for your actions and one for your view? So you can undo your camera moves if you want to. So, the question is how do I move my camera without affecting the actual projection of the image? What we have here is a frontal projection, a frontal projection of an image is we are projecting from the camera we are viewing the world. And what happens is that keystoning and perspective counteract each other so we get a proper image. There is another way to do it. I'm going to click on this button and create a camera for my scene. Now, the camera I created is actually in the exact same position as the camera I'm viewing the world through. And what you can do is change your materials projection to camera mapping. And of course it's going to be all destroyed. Then you link the camera to the projection and calculate this. So now, what we are doing here, watch this. What? So someone has turned off the cameras here, gotcha. So we're using the camera as the projector of the material. And this allows us to control the material using the camera, okay? And if we want to move the object now, I can't do it directly because it's going to lose that projection but if you make the camera a child of the object, you can actually do whatever you want and the texture stays put. Now, a lot of people can convert this to a UV map but sometimes, just sometimes, you may end up having some little issues. And because I do not want to have any issues, I leave it as is. One more camera in my scene makes no difference, I can turn off the visibility and I'm good with that. So, this is a simple way of creating a background from an image by eyeballing, the method of eyeballing. It's served humans for about 2 million years. Is that a deer? No calculations, eyeballing. And we're here. So, let's continue with a bit more complex stuff now. I have to refer to my, I have a PowerPoint presentation but I actually have to write it and I don't want to waste your time. So I'd rather go with my notes. So let's go and see what we have here. They were kind enough from the company I worked for, they're from Tantrum, to ask me if I would like some 3D models. And I was like, "Yeah, if you can find 3D models of the car." And they went on Turbosquid and there's this fantastic artist called Squire. I think he's Russian or something like that, maybe he's from the Eastern part of Europe. I don't want to insult anyone but his models, which are fantastic, the best cars you've ever seen outside the actual CAD data, the data that the factories provide, is made from photos. So I knew that I was going to run into some inconsistencies. So I knew that I was going to have to play around with the model. And a big part of this presentation is actually how we're going to fix that. So we went and we found the Turbosquid, where is the Turbosquid? There you go, Turbosquid. And let's go and grab the Escape since we're working on it. And there were 3D studio models, 3DS. And as you all know, Cinema 4D opens around, I don't know, 15,000 different formats, no, joking. All these, we can save and open all these formats. So this is the car which is all nice and dandy. Now the problem is that there are a lot of issues with the details. They just wanted that photo. They didn't want anything else, anything different. So, I just need the shell of this object, of this model, just the shell. So, number one and bring out your notebooks because tip time. You want to get rid of all the materials in a scene. You open this eye, you open the tags, you click on one of these materials and you double click on this and go edit, delete. Okay, that's simple enough. And then you go "function," "remove un-used materials." So no more materials in our scene. Close the eye. The next thing is I want to get rid of certain elements I do not need, antennas and roof racks and all that. So let me turn off my grid now and let's go and start selecting things, Alt+H to get my car in the center. So, very quickly I'm going to show you another little trick. If you go to the preferences which are right here, preferences. I think it's in the interface, these two entries, "insert new object at, " and "paste new object at." And it has a list: top, previous, next and so forth. This is very handy when you want to do things like copy-paste, where the object is going to be pasted in relation to a selected object in your object manager. So, this is how we're going to do this. I'm going to set this to "child." So that, for example, if I have a null and I select an object and I cut it and I select this and paste it, it's going to paste it as a child, which means that if this is hidden and I select all these, or random objects and cut them and paste them, I can paste them in a hidden null. So very quickly, I can hide geometry I don't need. So, what I'm going to do here is "expand object group," and delete the null. How is that going to help me? With my selection tool and "only select visible elements" to on, I'm going to select surfaces which I need