 So, good morning! Are you ready? Because did you put your seat belts on because you see there is basically the whole team is going to do a little section so they're like ready and pumped and you need to be ready too because it's gonna melt your brain. So let's go. In 2015 I went to a storytelling workshop in Rome and I met Metulon and he's a storytelling artist. He worked at Pixar for many years and he was you know traveling the world talking about the art of story and it was very inspiring for me to attend that event and it kind of planted a little bit the seed of you know thinking okay in the future maybe it would be great to work with someone that has such insight and passion for storytelling. And fast forward a few years that opportunity presented itself so we ended up actually teaming up with him and developing the project of sprite fright. The goals of the sprite fright project were fairly simple. One was to push our craft and knowledge in storytelling so we really wanted to do something that has some entertainment value and at the same time to always push the development of Blender and in all the things that we do with Blender but especially to focus on the pipeline and make sure that we can collaborate together more efficiently to try and you know raise the bar a little bit in the complexity of the production still make sure that things don't crumble and to share that. And that's like the main point that I want to drive here is like the sharing the production knowledge that's at the heart of the mission of the Blender studio that is really what we do. We put all this effort in working with Blender and in figuring out stuff and then we put it out there as much as possible. So the journey of making a film is not just the film itself and being able to bring it to the audience but is every step of the way and we try really our best to document it in many ways and to share as much as we can. This is for example on our blog. We publish articles, we publish interviews, we publish anything really and we have production logs that's like one of my favorite features that we have on the Blender studio website where the artists are sharing on a regular basis on a weekly basis the work that they're doing and we've for this project we've been sharing all the edits of the film. You can enjoy a hundred plus versions of the film from the beginning until the end including with the directors commentary so that's pretty incredible. And of course all the tools. I'm going to play a quick reel for the sprite fright kind of making off how some things came together and then I'm going to leave the stage to the rest of the team to talk about what they've been doing. Enjoy it. I thought he was going to introduce me. So my name is Jalte Hjelmozon I was the co-director on this thing but not only that I did other things. You're wearing many hats of course but these were this was a little team that was doing some of the story stuff but on different time zones so Turk and Matthew they would do a lot of storyboard sessions together and then in the morning when I woke up there would be like this whole batch waiting for me with like script updates. So every morning was very fun and then you go over it and you iterate and then you know at the end of the day after like a full work day I had a you know one to three hour session with Matthew we're going over everything making sure everything's clear. I have this tendency to press control S when I'm just working just to say apparently Skype makes a snapshot of it now and then so I have like a list of random things that I said. Anyway his original pitch was this kind of smurfs gone bad and the kind of tone of it was gremlins meets Sean the dead so it is horror but it's not trying to be necessarily gory but it's supposed to have this humorous kind of side to it but the original story I think I was looking at some of the original drawings they were more gory to be honest and it's kind of interesting so we we had to kind of find that balance but all of those drawings somebody had to put them in an edit so that was me putting on the editorial hat and start putting it all together and you know putting some temp score temp sound effects all of that stuff but not only that I was butchering Derg's drawings severely so in order to accomplish whatever I needed I was full on mangling it drawing my own stuff and doing what whatever was needed now once I had that animatic we had this kind of spatial problem because this is not going to be a 2d animation this is going to be 3d and then how do you start having a conversation spatially about everything so I made my own little rig like usually you do a floor plan you drew a drawings whatnot and Matthew started doing some of it but it kept changing and it got a little bit confusing so I figured hey let me just make something like this and I'll just animate it this is like ten times the speed so just so you got a gist of it so this became kind of at least the foundation of where the camera could be and we could try different things out and kind of spatially between the different sets where they should be and I have very little time and then as I'm doing the layout kind of moving on from there a lot of the stuff is still being designed so sometimes there are whole sets that don't exist and I but I know roughly from some drawings what they kind of maybe look like so I just start doing random stuff and sometimes I'm a lot of times it isn't using primitives and just figuring stuff out the this is like actually a more advanced rig than what I was using earlier on so you just have to kind of make do with what you have and for example oh she rescues a bird I'm like great where's the bird wow we don't have the bird so I'm just okay here's a cube you're to do and flappy things and you have a bird and it conveys the story so that's it you don't need anything more yes so in layout this is just an example of what the layout process looks like if you kind of take another camera and you look at it from the outside you see there's like a lot of spatial cheating going on and that kind of stuff but that's the fun of it