Lightning Talks — BCON26

All right, good evening everyone. Glad to see so many faces here in the theater on the second day of the conference. My name is Thomas Dingis. I'm development coordinator at Blender and I have the pleasure again uh to host the lightning talks this evening. We have 14 speakers here presenting their work and how they use Blender for up to five minutes. Um, for the people presenting how it works, I'll just announce your name. You can come to stage. You get a clicker for your presentation and a handheld microphone and then it's your five minutes of fame on stage. Um, you will have a small uh timer on the screens down there. So, please make sure you stay within the five minutes uh because we have 14 speakers and we don't have uh too much time. So, around about an hour. So, please make sure to stay within those five minutes. All right, I hope you guys are uh ready for some interesting talks. So without further ado, let's start with Timothy and the first presentation. >> Hi everyone. Okay, five minutes. All right, so I'm Tim um representing China's Blender community, Origin CG. Uh a few updates again uh from our side of the world. All right. So, first up, is everyone up for a challenge? Say yes. Yes. Good. Good enthusiasm. So, we have been running a contest for three years in a row now. For the past two years, it's been mostly for uh primary and high school students. So, this year we're opening it up to individuals, professional studios. So, it's imagination without limits starting from a cube. So, if you can make something with a cube and do a 30-cond animation with it, you might win a cash prize at the end of it. and get featured at the BCON in China in Shenzhen. So, what are you waiting for? Um, there are some pictures to show what last year's uh event looks like. Uh, these were some of our winners from last year and look at how shocked our judges look uh at the top up there like at the quality of the work done. So, yeah, you get to be part of this. All right. So, uh second thing I want to talk about Blender certificates. This is an ongoing project that we've been sharing uh across the conference. Uh we'll be talking about this in depth at the educator SIG tomorrow morning. So please join that if you're interested. It started with blender badges you see on the left. Uh defining different domains and a standard for designing curriculum for educators and teachers and then in Hong Kong and China we've been doing pilots for certificates giving them out through university partners. Um and we've actually issued our first batch. Uh you can track everything on our website here. Our first batch was given out through our university partner, the Metropolitan University of Hong Kong. So these students who uh study with us uh and with the university can then get a certificate that's recognized and built on a standard that's communitydriven and we welcome everyone to be part of this process and you can help us bring this to your region as well. Right. So last but not least, timer is running out. Can I get some sound, please? All right. So, the first BCON in China happened last year. We had 60 keynotes, 50 speakers, and over 3,000 participants join us over a long weekend. And it's happening again in November 28 to 29 this year. It's part of a bigger series of events, and we have our first speakers uh officially being launched here today. It's the first reveal. Uh we have a friend coming out of Pixar who's been there for 20 years, lots of experience uh from ILM. We have Andrew Price, the Blender guru, joining us in Asia. Uh we also have friends from Blender, of course, the Blender team and from Autoro. So, lots of new faces and familiar faces joining us um from across the world and from the Chinese community as well. I think this is the most important. We're hoping to showcase more creators and artists locally in the region. If you know anyone who would like to speak, who would like to attend, who'd like to join us, or if you'd like to join us, please sign up. I'll give everyone 10 seconds to take a picture of all these QR codes and then it's going away because I need to give up one minute for a special surprise speaker. All right. Thank you. Hi. Thank you. Yeah, I just kind of hijacked his uh presentation just to tell you about the Blender codes. It's a little game that has been running on these uh days and still on till tomorrow and afterwards. So, it's a treasure hunt inside of Blend Files. Just uh scan the QR code and uh start playing and let's see if you can get to level 20. I think people are around level 10 now. So, just catch them up and thank you. All right. Um, I was about to say a few words later how to apply for those lightning lightning talks, but you just seen one way. You just sneak on stage, ask another speaker. So, that's cool. No, I'm just kidding. Olivier is always welcome here. >> Thank you. Thank you. >> All right. So, next up we have uh Jovanni. So, here is your microphone and the presenter. The stage is yours. Okay. >> Thank you. Good