to keep. I want to keep them. So that's going to be the front of the car, a door, kind of the roof here and something like this, this, this, just roughly, good. Go around, I want this, good. So, I Command+X to cut them and I'm going to make a null and Command+V. So now I'm going to start isolating all the parts I want to keep, get rid of everything else and then start playing with the things I've got because this is actually around...the whole thing is 140 elements. So it's quite a few things we don't need here. So let's go select this and this, good. Do I need this? Yes, I do. I need this, maybe. And this, cut, paste, close it. Let's go pick up this. It's all about speed. This is the kind of stuff that you don't really need to think to do. I generally don't think, but for people that do think, if you want to just do things without having to think about them, this is the ideal time. You can listen to music, you can discuss with other people, do a presentation while you're looking around the car and pretend you're looking for something although you can't do that because your brain is pretty much fried. Cut, paste. Half of what I say is not true but that's not my problem. Okay, it seems that I've got all the large surfaces or outer surfaces of the car. Certain parts which are going to leave voids and holes, it's part of the process I'm going to show you anyway. So don't really stress about it. Excellent. So I'm going to select this, go all the way down, Alt+G it into a group, hide it and unhide the other thing. So now I have the outer shell of the car which unless... Oh, we just need the lights in the back, go quickly grab this and grab this. I think it's about symmetrical, cut, go here, paste, close, hide that. There you go. So we have the outer shell of the car, not bad. Nothing else. Good. Just check out the geometry, it's a bit triangulated and all that. But I don't scare easy. There's an inner shell here which I'm going to get rid of, excellent. There are a couple of other things we need to get rid of, excellent. The reason I'm not going to a ready-made model, I've got this ready, is that I really want you to see that the actual process of doing something like this which people may think it takes hours and hours, it's not. It's minutes. Now, what you tell your client is another story, okay because this is tough. I'm sweating, maybe it's the hoodie but that's again another story. Okay, I think I'm in a good state. So what I'm going to do, get rid of this, select everything and right click and do a "connect objects" and delete. So we have a unique mesh now that contains everything. Mind you, if we want to select a group of polygons which are connected, so if I didn't want the front, all you have to do is go to your move tool in your polygon mode and you Shift+double-click, there you go. If you Shift+double-click actually, it's the same as selecting one polygon and saying "select connected." So it's an easy way of going and saying, "Oh, I Shift+double-click and Shift+double-click and adding all sorts of parts of the car." This tip alone is worth your time. I think we can go and have some beer now and be happy. Okay, you're not laughing much at my jokes. It's either that you do not have a sense of humor or I don't have one. It goes both ways. Okay, so the challenge now is to find a way to simplify this model so that I can manipulate it in the easiest and most fluid possible way. Now watch where the problem is. The way I'm planning on manipulating the model is by using the sculpting tools. The sculpting tools can be used in...outside the context of sculpting. I'm not here. So if you right click somewhere here, I'm trying to find the right click over here, you can say, where icon size large so I have a nice big menu. Excellent. So, what I can do is get my pull brush and you have to middle click and drag upwards to change the pressure, middle click and drag. You have to drag, though. Middle click and drag sideways to change the size, or go and change the numbers. If I want to manipulate this geometry because the parts are unconnected, I'm going to end up with issues like those. So I need to create a unified shell of this car. So when I start sculpting, all the polygons are going to move because I need this to have no holes. We can do a projection, I don't want to see what's going on on the other side. That's why I need to make it one mesh. I'm not going to do the whole process because this took more than five minutes, more than five minutes, okay. So the number one thing I want to do is clean this up a bit. If you select all your polygons and you right click, I don't have right click. Okay, I know what's wrong. I need another tool. So what I'm going to do is select all these. I had the pull tool selected. I'm telling you, I'm tired, I can right click but I had the pull tool selected. I should have another tool selected. So sorry for that, right click and go "untriangulate." Click on this little gear button which will