Create Your Face (Blender Tutorial)

hey guys welcome back to another default cube cg matter tutorial and today we are doing something that i'm actually really excited about which is making 3d models of any face specifically my face using only photos so I took 6 photos of me and by the way this is better if somebody else takes photos of you so you can just stand still and the person revolves around you but I just went like this and the results still work or this results are still good I should preface this by saying I'm recording at 3:30 in the morning and I'm tired but I somebody posted this into discord so I thought this was something I should talk about it's really cool okay so we are gonna make models using photos and to do this we are gonna use an add-on because blender does not do this by itself without a ton of manual work so you're gonna open your favorite internet browser on this link it's going to be linked in the description you don't need a guess how to find it this is the Keen tools face builder add-on for blender 2.8 2.81 anything with a 1/8 and then a plus so even things in the future will work and this keying tool stuff right now is in beta meaning that it's free and then one day you might need to pay for it so try it now and this is also the same company or whatever that makes the face bill there's stuff for nuke and other stuff like that so it's cool that they made a blunder thing so what we need to do is click the blue button you click it it's gonna download the add-on you do that and the add-on is really just a bridge between you know this core library and blender to interact but we don't actually need to download this ourselves because the add-on actually has a way to download it for us and the whole reason there's this stuff is because of licensing stuff with blender but that's fine ok cool so now we have the keying tools add-on and we need to install it so I'm just going to open a blender and wait for that I'm gonna be using version 2.8 1a as opposed to 2.8 1f and let's go fullscreen ok cool so now we need to install the add-on before doing anything go to edit preferences add-ons and let me just make sure I don't have it installed already no good I uninstalled it just so I can show you this you're gonna quick install you're gonna navigate to wherever you saved it and you're gonna see Keene tools install add-on and then you're gonna see it it's gonna be disabled and we enable the add-on and then you're gonna see a bunch of stuff you know details about documentation bug reporting and all this but like I said we need the core library now I don't know if I need to download it again cuz I already did but to do this through blender you have to read the license agreement I recommend you actually do that but I'm just gonna say I did and then you click install online latest stable release is fine and that is gonna download it's gonna take a moment so of course I already did it so maybe it's like deleting and then reinstalling and all that okay cool so now it says everything is updated installed successfully so how do you actually use this well there's a lot a lot of things you need to do before using this properly but we're just gonna go over some of them so for example you need to make sure or at least it's preferable if you do eat everything in your default scene so I'm gonna do that next we're gonna hit n for the properties and you're gonna see face builder and you know whatever other add-ons you have sound like screencast keys but you should see face builder you're gonna click that and you may be wondering at this point what do you do there's so many options you click the button just you click it it's gonna load the interface and you're gonna see it's gonna give us a default model of the head you know it's you know it's a model as it's not I don't know what else it would be but it's a model and we have all these options of what we can do with it my phone is going off that is fine and basically what we're gonna be doing is using our photos to make this um face deform into whatever target face so right now this looks like somebody doesn't look like me so we're gonna fix that so first thing we're gonna do is we are gonna add our images so just click Add images navigate to wherever your photos are they recommend I think they said seven plus I'm gonna use six and it worked for me so I I would recommend getting as many high quality photos as possible make sure that when you take these photos they are taken with the same the same camera but also on top of that the same focal length same sensor size although I don't think you can change the sensor size if you're using the same camera but basically don't zoom in just take the photos keep them in the same resolution it will make everything easier open images okay cool so you're gonna see that we have six images being the six images we imported and you're gonna see the resolution here and that's the frame size in this case frame size is all the same which is good you're also gonna see the exif data I don't know if there's a way to pronounce that but that's basically metadata in your photos that's you know says a bunch of stuff like the focal length you know your sensor width and like what camera took it on so I took it with my phone a oneplus so there you go and yeah so now we have our photos imported the next step is setting our camera options so that is saying what sensor with did you take it with so you need to know that information and what focal length okay well let's say you don't know that you just didn't write it down before you did it okay then your next step if you don't know these is you look at the exif data and you can click read exif and choose one of these photos and it will tell you the data for that photo now in this case they're all going to be the same than we think that's gonna be different is the name of the source file as you can see it updates the source file but all the other information is the same same phone same focal length etc so you can use this exif data to put this stuff in here you know you have your