evening everybody. My name is Jovan Gallalo. I am a professor at Katana University in Sicily, Italy. And uh I'm going to show you introduce you some people, some person and a project. And uh what you see now on the screen is Mariah and she's the uh main characters in a short movie that has been produced by academia devel jointly by a team of students uh during the last year. We have a long tradition of doing this, working together art students with computer science students to produce movies, short movies. And of course, it's a uh didactical uh project. It's not uh something that uh too professional, but uh we believe it's it's a good produ production. And uh you have a QR uh code there so you can go and look at it on YouTube. Uh why we did this project this year? Well, in 1926 Theopor published the novel Metropolis. She was the wife of Fritz Lang and uh the year after on 1927 Fritz Lang produced Metropolis. I think everybody knows about this movie and uh we decided to celebrate the anniversary in this year uh because most of our team is made of girls and they wanted to recognize the talent behind Metropolis because seldom she is mentioned. So there's been a long work at the academia at school. Uh the students have been looking metropolis. They have been reading the book. They have been discussing a lot uh in order to produce a subject and uh uh go for the movie. And the idea is a reflection about relation within reality and virtuality within artificial intelligence and human intelligence. And uh you see here just two frames of the movie and uh I'm not going to spoil you the content of the movie. It's short. You can look at it by yourself. And uh I would like to show you some picture that we took during the production in a room of uh the academia and uh the people that work on this. A lot of work was done on site at school and a lot of work has been done online and uh I think this has been a valuable experience for students uh to learn how to work on site and online. The team was composed of for from teacher, four teachers, two computer science senior students who have been the technical director of the movie and 20 art students, junior and senior that did everything from screenplay, concepting, um painting, modeling, animation and so on and so on. And here are just some files from the production. And uh uh other than Marai which is the main character I told you there are several other characters in the movie and they interact in this futuristic environment and uh there is a surprise at the end of the movie and uh I think I've said it all and uh the final message is from our friend Marai come and see me performing and there is the code. Thank you. >> That's fine. Thank you so much. >> All right. And next up we have Frank Okay. Hello, Blender community. Uh, I'm Frank Russo, CG Wire CEO, and I'm going to introduce you to Kitsu, an open source production tracker. So, uh, first, what is a production tracker? Basically, it's a software that allows to grab data among an animation studio and to build report on top of it. And it has always been designed as a nouning software. And when we started the kitsu, we noticed that this software was mostly used in studios for communication and was not the initial purpose of the software. So when we created Kitsu, we we decided to make it first easy to use. So anyone in the studio can collaborate through it and communicate efficiently uh through it. And the manager uh to say it shortly, it's artist and supervisor friendly and manager are still happy because they have up-to-date data. So they can uh anticipate problems and report uh as they want. uh but because we like collaboration we wanted to push it at every level of the software. So we made it open source and we choose uh a strong license. We choose the GPL license and we choose only uh technologies that are well known by by the community. Uh if you want to contribute the the software is coded with Python and JavaScript. Uh so and uh and it's based on the Flask framework on the back end side. So it's pretty uh easy to contribute and if you want to set it up on your own machine we relied on battle ardent technologies and bricks that are uh really uh common to set up. So uh so there is a posgrsql and radius and there is a little exception with med search but it's an indexer engine uh indexing engine and it's uh but it's pretty uh it's a pretty good technology and it's optional and so uh why uh collaboration matters so much to us because we noticed that when people communicate efficiently inside the studio when they work together when they stay aligned We end with a better movie. Uh most of the time ship on time and uh and in the end when we all work together we are all winners. We we all go to a we all end with a good movie that we are satisfied with. And even better, it had another positive consequence is that we noticed that studios not only collaborate uh inside the studio with Kitsu, but they started to collaborate uh with uh other studios and that way they were able