give you a menu, a little pop up window which will get rid of "evaluate angle," press "okay." And Cinema 4D does a great job of untriangulating surfaces. Mind you, the reason this is so clean is that it was originally quads, okay? If you take an arbitrary, triangulated model, it won't be that good. But as far as it goes for these kinds of things for example, you're going to see some weird triangulation here but the model now is nice and clean. Good. So how can I go and patch this up? There's some very, very simple techniques we can use. One of them is in any component mode, if you right click, you'll find something like "close polygon hole." What this does is if you hover over an open part of your model where it finds boundary edges, it will try to create a polygon that closes that hole. But here, we don't actually have holes, we have rows. If you go inside, you will see that we have rows, there you go. There's no closed surface. The closed surface that's being selected now is the inner part of the light. So I need to do something to get rid of a lot of this detail which I don't need, namely all the little bevels that are folding in. I do not want that. That is actually giving me a hard time. So, then we're going to go to the least known tool of Cinema 4D. Anyone that has used the "Phong break selection," raise their hand. Yeah, smarty-pants. So the "Phong break selection" is one of the most powerful selection tools you will ever find in Cinema 4D. And basically what it does, it selects continuous polygons that are within a threshold of an angle. Currently, when you select it by default, it uses the Phong tag. But if you deselect the Phong tag and change the value, it will select groups of polygons that are within a certain angle. And many times, when something's designed, for example with bevels and stuff like that, you know that those curves are going to have a sharper angle, whereas the rest of the surface is going to have much more smooth angle. That way, you can go and start selecting parts of this model rapidly, okay? And charge your client by the minute. There you go. Oh, but I see people that may be clients in here. So what I'm going to do, and the other thing about the specific tool, the Phong selection tag, it actually has a great visual cue to show you which area complies to the specific rule. So, already you don't need to go and search things. I feel that that angle kind of works with me, work with me. So, let's go and do that. I'm going to just click and then Shift+Click, Shift+Click, Shift+Click, Shift+Click. And believe me when I tell you that all the inner surfaces are being excluded. So I'm just Shift+Clicking and even if I make a huge mistake I do not care. I've been paid the invoice anyways. Now, honestly, I mean is this easy or is it easy? And I'm not going to do everything because I've got the model somewhere and time is precious. Good. And this. Excellent. And all we have to do is, I'm not even going to do the undercarriage in this case but you understand how this whole thing works. I should have had the undercarriage as a separate object but I still have the model, I can open it and copy it and paste it. I'm going to invert my selection and I'm going to delete. Of course we have some points left around, select all, right click and optimize. And we have a car that doesn't have the bevels anymore, okay? That was a good four hours work. No, it wasn't. So now we're going to use another tool because I need to go and do the following thing. Let me give you an example here. Rule number one of presentations, when you have a screen in front of you, you don't have to look there to see what you're doing. You can get carried away sometimes. So what I'm going to do here, you can see that this is an open gap, it goes all the way down and it goes all the way to the end of the car. I want to create a kind of a hole, just an island. So I'm going to use a tool called the bridge. In edge mode, I'm just going to go and do this and this actually creates the edge of these two polygons. I've closed on side. I'm going to go here and close this side. And actually, I'm going to sit around here for a second and kind of create a few bridges so I can close off these surfaces. Now, accuracy is not what we are trying to achieve here because this is going to be sculpted and it's going to be further manipulated. Now that I've created those islands, going to do this here as well and here as well. It's good if you go with equal size polygons if you have a mild case of OCD like myself. My wife doesn't think so, though. She thinks I'm messy. It's all lies, it's all lies. Okay, I've closed these holes and what I want to do now is find a way to see where are these holes? Right? Where are these holes? Aren't you wondering where these holes are? How do we find them? Well, we go to the "mode" and we go to the modeling tool. Now, funny enough, these reside in a place where it's difficult to find. It's difficult to find, product manager of