focal length your sensor width and all that but let's say your camera or your phone didn't even write any exif data okay and then effectively you don't know it so we're just gonna click focal length estimation which is gonna predict or mathematically calculate our focal length what it should be when we do our whole head pinning process which you're going to see in a second so I'm just gonna leave this on even though I know some of this information okay cool so we've imported our images done the whole like focal length estimation now we need to do the long process of pinning which isn't that bad compared to modeling ahead we are gonna by the way you're gonna see that when we imported our images we get six cameras being each of ours so kind of looks like militia pointing at our head click the first one it's gonna get us enough in the view of the first camera click the second one we're gonna be in the second camera and you can see these background photos are changing and there's even some text saying you know what you should do down here but I'm just gonna say it so the way we're gonna align again in each of these cameras we're gonna align the head the way we do this is with pinning which means I'm gonna select a point on this mesh and bring it over to where it is in the image so let's say we are talking about this front of the nose right here on the model I click and hold so I'm gonna click hold and then drag and right now it's being pulled from the nose and I'm just gonna put that you know about where it is fine now the next pin let's say we're doing it on the I don't know what's that called the left wall in my case of the eye whatever I'm gonna click this tear duct I'm gonna click this and now I'm gonna rotate the model so the second pin does rotation and I'm just gonna line it up to where it is in the image now we're gonna do it again again each time it's going to refine more and more for the third one we are gonna keep messing with the orientation and the scale so we're just gonna line that up and already you see we're getting something that's half decent we're gonna make it much better but you know this is a start and now the rest of the pins everything beyond three is gonna actually change the shape of our model and that's a good thing because unless you look like the default model loaded in you're gonna need to do that so here's the other side of the eye I'm gonna click that and drag it over and it did a bit of a correction again it's wrong but that is fine because we're just gonna keep adding pins I'm gonna add another one over here going to take this like where the bottom of the nose the I don't know what it's called spring that over here and it shouldn't take too many pins for you to get this right you want to go through a very obvious features like the corner of the mouth before you do like random spots on your cheek or whatever so I'm just gonna line up the obvious things and of course this is a great example of how powerful this add-on is cuz these photos are sideways they're garbage there's hair covering like all the details we need so it's gonna be a good to end I took the photos myself okay so now we're gonna take the top of the ear and bring it over to where I think it is again the better you can do this the better the results are gonna be and I can see the mesh is really lining up so we're just gonna do a couple more top of this year bring that over we can also do the bottom of the year although I don't know exactly where it is and if you don't like what you did like for example I don't know about this one I'm just gonna control Z twice and it's gonna undo it so let's keep going if there's any more obvious ones normal you can do the chin but I can't see my chin in this case because of the facial hair the nose looks pretty lined up you could you know even adjust it a bit more and we're just gonna do this process for a couple photos but I'm only gonna do it in depth for the first two okay so I think at this point we've done all the main features and at this point all we have left is like major features like the middle of the forehead which is kind of hard to be like oh this spot on the model is this spot on my forehead so we're just gonna go for like the secondary features like maybe right here I feel like that should just be over here and the better you do this lining up there the better your final results are gonna be so I'd recommend spending some time doing this and you can also focus on areas like the neck you see that's part of the model I don't find a dad important because what I care about extracting at the end of this process is a model of my head not necessarily the neck as well but you can do that let's also add a pin for this ear and we can adjust this pre-existing pin okay I must say that's pretty good so that is the first image that information is stored now we go on to the second image and you see the backgrounds changed and this one's gonna be a bit harder because I don't know a lot of the you know spots if exactly were my eyes and stuff like that but if we go back to the first photo you see that information still there so it saves our pinning you see this icon by the way we should do a save let's save I'm gonna call this should have saved earlier cool okay now let's do another I'm gonna pick so I know approximately where the front of my nose is I'm just gonna select that bring it over again first pin is location second pin I'm gonna go for this side because that's the one I can see second pin is just you know rotation scale all that so that's about there third pin is a bit more of that so now we're gonna pick the other side of the eye again so somewhere around here and of course we can zoom in to make that accurate but that's fine and now we're starting to deform our mesh with the more successive pins so now it's line up the corner of the mouth and this is a great image to talk about because I can't see half of the information here but we'll try our best let's do the top of the ear that's an obvious one and the bottom aka the ear lobe some people call it whoops by