it's a four dimensional puzzle per se just to touch a little bit upon the cinematography it kind of does feel like two different films one of them is more calm and quiet when it comes to the camera movements that are very controlled and the other one is a little bit more chaotic so it kind of goes into this wondrous steve Spielberg ask thing to more of a evil that horror kind of thing and there were a couple of shots that were like inspired directly by some choices this was as far as we actually took the whole at the right thing I pushed it further because he's really something about it and I pushed it way further and it became like a parody of itself so it was a totally back down so I think that's the only thing that remains in the film the actual editorial there's so many things that you have to handle there's also you know there's the scratch dialogue and the script is changing the dialogue is changing so you have to redo all the recordings and you have to splicing together every single potential sentence has 20 different takes and then you start splicing those together into the take that you want so for example this sentence was not a one-take thing she was never in the same room saying just this one sentence it's taken out of many different takes recorded at different times so there's a lot of like this there's a lot of this that goes on with for editors like oh she didn't say it this word in plural but the take is nice can you find an s I'll find that s so you might think of all of these things as a kind of a nice conveyor belt but really it's a little bit more complicated and there's way more kind of symbiotic relationship back and forth and usually the editorial sits kind of in the middle of it and it gets even more confusing but once you have it all together like you kind of snap the puzzle together it's very satisfying it it kind of becomes this satisfying piece of artwork speaking of artwork here is our art director and the good old chair okay we don't have a lot of time so sprite fight was easily one of the most complicated productions that we had at our studio we had lots of characters lots of props and not only that the characters went through multiple stages of wear and tear which made it very complicated because that's something that we never did up to this point like some character that gets hurt that bleeds everything so I'm gonna talk a little bit about the early production process shortly even before we got on board with this a lot of this complicated production the magic trick was to make something as complicated as that look very simple so it was an exercise in simplifying things so before we saw anything Ricky Nerva the production designer he's a character designer at Pixar also production designer off up he spent a lot of time with Matthew developing character concepts and dueling around so there were literally hundreds of drawings of these characters in different stages and the characters had to you know they're they're the main cast we had five five five character five main character seven ten billions of sprites so they at some point we we have bothered down to okay this has to happen in Great Britain because we want to make it more European so this was the line up that Ricky made for the characters and okay so where do we start at the beginning it was only it was only healthy me Matthew and Ricky and how do you tackle something as finding a style for a film like that so the easiest thing to start with is not with the characters but with a prop so we actually decided to start with the boombox of Jay because Ricky did a little little style sheet of that and we wanted to figure out okay how do we make this movie work in in three dimensions how do you find the style and the language of the shapes of everything so this was my first pass of it very rough just blocking out shapes and everything and following the design pretty faithfully and you can see it like there's some crooked shapes here and there but we the core of the design language is there it's a chunkiness that's what Ricky wanted so the detail is a little bit bigger than in reality because the characters are also very stylized their hands are bigger and they have these simple shapes that are based on so the props also have to reflect that so the sprite for it is somewhere in the middle between something that is a model and a realistic representation of that prop this is what the final boombox ended up being you can see it's a little bit more straight we wanted to make the shapes not too crooked so it's not wonky and all skewery and so on so it's a little bit you know more chunky that chunkiness is still in there textures are also simplified the local contrast is simplified and that's kind of the language that later on we apply to all the different props that level of magnification of detail make it look like a toy with Chris Matthew kind of comes from a family of toy store owners so with the characters it was a little bit more complicated because it was a big cast and unions going to show more of the actual modeling process but very early on it was about these simple shapes that reflect the personality of the character that's what the drawing that Ricky made and then he asked me to just take that into 3D because he was a 2D designer and he didn't really know how this is going to work in 3D how the volume is going to work how the character is going to behave in relation to each other and that actually reflected back to his design process because he kept working with these volumes and tried to push things a little bit back and forth and this was the approved character sheet for a long time until we got on board and even iterated more but Julian's going to show that the environment was a little bit more complex because Ricky wasn't that you know he wanted to give us a little bit more leeway with the environment design so this is a very early environment tests finding out some you know wonky shapes that work with the character shapes we did a few of these tests but we really wanted to do something that fits more into the film and also test the different day times because we