Maxon. So, there's a wonderful, wonderful, if you have Cinema 4D just go to the "Mesh Checking" and turn the button on. Then get drunk and watch the show. Anyway, what we're looking for, the mesh check is a fantastic tool that allows modelers to identify certain attributes of the mesh that may deem problematic what you're trying to do, 3D printing requires water-tight measures and all that. You can't have reverse polygons and although everything works fine inside Cinema 4D, when you want to use that in other software that are not so good at handling all these cases, you can go and isolate all those cases. But, we're going to use this in a totally different way. I want to find the holes. So I'm going to turn off everything except "boundary edges." So now, this gives me a green color of everything which is a boundary edge. So my eye now can concentrate on these areas. Had I done a good job of closing all these up, I would just see the islands and then everything around but, nevertheless, if I go to my "close polygon hole" now, I know where to hover my mouse in order to find these areas. Now, when you are doing "close polygon hole," what it's created is a huge polygon. And as you can see, it closed that hole, closed the hole. Now, someone's turned on the N-gon lines. And you can go here and just wait until it flashes and you click, wait until it flashes, and you click. See, I didn't close the hole so it's selecting the whole thing. But once you've created those little hole islands, you just go in and click and it closes the holes, click and it closes the holes and you go on doing that until you patch all your model. That's another six hours work. So, when you get to the point where your model has all the holes patched, make sure that you triangulate your N-gons. So you select these holes, you right-click and you say "triangulate" or remove N-gons, okay. Or sometimes, you may even need to go to your knife tool, cut through it a couple of times so that your triangles are a bit more symmetrical. So now I can right click and remove N-gons and this looks slightly better. Let me open one of those models where I actually spent many hours fixing. Open, and let's go to “Processed models for NAB.” So boundary edges, this is the model and what I did after patching those holes just because I forget to tell you, I actually went to my top view because we know that most cars, at least the ones made in the Western hemisphere are symmetrical. So what we need to do is cut down our car right in the middle. So you find those points which belong to that part of the car and of course I forgot to turn this off, so let me do this one more time. And cut this and go all the way down. I'm not going to show you exactly how it's done. And we have a half a car. Once my car is water-tight, and let's go and check our "mode," "modeling," turn this on, turn everything off except for "boundary edges," this is the state where I got it to just to show you that once you've done all your islands, you're going to have a couple of holes left which you can't really see, just turn on that tool and make sure you find them nice, green color. And sometimes it doesn't look like a hole but it is. Nobody really cares. There you go. It's so easy to isolate them especially if you don't have the polygon selected like I do. There you go. So pretty much we've patched the whole car in 12 hours and 30 minutes. The next thing we can do is put a symmetry object and just do that and that's actually the wrong model. And we have a whole car. If for any reason these points are not right in the center, that's not a terrible issue. We're just going to go and select them and move them to the 0,0 point. I'm going to select them all and make sure that the 0,0 point and that's that, no problem there. I'm not going to do it now because I've got a ready made car. Let's talk about the re-projection now. So modeling, are you happy with modeling? Good, applause, you love it. Come on, don't be shy. I'm a vain person, you know. Thank you, that was depressing to say the least. Okay, so let's go to our original model. We're going to go back to the 3DS and all the hard work we put in here for hours, I'm going to delete it. Now, I didn't do everything in the right order but that's not a problem. Luckily, Cinema 4D allows us to do certain things extremely fast. So, we won't have an issue with that. So what I want to do is select my whole car, whatever, I deleted all the unnecessary things. I've made it, I've connected it, excellent. And then I go to my favorite tool, the "Axis center, " which actually takes a model and centers the axis based on the geometry. Then I'm going to center my car. I have two cars here. How did that happen? No problem. Select everything, deselect that. Good. Did anyone realize what happened? Why did I make two cars? Interesting. So I have this car, how does re-projection work? The answer is very similar to the plane. Only we need to add a few more hours of work. So what