to uh to achieve their vision. Uh they were able directors were able to go to the to to reach the goal they wanted with their movies. Uh and that's and that with even with small budget. So we ended with many uh beautiful movies or productions or TV series were done with Kitsu. Uh here are a few uh I selected only uh the only movies who were done with Blender also and we we ended with u more more uh a bigger diversity of narratives uh because thanks to Kitsu uh small teams can uh can build very cool movies and can make very cool movies and so it's uh it creates more different narrative than the usual Disney ones and so it's Uh yeah, it's a pretty cool achievement and on this slide you can see that we have very different styles. Uh it can be TV series for children's, it can be political movies and and also it can be fun short films made by Blender Studios. Uh all that uh that's all for Kitsu. just want to share you a few dates uh we have uh for u for for for our own events because uh we want uh we also have a community and we we are happy to gather this community uh in the same uh place. So the next one we we already organized a kitsu summit uh this year uh at the beginning of the years the next one will be online because we are a small team so we can we can't organize events like this every years but in two years we will meet again in Paris uh in February so you are more than welcome uh we will talk about production pipelines uh production workflows and there there are always a lot of blender users uh at the kitsu summit So it will be very interesting uh and you will be uh able to talk about the future of creative collaborations. Uh that's it for my talk. Uh if you want more informations you can uh connect to the website and see uh exactly what does kitsu and you can also contact us directly and uh we will be glad to uh talk you more about our software. Uh thank you Blender community for your attention. All right, so next up we have uh Sam. >> Oh my god, it's Sam. Hi Hey. Okay. So, uh hi, I'm Sam. Uh I'm a 3D artist at Tomorrowland, but uh in my spare time, I um create short films, uh animated short films in Blender, obviously. Um this is u my first animated short film that I'm actually on track of actually finishing. And it's an uh obviously a pirate adventure called Captain Redbeard and the Kraken. Um, I started the project uh like nine years ago, which sounds like a long time ago, but um, this goes really slowly and I've only recently started enlisting the help of like these really talented artists who can actually help me with character design, character modeling, uh, animation, stuff like that. Um, and then Blender 5.2 D came out with its long-awaited hair dynamics and it turns out to be a node group um inside of a node group and then you look for the solver and it's oh it's over there but it's actually another node group and you find the actual solver it says experimental and you don't know a word that it mean that that it's saying and um when you apply it to your hair then this happens and I'm like the vendor conference is in two days I don't have time for this so uh you go back to mesh hair and the bump map and it doesn't deform and I'm sorry but I will fix it at some point. Um yeah, this is the first sketch I ever did for it in 2017. Um starting like another project uh I I have many but like um like I said just getting the help of these talented artists actually make these things look good. Um great animation by Vincent. Shout out to Vincent. Uh don't steal him. I'm so glad this guy doesn't have hair to be honest. Uh that's less work for me. uh still debating on whether or not to give him clothes in the in the final film because I kind of like the skeleton look. Um gives it like this this eerie piratey vibe. Um he also lives in the underworld which is like a very gloomy uh sort of contrast to the to the warm lighting of the rest of the short film. Um but the most complex thing I've been working on uh for these past few weeks is this town. And I have like a weird little thing with um like vertical cities like metropolis kind of style things that you see in cyberpunk genres, but I wanted to make it like makeshift and build out of shipwreck parts. Um so yeah, I've been cracking my skull on how to build these things, how to start building these things. But like with any art, you just start with the big shapes. And I've added a small layer of like a wooden planks just um to give it some detail to build your detail on top of of um it's just a node network like it's not even the most complex thing but um you can just draw simple curves and they will all turn into blank instances. Um and everything else is just like intersecting boxes uh seeing where the design takes you where where it really wants to go. Um and then adding like housing modules. There's like six or seven different ones. Um, and it will feel like a city soon enough. Like if when you add uh pathways, walkways, um things that make characters actually move through this uh this whole thing, some