the way if you ever accidentally exit the camera view because you didn't lock it you just click this button again and then try not to do it again let's line up the chin I feel like we know where that is and then we could do the lips again the thing about this is it highly highly depends on your image so like this process is very different from a front-facing photo we can bring up the eyebrow and bring the forehead in a bit and I'm just trying to show you that even though I'm kind of doing a garbage attempt at this it will still give pretty good results and I'm very impressed with this pick it up before it leaves beta I don't know what what's up with it I don't know if we're gonna have to pay for it or what but I I'm not in the business of finding out okay finally let's set up our nose and is that looking good that's looking good maybe just bring this face in a bit and you can generally kind of tell if it's doing a pretty good job just from eyeballing it but if you take better photos it's gonna be fine so I'm gonna save and at this point I have four more photos to go through so I'm not gonna bore you through that so I'm just gonna fast forward that process and talk to you in just a minute so while I finish up the pinning process I just wanted to tell you about my patreon which helps fund more tutorials like these are really apprec-- already out of time ok and check out my patreon ok I would say that at this point I have done everything I can so I've gone through all six images and at this point you should also be done assuming you're doing this along with me you have now pinned everything so let's actually just preview these so we have the first the second the third the fourth maybe this one could use a tiny bit of adjustment let me just fix that chin the fifth the sixth and the whole point of me showing you this is saying I did a horrible job at this so if we even get half decent results out of this then you know this add-on is really really good and my testing has shown that that is the case so at this point you are now done with your pinning and I guess I should also mention I mean I could have mentioned it before that you can also affect the pin size and pin handle radius and this just affects how the adding pins will deform right like what kind of radius of effect does it have kind of like soft selection I guess you could say in some sense but yes so you can mess with those but everything I did was with these default settings and you can also mess with the color of everything I haven't played around with this I'm assuming this will affect the you know yeah red wireframe and then you can affect the color of everything else and this is transparency or something yeah okay cool so you can mess with all that but assuming you are now done with all your pinning the next step is I mean the next step is you could be done theoretically like if we go into the 3d view it's kind of oriented weirdly you can see maybe you can't tell but if you seen a lot of default cube a lot of CG matter you can tell at least I can tell that this looks like me you probably can't because you're not familiar with my head but we'll texture this as well but you can see that all these cameras have now been positioned around the head my head specifically in a way that it's the same orientation as where the pictures were taken so for example this camera right here has a top angle and I think that says it's the fifth image sort C this should be looking downwards on the finish nope one of these images okay whatever one of these is facing from below and we can see I'm just trying to prove a point here we can see that this is probably the corresponding camera so each of the it's kind of like photogrammetry so I'm gonna stop talking about that so at this point you can take this model again it's 3:30 or something for me you can take this model and export it and do whatever you want with it or what we can do is move on to the texturing step which is really cool it projects some these images back onto the mesh using some math it's not as easy as project from view basically because it needs to average out all six photos and by the way you could also take photos from back here although it'd be difficult to align okay cool so now in the texture tab we have a bunch of settings first of all what resolution do you want to output in do you want a 2k texture a 1k texture etc I'm gonna go for 1k so divided by by 2 gives us thousand 24 5,000 24 so this is a pretty low resolution texture but that's fine with me next we need to talk about what UV map you want for your model again this is a model it needs some UV map and that's gonna determine how our texture is kind of created because it's gonna form to that UV map if you understand what I'm saying it's gonna basically if you look at this texture it's gonna look like an unwrapped version of our head and that unwrapping is determined by this UV so butterflies where it kind of splits your head in half and flans it and you have a bunch of other ones I'm gonna keep it on butterfly which is fine but you can use any of these UVs and then you can either just create texture which will do what you think it does or we can actually mess with some of these settings and I'll just quickly describe what I understand about these so angle strictness it goes between 0 and 90 I think 0 means that every point on the head so every pixel that we're making for our texture is gonna be kind of averaged out between these six cameras so it's gonna look at the spot see which cameras can see that area and then it's going to take all those you know pixel values all those color values and it's gonna average them out so and that might be a horrible idea if the lighting conditions change so it's gonna look weird if you have it at 90 it's only gonna look at the camera that's kind of the best one in terms of facing it and just inherit that color but that's gonna look bad because it's gonna have a bunch of seams so you want to keep it somewhere in the middle they recommend 10 to 20 I think they say and then expand edges I think