have this we have this bright side of the film and the dark side of the film so we decided to do the first environment test with the U tree it's a it's a tree type that's found in in Europe in graveyards so it's very spooky that's the tree that Jay actually dies in and we we got this drawing from Dirk and we pretty much transferred it into a free model Simon made some rough shaders and we tried to put this together okay when I have kind of realistic lighting realistic surfacing but stylized shapes this is at daytime and this is at nighttime so at nighttime we removed some of the detail and and made it more skeletal which that was great because now it wasn't just a conversation anymore it was actually a thing that you see and you can look at it with your eyes and also at the same time character blocking was going on so we were able to put the characters into that set as well and see if it works in that case not so much because we figured out okay the characters are so expressive how do we make it with the environment when if the environment is equally you know busy it's gonna it's not gonna work because you're not gonna be able to focus on the characters so we had to really find out okay what do we remove how do we simplify things and how do we make the characters read a little bit shortly after that we had a concept artist called Todd Polson work with us for a brief period of time he actually had a good point and he was seeing these environment renders here and he was thinking okay how can we stylize the shapes even more so he made this drawing of these very stylized mushroom shapes and I tried to kind of make a little scene with that to see how would it work you can see the sprite character here in the background and so we have this little scene with grass and the sprite in the background okay with some light and some cool shaders on top of it yeah it looks a little bit too busy because the grass is just very noisy we were not focused on the sprites really so the thing was okay let's find a way how to remove a lot of that detail so we use clovers instead of grass and we use bigger patches of soil and that's how that scene came to be so this was like as a kind of testing ground to you know have a conversation how do we how do we find those chunky shapes in the environment how do we make the character read how do we use light and color values to make the characters be separated from the background and how does the chunk of vacation work with the character later on that became kind of a test graphic a banner graphic for the blender cloud we polished it a little bit further and this is how it ended up being so next it's going to be Julian our character artist all right thank you very much so I was mainly responsible for the sculpting and modeling side on the characters for half of the production which of course because we had such simple character designs was actually super easy to do you just basically start from a sphere and sculpt the final character so I mean the elder sprite is basically a sphere so it's easy right but as it turns out there was much more to it and was quite different than anything that I've done before I would actually start say there's a bit of a misconception that I was living by and I think a lot of people do and that's like that you typically think you start out with a 2d concept and that the concept artists and the all the people involved in that are figuring out the design and then the 3d artist comes in and recreates the design as faithfully as possible but so I basically snatched the at the time latest approved concept art from the sprites this one made by Ricky Niava and I made a sculpt of it and this is what came out of it in like half of a week I threw it in there as an appetizer and I thought okay this already looks kind of good but it led to a lot of confusion it's not really what I was expecting so because the plan was actually to not take these drawings as holy but they're sort of the the starting point the conversation piece to find the core design principles for each character and then we would go into 3d and have a whole collaborative process of actually figuring out how the characters are supposed to look like so after all of that process this is what came out of it the final 3d characters and it's really interesting like the the first take away the first lesson that I kind of took out of it is that you really need to go into 3ds soon as possible because the character designs changed completely some of them were the original designs were completely scratched so you start to work more iterative iteratively and then refine them as you go along and even like the simplest lineup at the start helps a lot to give you a base to start off with these really simple shapes and then adding more complexity on top the the second thing I really learned a lot early on is you should really develop all the characters at the same time we were lucky that we didn't have a lot of people so I actually got to work on all of the characters and sculpting but I still tended to go ahead and and work on one character first and whenever we did that whenever we rushed ahead and worked on for example Ellie over here in isolation we noticed later on she she doesn't fit at all with any of the other characters anymore and so we went back and readjusted the design heavily changed it and tried to progress all the characters evenly at the same time so to keep the style consistent this basically ended up in similar issue as with this right in the beginning where it just didn't fit what we ended up doing another important thing though is this this sounds kind of strict but you should really not be too strict with this workflow sometimes we just went out of a way to make a fully post sculpt to just try something out with for example in this case with the mouth shapes how cartoony we can go and this gives you a lot of ideas on how to get an appealing design how far you want to take the style but a lot of these sculpts were completely throw away like some of them were never used but that doesn't mean you're wasting time actually taking a bit of time to make