I'm going to do here is first of all, go to Photoshop, find our vehicle. I think it's this one. There you go, not bad. Double-click. And what I'm going to do is make sure that I have a background of a certain color. It's good to have a background because when I see the edges of my car, I want to have a very good visual cue of if something is part of the background, if it's a hole or not. If it's just white, if it's a mask I won't be able to see anything, or if it's black or something. So you put a very, very bright color here. The other thing you need to do is make this RGB and don't merge, I don't need to merge. I'm going to make a very simplified version of this. I don't even need the shadows. Everything else is fine, just do that. And I'm going to save this. I can even crop it, depending on if you're going to use other elements of it. It's better to have it cropped because everything else is just visual noise, we don't need it. Now, one thing that you need to remember is that at some point, I'm going to retouch and get rid of the mirrors because the're going to see the problem now. They're going to look like elephant ears when it's reprojected. So let's go to the desktop and let's create a new folder, "Noseman rules, " not rollers, rules. And save the other car in the text folder. Good, okay, whatever. Is it saved? Make sure it's saved. Then I'm going to go here and save this as desktop, go to by "date modified," "Noseman rules," and just save it in there. So, then again, I'm going to create a new material, put it in the color channel, in the end you can put it in the luminance channel or something like that. I'm putting it in the color channel because it's going to help me compare the car details with the image details. So I can see where exactly everything is. If it's in the luminance channel, I won't be able to see the actual model undulations on that image. You're going to see the difference in a second. Let's go here. Click on this button. Go to the desktop, "Noseman rules," "texture," load this in. Again, let me do my little thing here, my OCD thing, 4019 x 2482. Next time I got like this, can someone read the numbers for me? There you go. Interesting to say the least. Here's the car, okay. So what I'm going to do is put a frontal projection, just like before. And make sure that number one, my editor has a higher resolution so I have a bit of a better preview. You can see nice and sharp. And number two, don't forget to flip the image, okay? It's imperative that you flip your image if the image needs to be flipped. So, "flip canvas horizontally," save this, wait for the blue line to disappear, go here, double click, click on your color, click on this, reload image, flip car. See, I didn't have to do it. Look at the color undulations here, we can actually see the wipers, we can see the mirrors. If this was in the luminance channel, we wouldn't be able to see those because we'd have can't see them because luminance channel means we don't have any color undulation. The color doesn't change. So that's why we do it in the color channel. Okay, Eyeballing 101 when it comes to complex objects. And that is, number one, find a common point. So you can see the bottom of the light and that's the image at the bottom of the light. And you're like, "Okay, I got it." If I rotate this, it will fit. It will fit in your dreams of course. So, we need to make it bigger and you're going to try and do this and it just won't work. And the reason is the camera lens. And you call your client and you say, "What's the camera lens?" And they say, "What?" And you make a camera, you activate. I'm just joking. And how can you find out what the focal length is? And there's a rule of thumb, the wider the lens, the more perspective you have, okay? So the smaller the number, those mm's, which means that anything that's closer to you is going to look much bigger than anything that's further from you. And using that rule, we're going to eyeball our focal length. And the people that write algorithms to do that can take a holiday. So let's see what is bigger. Is my model bigger? Or is my photo bigger? So what you have to do is just bring your car to have a certain size here and you can see that obviously the model, I mean this wheel is a wheel of a truck. And although the back seems to be in the right size, this wheel is huge. That means we need a more zoom lens. And you go here and you say, "Oh, how much can it be? Sixty?" And you do this and you scale, and you move the common point and you bring the back here and try and find it. It's still too big. Anyway, and you end up with the number 100 after at least six hours of hard work. And what you will see with 100 is that funny enough... Let's bring it here. Look at that wheel. I'm looking at the wheel arches. I'm close enough, okay. That's what I want to do now, I want to be close enough. Forget about the back wheel and all that. This is kind of getting there, kind of getting there. And believe