vegetation. Um and everything else is just really decoration like windows, doors, props, some some hanging laundry. I just that that people actually live here. Um that that sort of will sell it as a city at some point. Um this is only 10% of the whole set. Uh, so I may be like over trying to overachieve here. Um, but I've wanted to build something like this forever. And only a few weeks ago, I I've finally it finally clicked like, hey, this is how I need to start. And uh, yeah, that's most of what I wanted to show to you uh, today. Um, this past summer, we did a Kickstarter for the project, which tanked horribly because I'm very bad at social media. So, people have been telling me like, "Up your social media game and talk about your project." and I'm like, it's not finished. I can't talk about my project. So, that's what I'm doing right here this week at the Blender conference. And it's it's very like stressful and everything, but if you if you want to stay updated on the project, um join the mailing list. There's the QR code. Um follow me on Instagram, follow me on on YouTube. Um if you want to stay updated. Thank you. All right, thank you so much, Sam. And next up, we have Doris. So, thank you. And hello. Um, so my name is Doris. I'm a professional lighting and compositing artist. However, I'm a not so professional costume maker as well. And therefore, I keep running into usages of Blender that are not quite as intended. And today, I would like to talk about my current project, which is building a costume of Nyx from Hades. Now, I love the game. I love the style and I would like to preserve the style of the game as much in the costume as well. And this is a big challenge because it is matte flat. There's very big flat colors right next to each other. Lots of detail, pitch black color. And I was wondering if there is a surface material that I could use both on the hard surface as well as on the flowy fabric parts. And I think I found something that works for me. And it's basically making fabric into a reverse UV map. I've done this before by painting fabric pattern uh like pattern pieces before assembling the garment and that was quite successful. However, with the style of Hades, I cannot quite do this. So, I did some tests and I came to the conclusion I want to use digital fabric printing. So, how does this work out? I start with a photogometry scan of both the model and myself as a reference for scale test. Um some base modeling. I think everybody knows what's happening here. And then I have to slice it up for at least distortion possible. Um and the paper pattern tool is helping me keep this on track. Um and the topology looks terrible. Now when I had it cut up into its islands, I had the problem of how big do I scale it because it's very abstract units. In UV space, we have 0 to one. In pixels, we have 0 to however big your texture is. And then in centimeters, we are printing from centime 3 to 141 or whatever the printable area is. And for centimeter to pixels, we have DPI, which is straightforward. I used 300 dpi to get a crisp texture from pixel to UV space was a bit trickier because that is tixel and Blender doesn't natively display the text density. So how many pixels your uh island is taking up but there is plenty of add-ons that can do that. Here I'm using text tools. And once I've figured out how many pixels per unit I want to have, I can consistently scale all of my islands. And with that done, I can build a UV layout, a little bit of spacing in between because I will have to wrap the fabric around edges. I do some texture painting to mark out all of the important parts of the design. And then I take the texture into a vector program because it makes it a lot easier to clean up the textures, the edges and have consistent colors with the color palette which is another problem that came up because we have Blender with AGX and Filmake few transform in RGB color space and I need CMYK fogra 39 coated proof. I didn't know that. I was like, "Oh, well, there's an easy solution to that, and that is just buying fabric swatches and then writing down the CMYK numbers." And that worked perfectly fine. And this is how this texture looks. It is 140 by about 3 m big. It is about half of the costume. And while I send this off to the digital printing company, I do a little bit more 3D modeling to get the base models that I will later wrap in fabric. I will also have most of these 3D objects printed in TPU. And I will use craft foam, very straightforward to do the flat surfaces. And for the craft foam, I already have p paper patterns because I did a UV layout with the paper pattern extension. And this is how the process looks like. Um, so essentially the fabric is applied to the different objects. And here you can see me do one piece of that. It is a layer of craft foam, thermal adhesive, the printed fabric, and then I do the black details with um velvet because it captures most of the light and some ironon vinyl details for the neon splotches. And here we are right