these experimentations to make actual expression and post tests tells you so much that you will save a lot of time in the long run on trying to fix issues that it did you didn't foresee for the characters we actually went a lot with design principle that Ricky called simplicity in a nutshell it's like sort of starting with a simple design like with elder sprite over here with just a sphere and a hat on top of it and then when you start to add more detail on top you try to not disrupt or distract from that original design language so we tried to go for that but it's really important to remember that doesn't mean you should go for simple shapes doesn't mean that the 3d model needs to be simple it just needs to look simple there are so many invisible compromises that are like when you see the characters it's just like it's a sphere like it's just simple bucket shapes but there are many breaks and changes in the shape in three dimensions to make the character as appealing as possible from all angles and appeal is really needs to be the deciding factor in this so as we went along we started testing the character sculpts even more making like temporary retopologies and sculpting on top of it trying out all kinds of expressions and post tests making entire sheets these were mostly based on actual the actual storyboards of the movie or any sort of expression or moment from the movie that we needed to see if it works with the style that we have if it's appealing so we collected all of this information of what we're aiming for and created a style guides and this really encompassed all of the takeaways for how these characters are supposed to look and emote and deform and that got passed on always and refined further to giving it to the next artist like for example taking it further into retopology Angela Jeanette really did a fantastic job to create a topology that suits this crazy cartoony deformation and without some first expression tests we maybe would have missed that and it would have lead led to some less than ideal animation later on so but that's basically my part of it I'm gonna hand it over to Vivian so yeah I did visual development on sprite rights but I came up on production I came on board a little later for the pre-production and everyone else when it was already going at full force and the characters were looking roughly like this and you can see that Ellie for example she's not looking quite like herself yet she's got these pointy elbows and knees and she actually was called Emily at that time and I was tasked with making the characters sit better in the environment and you know fit better with each other and I think by by the time we made this concept art we were quite successful with that and yeah how did I manage and what was it like to work on this production yeah he said that making a movie is like laying the tracks while you are already riding the train well from my point of view it was more like well there's five speeding trains and I am meant to decorate them and by the napkins and curtains that I'm choosing they somehow magically end up crashing into the same beautiful sunset and yeah so it means that in a way nobody waits around for concept art to be done to continue working everyone is constantly making stuff so for me it was really important to squint very hard keep this white gaze approach and but also at the same time somehow be very detail oriented so what really helped was that we already by the point I joined had this emerging gold standard I think Rex and Phil were always working everyone always had the hots for Rex and Phil and we kind of could you know just look at those and kind of derive from them what does the tree look like what does a bird look like and yeah like that's something your local concept artist can probably do for you and I tried to with every bigger concept painting in form as many departments as possible at the same time so we got something to talk about and nitpick and this one for example as you saw what's recreated in 3D and I think it overshot the style by a little bit but it gave us a lot of information but also you really really don't want to take too long with concept art because when I was still shading the little bark detail Andy was done with the tree it was already approved by the director so yeah you probably as a 2D artist want to focus on stuff that is just quicker and 2D than in 3D whatever you can do there which for example here is costume variations like for some time I get really obsessed with her little miss piggy feet and I designed all these 80s shoes for her and that's just kind of quick and I also tried to help out with effects because really needed to know what what would it look like to have these melting sprites and maybe we would have animated fire maybe not and I also did tons and tons of draw over as like hundreds and paint over us but I mostly ended up chonkifying everything because that just you know that worked and it's I think it's quite obvious but the bigger the sample size from your universe becomes the easier it becomes to just generate designs for your designer and you know once we had the human characters done it became way more streamlined to do the animals and we could go with a way more default process as in you know you present options to your director they turn down your favorites and then you move on or you end up crashing into your own little sunset with bird age which you know I'm willing to do that but and then you know by the time we got to the props it was even easier because we had you know a lot of the other props already actually and you know that only took two days because you know we know what the world looks like and I got a shit ton of wonderful 80s references kind of easy something that was way less easy is making the color script because the movie was still constantly changing and as you can see here these bunnies at the very top they didn't make it into the movie the ending there that was cut what is in the color script here and but it really helps to just paint your characters over and over very small in different like circumstances kind of helps