me, because this image has been probably skewed in Photoshop, maybe not but I can't guarantee it. You cannot spend half a day hoping that no one touched it in Photoshop and there's no way for you to measure it. You need to go with hope. So you say, "I'm happy with this. Let's lock it down." Select this, "camera mapping." Actually, that's cool. Send this to the client and be like, "I was in a weird creative mood." And just turn this off and now a car. And here's Bugs Bunny's ears. Now what happens with the mirrors is that because they're protruding from the car, that means that we go around because of the projection angle, you're going to see them on the car. So, all you have to do is just go to Photoshop and Photoshop them out. I mean, the cars I used, the final ones, look like this, Ford Preroll and we go to Ford, let's go the Escape, text and this is the final image I projected on the final project, retouched a bit. I'm very good at Photoshop as well so I can just do these in my sleep. Mask out these, get rid of the bottom parts which I added as models, put some mask there, got rid of the top. This car had a rack, so it was the only car that I actually had to add something on top, it took me at least four hours. And then I used the model, the mirrors from the actual 3DS model, unified the model and just reprojected a second image. This was the image I used to reproject on the mirrors, that's why I haven't touched anything else. So, project one image on the mirrors, one image on the car and you're good to go. But what you see here is the basic idea. Of course, you're going to ask me, how are you going to fix this? You're going to ask me. I knew you're going to ask that. And because we have a static projection, if we change the model, look what happens. Go to "plastic, " go to your sculpting tools. I love these tools. I love these tools, man. Let's go and use the pull tool, good. And let's change that. Now, I can go close and by pulling your geometry... Oh, this is not the one. Sorry, sorry. This is the old model. No problem. Processed models for NAB, and let me show you that. Where is my final thing? Okay, this is because... I'm going to show you a nice little tip here. I think I'm missing a material here. There you go, I'm missing a material. Did you know that you can go to your "texture manager" and you select this and you say, "Okay, relink textures." And you're wondering, "But where am I going to look?" If I knew where it was, it wouldn't be missing in the first place. You just select one of the folders where you pretty much think it's in there and you press open, it's going to search in all those folders and it's going to find it. And Mary Ellen just went... Okay, it's the cool stuff. Okay, I need to find that car so I can show you that sculpting thing. Come on. Is this it? This is it. We did find it. Okay, this is not the bridge action I did but you understand what's going on. I'm just going to make this editable and just drop this here and show you what the repercussions are from the sculpting. So, pull. Now when I start pulling this, you can see that the geometry is moving. And I'm going to put it in the luminance channel, and now you're going to see that it just works. So from a specific angle, from this angle here, regardless of how much I messed with the geometry, what you're going to see is a proper car. Now, you want to see how it looks like from another angle. Weee! Okay. But if you combine a decent reprojection and you actually spend your 25 minutes and then just find tune the sculpt because if you see how the actual cars look from the final ones I did, you will see that they're very decent models. So let's see that. Anyway, the sculpting I don't forget, if you want to push it back down, you press Command and it pushes it back down. Pull it up a bit, just go around and bring the geometry so now I'm pushing the geometry. When you see the color coming through, bring that over. Then I'm moving the texture on the model by changing the geometry, okay? And it goes the opposite way, when I dig in, geometry moves that way. And funny enough, from that particular angle you need, it's going to look perfect. And then you're going to ask yourself, "How much degree of freedom do I want? What did the clients say?" If the client says, "Make it look at it like it's kind of skidding in." And you say, "Right." You take your car, you have a couple of drinks, you go out, you skid your car, you crash, you measure the distance, you calculate, 20 degrees. And you're like, "You know something? It can hold. It can hold. When are you releasing me, officer?" That's what you say after the crash. And let's open the Escape in it's final form. So, it's Escape, Escape, good. And you can see, where is my camera? I can go to cameras if I can't see it. And use camera as default. That's how my car looked after I measure the 20 degrees. But my wife didn't because I parked it sideways. So this is the car, the actual final car. It's perfect for