now. Uh quick conclusion, it was not the most efficient way of doing this. Um but there's a lot of potential I think with the new updates in Blender and also maybe using this on paper on rigid models. And if you're curious, I brought a few of the pieces and you can inspect them up close if you would like. Yeah, thank you so much. And next up, we have Jacob. >> Thank you. Thank you. Oh, how do I get this permission? >> Is that okay? Was I supposed to do something different? We'll see. >> Okay, we'll find out. Uh, my name is Jacob. I use Blender to make VFX. And today I wanted to talk about uh VFX lens distortion workflow. And for those of you that are not familiar with the standard workflow, I made a full tutorial and today I'll just play the intro to that tutorial because it'll describe things better than I can live. Uh so here we go. In the world around us, straight lines appear straight, except for when you film with a camera lens that makes straight lines appear curved or distorted. This is called lens distortion. Now, as a VFX artist, if I want to put myself in this CGI environment, lens distortion poses a unique challenge. That's because while a real camera lens produces images which have distortion, a virtual camera lens renders images which have no distortion. If we try to put these two together, things won't align properly. So, the big question is how do we merge this imperfect distorted footage with this perfectly undistorted CGI world? Well, first we start with the footage and we remove the lens distortion. So now both the footage and the CGI environment have no distortion which means we can properly merge the two together. And finally we reapply the lens distortion so that this final shot has the same distortion as the original footage. So the workflow is undistort do the VFX work redistort. Simple to understand but a bit more difficult in practice. So today we'll dive into this workflow and I'll show you a custom Blender tool I made to help you handle lens distortion the right way. Yeah, there we go. So, thank you. You can check out the full tutorial on my YouTube channel. We're not going to dive into the workflow today. Um, but I'll just uh highlight the plugin I made which is called Undistort. Uh, you can find it for free at this link. And basically the way it works is it's a plugin for Blender. So, you bring in this image. This is uh an image I captured with a fisheye lens pointed at the checkerboard grid and the plugin detects the corners. This is all working in the camera tracker. So these are just tracking markers. And then the plugin uses that data to solve and remove the lens distortion. And this plugin also supports um creating ST maps and handling overcan. So here we are into the compositor. Here's the original image. Here's the image with the distortion removed, including the overscan area. And then here's the image redistorted. So, this would be the same as the original. And if you're not familiar with ST maps or what overscan is, again, that's all described in the full tutorial. Uh, finally, one other neat feature or oh no. Um, maybe that's not going to show. Okay. Anyway, uh another neat feature of the plugin is that you can calibrate instead of with a single image, you can use multiple images. Um and that solves the distortion much more accurately. Uh so maybe this is good not to show you. It's a little teaser to go watch the full tutorial. Um so again, if you want to check out the plugin, it is free. There's the link and it's kind of in beta. So if you have feedback, please email me at that address. Thank you very much. Thanks. So, next we have Christopher. >> Um, it was an extra video, right? >> Yes. Yeah. Yeah. Okay. >> Can you load that in the background? Yeah, sure. >> Yeah, that's awesome. >> I will start speaking. >> Hi, my name is Chris. Uh, I'm a 3D VFX artist uh from Vienna. I'm also known as Hexabar. I was just on stage to present uh the how I restored the hairline for Seth Rogan over the last couple months. But I have a secret passion project which I'm really excited to show you because it's a music video I did with a live uh with a pop punk uh band in Austria where I took the band and transported them into a video game. And I did that by scanning their heads from our live action shoots and created their counterparts the digital doubles uh which the with with the mesh to metahuman workflow. Uh they also needed uh some really high quality clothing like exactly the clothing they had on set and I achieved for the first time like some really impressive results for my terms. And uh one of the main parts I stand up here is to encourage you to uh use the power of UDIMS. That's how I achieved this like leveling up of my quality of work. Pun intended. And big shout out to Matt Curtis from Bold Graphics. I don't know if he's here but I love you man. He encouraged me at BCON Austin to learn UDIMS and I have this newfound appreciation for UVs and it literally uh yeah leveled up the quality of my work and I want to use this space and