you to boil down the color identity of your movie and I think this is almost the Lego test where you check like which kind of colors are really relevant for your character and then you try to remove a lot of the other stuff that is just busy and what it also helped identify is this green on green issue where you know green characters in the very green environment and you kind of try to come up with solutions for that and then you end up making the ground brown and you're done like magic right it just works and yeah that's when you hopefully ended up voicing all your concerns all your ideas and you end up with something like this as a concept artist and then you can just watch from the sidelines as the movie gets done right hey guys I'm Demeter and I rigged all the characters on sprite fright and the way that I did that well the first we had to consider what's going to be the tricky thing about this project and that's was that there's quite a few characters so there's of course the five teenagers but then there's also other sprite and the sprite and on the side there's the much of forest creatures so it's a lot of characters in a limited time and we wanted to keep the rigs consistent at least between the teenagers since they are all humanoid and quite similar but even the sprites and consistency is important so that's you know animators don't have to relearn seven different rigs obviously so the solution that we came up with kind of before this project but mostly for this project was procedural rigs which of course many studios use at this point and it's kind of the standard but for the Blender studio this was relatively new at the time and so the workflow for what I call procedural rigging you know in the future hopefully one day maybe soon it's going to be a node-based thing but for now this is a just a python add-on so many of you might be familiar well those of you who are familiar with rigging might be familiar with rigify and we basically use rigify but extended it with our own features and that set of features was in the end called cloud rig and so the way this works is you create something called a metric which is just a simple skeleton that kind of defines the characters proportions you know like where are their elbows and where's their spine which is pretty simple and then additionally on top of that the metric also contains information about it's like parameters and values about what kind of control rig should be generated based on that metric so and the actual behavior of that is implemented in python and you just click generate rig and you get yourself a control rig as you can see and so the tricky part is just implementing the code is one thing and then getting the right parameters that your animators need on the metric and they just generate the rig and that's it and then an important tech that we used for the facial deformations of the characters was the shrink wrap modifier which basically just snaps one surface to another and so what I'm not showing here is there's a hidden helper mesh for wrecks in this case but also for every other character and that helper mesh is only being deformed by the head bone and the jaw and nothing else so it's not being deformed by the mouth instead the mouth the actual mouth of the character is of course being deformed as you can see and then after that that gets snapped onto that helper mesh and this helps preserve the silhouette and keep it clean even though the bones and the weighting and the deformations might not initially be perfect and so besides that let's talk a bit about the forest creatures because I think they're kind of fun so this is Mr. Snail my favorite character in the movie and his rigging was actually quite simple we just use the lettuce for for the cheek puff which I can highly recommend oh man an apple ID do not allow go away oh boy I don't know if you guys saw that maybe not so yeah that was the snail it was pretty simple in the end the animators of course wanted to do all kinds of squashy things on them so they also have like a giant lettuce that encompasses the entire snail to squash it and then next up an interesting one was the spider I believe most of the shots of the spider were animated by Monica Eggers if you've met her then you know she's freaking amazing and so she did a fantastic job and the tricky thing about a spider was this zigzaggy legs that had to be working in IK and also had to straighten out and get squashed and in some shots and it was amazing to have a procedural rigging workflow for the spider because it had eight legs and those legs were a bit tricky and they did need quite a bit of iteration to to get right and so I was very happy that I didn't have to do that iteration eight times over every single time that we would make a change I just made the changes in the code and then I regenerated the rig and all of the legs all of the legs were updated with the new behavior that the animator asked for and then there was the bird and this one was tricky in a completely different way in that it's of course covered in a bunch of feathers which was very annoying but because all of those feathers there's I couldn't really come up with any magical way to not make them constantly clip into each other so we just kind of did it the brute force way and yeah we just set up some shape keys and action constraints to try to fold the wing down with as little clipping as possible and then there was a lot of iteration on that because of course people kept running into imposes where it would still clip but yeah we did our best and I think even if there is clipping you can't really see it in the movie and I also made a special rig type just for each individual feather so that they can be controlled in a nice way yeah and okay and then finally the final interesting thing that we kind of started doing in sprite fright was rigged hair particles it was always possible to sort of rig hair in blender but the way you would have to do that well the only way that you could do it before as far as we knew was that the entire hair strand would just follow the surface that it's growing out of which