what we need and a bit more, always make it a bit better. It's perfect. So, just add your motion and the particular scene is ready. So the projection you understand the concept? Okay. Make your frame as big as your pixel resolution of the image and put it in frontal, eyeball it, play with your camera focal length to adjust the car. If you get used to this, it will be a big part of your workflow if you want to put objects into 3D scenes and so forth, you can use the tools. But some scenes are made for you to do it by hand. The joy of rotating the camera and nailing it and saying, "I'm better than an algorithm." So, no questions about this? Because now we're going to move to the other step which is how do you model a rim that has radial symmetry not divisible by two? Yes, please? So the question was, do I manipulate the projection to fit the model? Or the model to fit the projection? Or both sometimes? So, you manipulate the projection when you're doing the original placement of the car using the frontal projection, that's the projection manipulation, basically. Once you've nailed that, that's locked down and then you find tune the position of the geometry to comply to the actual projected shape. And the better job or... One thing I was working on, and I'm going to show you retopology, a big thing. I didn't have to use these models. They were just there and I thought, I'd use them and it kind of catered for the situation. You can actually model your own low polygon car that will accommodate for the particular case fairly easily using the polygon pen and so forth. And if you want to see a bit of the polygon pen and how we do reprojections, by a raise of hands, who wants to see retopology using the polygon pen on a car model and so forth? By raise of hands. Raise your hand, Steph and Mary Ellen. All right, now we have... You know how to do retopology. Okay, you don't. Good. Well, I'm just going to show you a couple of tips. It's going to be a couple of minutes. So let's go back to our 3DS model thing because the more options you have in order to do a project, the more you can choose depending on what the requirements are. Some processes fit different projects differently. So let's go and find our Turbosquid models and get the Escape and load the 3DS. So this is going to be very, very quick because it's relatively easy to do. And I'll do this for the third time in a row. Unfold this, go delete, they get rid of all these, okay. I had to do that. I mean, the music is too loud. So since we already have a model, sometimes, we'll do something which is called retopologizing which means basically we use something that's already there but it's too high resolution to be useful for what we are doing. And we're actually going to model on it by using that underlying geometry. And the ideal tool and one of the best in the industry, drum roll, is the "create tool," "polygon pen." All you have to do is go to one of these modes here and... Sorry, polygon mode, excellent. And with the polygon pen, we're going to put the reproject result. So we're going to go in point mode actually. And I don't even think we need to be in a mode, polygon pen. No we do need to be in a mode. When you use your polygon pen to paint using points, with the reproject result, you can see every time you press your point, the point actually goes and projects itself on the geometry. So you don't need to be in any particular view, you just start drawing points on your car and the polygon pen will actually stick it on the same surface. And what I'm going to do now is because the polygon pen is extremely smart, just like I am, I'm going to Command+Click and drag edges. And as I'm dragging these edges, you will see that those points are, see that? That is because where this point was, there was a hole. This ended up here because there was a hole in the mesh over here. So just make sure that when you're doing this, that your points don't end up in the void as you can see because reprojection means it's firing a ray, as we call, a trajectory and hits the first geometry it finds. If it's a hole, there is no geometry so it goes to the other side. So just be prudent and once you get the hang of this, oh my. In a few minutes, I mean hours, you'll be able to model a car that has enough geometry to accommodate your projection needs but not enough to give you a headache. So you can get way retopologizing this car in, I would say, half an hour, 45 minutes and it's going to be very streamlined and you can keep the rearview mirrors if you think as a present. So this is what you have to do, points mode and make sure that your reproject result is on and everything else is an automated process, okay? Now, I'm going to add ever so slightly to this because it's a bit cool what I'm going to show you. I'm going to turn on my Gouraud Shading lines, and I'm going to do something odd. I'm going to space this out. I'm going to just pull it out a bit. I'm going to get rid of my rearview