what this one does is it basically you're gonna get a texture but since you don't have perfect coverage there's gonna be you know gaps on the bottom of your chin behind your ear and I think what this does is it's basically like a margin and it will extrapolate a little expand edges then they say that these two buttons are experimental features equalize brightness will equalize brightness from the various photos which could be useful could really mess something up it's an experimental feature and then equalize color so all that explanation and I'm not gonna change any of these but it's important that you understand what what these do and when I create texture it's gonna take a very very long time so I'm probably gonna skip ahead but make sure you pick whatever resolution you want and I'm gonna hit the button see in a second oh never mind dad does a little awkward forgot about this you hit create and then you basically pick I'm dying of embarrassment you basically pick what photos you want incorporated in this process do you want all the photos to be incorporated in this texture building or only some of them I'm gonna keep all of them enabled and then it's all I'm also gonna make sure that it automatically applies to the created sector so it's gonna make the material for the head and apply the texture to that material so now I will see you in a minute okay so just finished loading like a second ago and I would say that was like in real time something like 15 20 seconds at 1k you can imagine if you made a 4k texture that would take a lot longer but 1k took about you know on my computer 15 to 20 seconds and now we can see this kind of it looks a bit scary but honestly considering that I took these photos myself and I tried my best to look forwards this is an amazing amazing result right you need to remember if somebody else takes these for you and you don't have facial hair and you can do the pinning with a spending more time on the pinning this is gonna look much better but you can definitely tell that looks like me I mean again you don't know me perfectly but and it definitely has high accuracy in the face and the head should be pretty accurate but of course we didn't take any photos that come from this angle or even from the perfect side or anything but these are really really good results and maybe you're like okay I know what I look like doesn't look exactly like me you need to make sure that you remember that the way your face looks is highly dependent on the focal length of your vision meaning that in this viewport right now we're looking at this with a focal length of 50 if we change this let's change it to like something like 70 this is gonna make your face look a lot um I don't know a lot wider a lot blockier whereas if we changed it like this and then zoom in it has a different look to it so at this point it really really looks like me to my eyes so I'm just gonna keep this at 35 and you're gonna notice that the most amount of distortion and the texture happens in a couple places it happens on the pupil area and around the eye why because I wasn't looking exactly fords every single time especially since I was kind of like looking at where the phone was went when I was holding it another thing that might be a bit warped his hair if it moves my hair's pretty short so it actually stayed in the same place and you should know it that you know it's a 3d model of your head you need to put a particle system on this if you want to hear that isn't you know flat like a texture and then you can also have some distortion in other areas but notice that it actually filled in some areas that there's no way it had access to like I don't think any photos were inside the caverns of my ear plus you can see it's pink which means it extrapolated the skin next to it because if it actually got the texture it would be black but then if we go far enough to the back you know there's no saving this so you just need to take more photos and what I was talking about with this UV unwrapping in all this we did butterflies if you go to UV editing you can see you can see the unwrap we have it's looking pretty scary I think I can hide everything and then this this is a thumbnail this is the texture generated and you can see it generated the texture for this unwrap and you can see there's some like you know some weird stuff like there's some bridging between the lips and I would imagine that's the expand' edges basically expanding the margin it's extrapolating the texture but you can you see this is actually a crazy nice texture just for 1k - Kay we'll look you know it will retain a lot more of that detail make sure you're in rendered mode and I'd imagine that in the shading workspace yeah just made a psdf evident already have one yeah I'd made it you can see face builders in the name and then just put the texture in the base car but um yeah that is pretty much all there is to it of course I I keep saying in this tutorial that you can do it better but this is a perfectly fine result and you can see it has nice topology and all that again I'm not a 3d model or anything but I'm pretty sure this is pretty good topology it's a it's a bit high density but it's good top oh you can use this for I don't know fine just one application already is this but there you go you now know how to make faces using this add-on I don't know what the state of it is in the future so if you're watching this a couple years from now and it's no longer in beta or a couple months or a couple of days then I don't know that I can do for you but if you enjoyed this tutorial I would really appreciate it if you checked out my patreon because that is how you fund these tutorials you are definitely directly contributing to me and of course you get benefits but that's not really the point and yeah that's everything I got for you guys

Channel: Default Cube

Views: 666,010

Rating: 4.9584346 out of 5

Keywords: face, head, 3d model, blender, tutorial, keentools, addon, facebuilder

Id: sf88UeC7LmE

Length: 23min 33sec (1413 seconds)

Published: Fri Dec 27 2019