is not great for long hair so Sergey ended up packing us a quick blender patch and some Python scripts that allowed us to create a sort of binding mesh for the hair and then we could deform that binding mesh and the particles would would follow it and so that's what we used for all the characters to deform their particle hairs and that is it for me and next up is my third favorite animator okay I'm gonna talk a bit about animation and how the animation style came to be when we were looking at the initial designs the first question as animators we asked is like how do we marry the animation style with the actual characters and it's not only about the character design like if you see Phil on the very right side you don't see it's Phil but it's a very different shape than for example Rex or Ellie so to kind of like support that in terms of weight but also in terms of character personality which for example Ellie is a very insecure girl goes to this arc of being a hero in the end and Rex is this macho guy so next is also the story like the story has different pacing we go from a lot of dialogue to action scenes and how do these characters behave in those different environments different scenes now you can talk endlessly about animation but for me once the one rig was ready somewhat ready we I this is my first animation test I did with a somewhat working rig and a lot of things could be like one of the great things about this is that you can start the conversation but it also pops up more questions like do we want smears we want to animate on ones on twos and the same goes for all of the sprites like how do these how do we keep these sprites jolly and then ultimately turn them into evil monsters one of the big decisions that we had to make is animation on twos what does that mean one second of traditional film is 24 frames 24 drawings and we basically decided to hold every frame for two frames so we end up with 12 drawings per second so that's that's called animation on twos now here is a comparison where we animated do we have sound on here okay oh okay so the left one was animation on ones and the right one was animation on twos and we decided to animate on twos a bunch of reasons one of them is to really emphasize the toy feeling of it and really go towards the stop motion style of animation but also you know the budgetary reason of not having to animate 24 frames but only 12 and also not have to care about any motion blur now that while still having a 24 frame movie 24 frames per second it came with some some issues that popped up mainly with camera motion because the camera motion is moving on wait this is playing insect so the camera motion is moving on once while the animation is moving on to so you get a strobing effect does it play come on and so what we did is basically it's a little clip so what we did on camera motion is animate the root of the character the root bone along with the camera but the local animation the arms and the legs and made animation on twos to keep that style consistent but not make it too jarring for for fast action so that's something we did here you see the characters moving out of frame on once while the animation the character animation itself within is on twos now with here we go with one sec with the animation on twos we lost the motion blur so we had to you know implement all traditional ways of emulating motion blur by adding smear frames distorting the meshes and using crease pencil to enhance or like draw extra extra lines and and these smears and multiples you couldn't apply everywhere it would be a big mess but really had to be earned per action and another thing we applied let's see here we go is adding the grease pencil and converting them to geometry for effects like the bird poop for gooey melting sprites and that was a great way to to have full control over the simulation so to speak but also a quick way to iterate over shots over the effects so it was for the director you don't have to wait for endless endless simulation time but really could draw in some some quick sketches and then later refine that for yeah well whenever it was approved so it was a great way to also stylize the effects and I think they look great in the in the movie and fit well so next will be Pablo and you will be talking about some facial lip sync animation hello I'm Pablo and basically gonna continue a bit with what Rick was saying about how we decide to stylize everything but this time on the facial so and Julian started to make some of the sculpts to try to figure out the simplicity on the characters and this was the really help us in the moment we had a rig to start to push the poses and see if we could achieve something like that with the technology we had and this process was a really long process between all the departments and so into the director we end up having like a like a group of of expressions that we wanted to recreate with the rig and we were thinking that they were precisely the the expressions that they were gonna be through the movie and the characters so we have this this already animated so it was really really helpful to have the scopes and see if we could recreate them with the rig in motion and they were gonna be holding up or not and we did this with all the characters so yeah we tried to also like really go it's not playing no yes so we tried to push all the all the characters and really aim for the simplicity and aiming for the mouth almost being like to the mouth so that was how we really push everything and really like even like making them working for the view for the camera view but in the moment you were turning the character it was exploding so it was like really like yeah just working for that frame in that camera moment so I'm gonna also all these all these poses later we didn't have to recreate them all the time so we did use the post library this was really helpful to keep consistency during all the production and with all the animators we had so yeah all these expressions they were great and it was a preproduction worldwide that we really used through all the production I'm gonna guide you a