mirror. So, Shift+double click somewhere here. It's not working. Oh, it did work. So I'm going to select these, make sure I'm going to select "connected" and get rid of that. I don't need this so I'm just getting rid of that. So select, "connected," get rid of that. Anyway, we're in a good position. If I take this polygon, let's assume you realize that, "Ah, rats. I wanted high resolution." No problem. I'm going to subdivide this thing. So right click, "subdivide." Good and let's say we want one more subdivision. That's the shape I want, that's what I want. And I didn't want to go and paint all these one by one. But now, I wanted it to comply to the shape of the car a bit better because the polygons are small, I'm going to get more resolution. No problemo. You go and get what's called a "shrink wrap deformer" which you put under the polygon. And you tell the shrink wrap that I want the collision object, the target object, to be that. And if I turn this off, you will see that it has been reprojected on the car. Did you know that Mary Ellen? No, you didn't. I feel much better now, for some odd reason. So what you can do, once you retopologize the car with very big polygons, you can actually subdivide it, then use something called...let me undo because the projection with the shrink wrap deformer is done on a normal basis. So a normal is a line that goes right through the center of the polygon and it indicates the direction of the polygon. So, if I was to... Remember, I just moved it out? What if I've retopologized the whole car? So let me get rid of my rearview mirror one more time, there you go. So, remember this little guy was here, okay. What if I want to move it away but not in one direction? All you have to do is select your polygons and do a normal move. This moves each polygon in the direction of it's own normal. So click and drag. It's what the extrude tool does but without adding the geometry. So if I had a car shell, and I wanted to bloat it so then I can subdivide it and then shrink it down using the shrink wrap deformer, that's how I do it. I would not move each polygon or do something like that. I will actually do a normal move. So it will kind of grow bigger. So that's another way to do retopology using the shrink wrap deformer and just using your polygon pen and it's an interesting avenue to investigate. Good. Let's move to car rims. Now, the main problem with rims is that nobody knows which ones they're going to use until the 11th hour. So, certainly we're not going to use the rims that were on the 2015 model which was the one we bought on TurboSquid. So we have to model the rims. Now, you can scour the internet or the interweb and find all sorts of interesting photographs. And that's what actually saves the day. And let me show you some references. It's not rules, it's this one. There we go. So let's go to the downloaded references. And this is the one I'm going to use, this was the final rim and this is the one I undistorted in Photoshop and this is the best rim I found for this one, the best image which is atrocious but it's good enough to work. And then I have different references, this one, this one, this one, this one. What else? This one was very good to get the volume, plus the fact that there are very small parts of the visuals so I can get away with quite a lot of stuff. And then I found the manual. This is the other rim I modeled which was much harder to do but we don't do hard today. So let me show you how to approach modeling a rim like that. I am going to open only one file and sorry for moving back and forth. So, download references. I think it's number one, just so that I can tell you what I'm thinking of when I'm making the rim. So, number one, try and find a face-on image of something so you can get the basic idea of the shape. It's very tricky to understand what's going on especially if you have radial shapes if you see them under an angle, believe me. It took me several hours to figure this one out. So, what I'm going to do is go to my front view and I'm going to load by going to the options configure, or pressing Shift+V, go to the back and load that image. So, preroll, NAB presentation, download references, rim. It's not that one of course. It's always the other one. Good. So now I've loaded an image and you can load images in your orthographic views, okay, but you already knew that. Good. Now, the good thing about this is I can rotate it, I can change the transparency so it's nice and discreet. And I can do things like scale it and move it. So I can center it, put it exactly where I need it to be. Okay, now let's do something which has nothing to do with 3D, it has to do with logic. This has a five point radial symmetry, okay? But they're not all equal. So the designer thought he'd make our lives difficult. How do we find out? How do we work with this? Well, one thing we do know it's a radial symmetry, five part. Number two, if we take any fifth slice, that has symmetry
Resume Auto-Scroll?