bit through my lip sync process at how I do it not the best but let's see and I'm gonna play the clip first this is a shot I like to start just putting the key poses on the body telling the action and telling what the character is gonna do and then I usually do facial pass with a strong sounds or strong emotions that they are going to be close to that poses so like this I have some reference of the characters is gonna be angry she's gonna have the mouth open mouth close so from then I put more information more poses on the character when I'm comfortable with something that I have and I think that is kind of representing already the full shot I saw it to my supervisors and then we have notes in this case in this project in this project it was mostly pushing mouth poses bigger most of the times because we wanted to go really cartoony so it was push it push it push it I continue this process of adding more information refining the body and refining the facial and we go go and on and on and on and on with the retakes with the with the with the notes so like pushing pushing pushing pushing until we end up we don't have more time to finish the shot I mean it's like the deadline of the shot and then you have something like this but yeah this is the last already everything Polish has moustache well everything is animated so that's what they end up being on the movie that's my part hi guys I'm Bo I'm 3d artist on sprite fight I will talk a bit more about the practical side of how we approach lighting so so planning really helped out in most of the cases for the sequences for instance when the students here set camp and he would create an initial lighting pass of the establishing shot and the plan is to copy and paste and refine it this particular shot to the other ones so this creates a better workflow basically and the values are mostly consistent so once we have an entire lighting pass we'll throw it into kids who so here we can also for animation but in this case for lighting we can just throw it in there and we can compare them we can compare it to we compare to the values and it has an entations which really nice and this way we mostly iterate it on the lighting so but before we can start lighting we have to of course build it so quickly go into how this is done so we use the shop builder tool basically combining the animation in the set and we added a predefined sky which has like this basically like a ramp so no HDRIs and this way we created a pin system like global emulation well it's not real global emulation but yeah color and if there are any gaps we will fill them with bushes and trees and basically start adding key lights and shadow casters as we call them or light blockers so the lights in sprite fright these are the lights we used have their own attributes and purposes so the key light especially for this particular sequence was to mimic the warm sun when they were setting up the camp yeah and then rim lights we placed them behind the characters to separate them from the background the fill lights we placed above the characters usually to create more dimensionality to also fake a bit of the global illumination on them as well and to soften the characters faces we would add soft fill lights on the dark side of the characters faces to basically create better appeal and then yeah as I mentioned the shadow casters or light blockers so light linking at this stage we didn't really use so we use meshes and the nice thing about it as well is that we could also like block glossy rays so if there were objects that were too shiny we could also for instance turn off the diffuse and only have for instance well it depends on the effect so we have more control for the final image finally we add atmospheric effects basically an emissary shader with a 2D texture that emits lights to fake light shafts basically so for me this really created a really nice structured way to work from so trying to maintain within the style and a vision of the movie and using these guidelines the way we use the lights how we place them the predefined presets and stuff like that really helped this is another thing it's just a contact sheet I used it a couple of times to compare the values in a sequence so because okay well this shot needs a bit more key light for instance created by Paul yes over there so anyway but besides maintaining consistency is not always the goal to have that sometimes you really want to do more cinematic based lighting which where you can break the consistency and continuity to improve to improve the story and the character dialogue for instance so if you take a look here at these shots especially near elder sprite we see a shadow there and somewhat different lighting direction when compared to this one now you don't really necessarily notice it when you watch it but because you want to focus on the story it really emphasizes it so yeah this way it really depends on the scene and what you value yeah the talking about values the master of values see more that's it I'm Simon we're already running out of time a little bit so I'm going to try and speed through I was doing shading and some of the effects on sprite right I'm gonna talk a little bit about the shading principles so once we had the style guide set and knew what the movie supposed to look like we had this like toy like feeling and want to chunkify all the details with the style guide for the shading principles so one of those aspects is that we wanted to have the surfaces have really large enlarge in patterns so they're larger than life to make it a little bit more small scale and tar like now thing was chunky details to also not distract with a lot of small noisy detail and then keep the specularity overall pretty low for the appeal also and alongside that also make the mental values quite rough overall and then also keep the bump tater the bump detail quite low and profile and within the surfaces have a relatively low contrast we have these strong and clearly defined colors that we don't deviate too much from within a surface and then also what we've been was working on defining the palette in a way that it's clear and we don't deviate from that too much and then for the character shading specifically something that we wanted to focus on is not make it too realistic in the sense of actually skin but still make it flashy and alive so it's very stylized we don't have don't didn't want to have any bump but have like a slight marbling effect just to make it a little bit more appealing and a subtle SSS to make the characters come alive but not too realistic in the sense of actually creating skin and in terms of colors for the character specifically we had these very two different moods and the daytime and the nighttime and to be sure that we can hit the right style in terms of colors also we actually baked a different color palette into the shaders themselves we could switch between a night version and a day version and depending on that switch they would use different colors in the shaders directly so we didn't have to rely on the color grading afterwards to make that mood hit but we could also just do it in the beginning of the pipeline an important aspect of the shading especially for the story was the character arc that the main character was going through from in terms of the distression where she's clean in the beginning and then yeah bruised and everything in the end and that was really important on the story level for the main character but also for all the other characters it was pretty important to marry them into the environment and make sure that they work together with the with the world and the story and for that the process was that Vivian was making drawers for all the characters and I tried to implement all of those different details into the shaders directly and yeah there were lots of smaller things we did here in their scratches running mascara dirt and ripped clothing and for the main character that was a little bit more tricky because it needed to be several different stages throughout the movie to make sure that every single stage every every interaction with the environment is reflected in the in the distression of the character and for that we built it into the shaders it's a as well with a bunch of parameters where we could actually change the sliders around for every single shot to make sure it's all consistent and yeah it's the right frames and then a little bit about the props and the shading in general so I there were a lot of props to be shaded so I like to also in general rely on a more procedural workflow or create a bunch of base shaders that and can just copy around then and then very easily paint in a bunch of masks and also for the for the detail and the patterns also rely on a procedure workflow where you can get quick iterations make changes here and there at every point in time the process basically that was really helpful and also combining that with effects that we would use something like geometry notes for which was the first project we had the opportunity to use geometry notes and we did that a lot in the environment and then also the shaders could be informed by for example here the moss that we're scattering around and speaking about geometry notes they're all over the place in the environment so that all the leaves scattering the grass the moss some of the clouds it's all using geometry notes which at the time it was a very different system so it was a little bit tricky here and there but that was really really useful now than that about the environment it was mostly just about making sure that works with the characters that the characters fit into the world but doesn't overwhelm everything and make sure that the characters can really shine and then a little bit about the effects so for sprite right a sprite fight we were using various different methods for the effects one thing was simulation that was what Andy was in charge of pretty much we had a bunch of simulation here and there and then another thing would be animated props where we just have a have a rig on a prop and then the animators would take care of actually making sure that looks like a nice effect another thing that's something that Rick also mentioned we would actually frame by frame animate some things like with grease pencil the the the splashes from the sprites for example and then turn that into a mesh that we can then render with the lighting of the scene and then last Lee I think the geometry notes effects so I tried to wherever I could actually incorporate geometry notes for like for example to create a prop we had the hairspray effects with a bunch of sliders and then paste them into the shots and really easily adjust them without having to actually manually do anything on top of that then a lot of the melting effects and like also for example the snails scattered around on the body of Rex I want to talk a little bit about the melting effects in more detail so here's how it looks in the actual file so it's all a system in a single effects file that we then link it to the lighting file and there's various different things going on to figure out how that's supposed to look also Vivian made a drawing to figure out the different aspects of it to still also make it appealing and then we just looked at the different elements so it was supposed to expand from the salt and carve into the material droop down have an actual change in the color of the material to be very very visible create some bubbles and a little bit of smoke and then I tried to incorporate all these individual things into the geometry node system and here are the different layers of how that looks like this is the base animation that I got from the animators and then these are the different elements that are layered on top with an iterative process and that's the final render and that is it and I think that was okay before we wrap up of course please give another round of applause to the blender developers because they have done amazing work to make this possible okay thank you for coming one final note today at 5 p.m. in the special interest group room there is a blender studio ask us anything so if you have any question you can ask us anything over there so see you later