this uh place now to encourage all of you to learn UDMs and maybe you will find the same appreciation for UVs I have now. So okay. All right. So, uh, you are about to see the music video and it goes live right now. And, uh, I created some more outfits for the singer. He has a journey he's going on. And I did that for all of the props, the instruments, the environments, everything you're about to see. And even the skin, I wanted him to be gnarly, to be nasty. And I even did the hair, not only for Sethroen, but also for the singer. And yeah, I even used VFX uh instead of After Effects with Blender to uh yeah, transport them in a different world. Um I used uh also a big shout out to Pure Polyjord for his awesome glow map tutorial which saved my ass countless many times now. And yeah, you can see the VFX I use here. If you like what you see, if you have questions, um please reach out to me. I'm Chris aka Hexabar. And now I'm so excited to play this video. This has been months in the making. >> Yes, ladies and gentlemen, plates of misery. >> I guess I it up this time. And I guess you thought I lost my mind. That's a place of misery. And that's where you got to find me. Come find me broken on the floor. Too much to take. But I beg for more. to those pills, but the won't stop. My time is running out. It's so frustrating. I can't think straight and my mind is racing. My life is fading. I'm a mess. That's what you're saying. I guess I it up this time. And I guess I lost my mind. That's a place of misery. And that's where you got to find me. I guess I'm gone crazy. And I guess I'm messing up lately. That's a place of misery. And that's where you got to find me. A new day means a new breakdown. My job got me stressed out again. I'm getting sick. Why can't this ever work out? It's so frustrating. I can't straight and my mind is racing. I'm so jaded. I'm a mess. That's what you're saying. I guess I it up this time. And I guess you never lost my mind. That's a place of misery. And that's where you got to find me. I guess I'm gone crazy. And I guess I'm messing up lately. There's a place of misery. And that's where you got to find me. You got to find me. You got to find me. You got to find I guess I it up this time. And I guess this time I lost my mind. That's a place of misery. And that's where you got to find me. I guess I'm going crazy. And I guess I messed it up lately. There's a place of misery. And that's where you got to find me. I guess I it up this time. And I guess I lost my mind. That's a place of misery. And that's where you got to find me. I guess I'm going crazy. And I guess I'm messing up lately. That's a place of misery. And that's where you got to find me. >> Thank you. All right. Thank you so much. And next up, we are going to have Sebastian. >> Hello. Um, my name is Sebastian Kunik. I'm from Blend of X. We are a small studio in Leipzik in Germany. And among other things, also known as the doors, uh we build and create uh interactive installations for entertainment and education for museums, castles, exhibitions, stuff like that. Um, two years ago, uh, Claudia Febach, the musiologist of Castle Milildenstein in Germany, uh, asked us if we could help with some, uh, face reconstruction and create an interactive app for an exhibition they were planning. And that exhibition is about, um, a skeleton they found like uh, in 1965 or something like that. And they uh for a long time they thought this was a a woman like a um the skeleton like a 100 years old or 500 years old woman. And so there was this legend of this dam this castle maiden um that lived there. But then uh eight years ago some um scientists or anthropologists uh had a closer look at that skeleton and turns out it was a man like always. Um uh and um so they inspected it further and did some um more science stuff on it uh to to look at the teeth and the the skull and the way the the skull was worn down and stuff like that. So to just to find out who was this guy and um so this was a very interesting process. So they decided to make an exhibition out of that. And one of the things they did was they gave the skull to um to this uh anthropologist forensic company and they uh did the uh reconstruction. So they calculated the soft tissue marker depth and all this um forensic face reconstruction and stuff and um they came up with a rendering and the musiologists asked us if we could do something about the textures and the hair and we said yeah that might be good to do. So um unfortunately we we couldn't really use the topology due to unwrapping and other things. So we decided to start from scratch and um so we got the skull from them and uh we did put in the the soft tissue markers and um recreated the the skin from scratch. So we put in some muscles and then of course the skin and then because I wanted to have hair um and test out different hairds or hairstyles uh we do we dived in into the new um hair system. the hair uh geometry node hair stuff and I had to get myself familiar with this. Um but it turns out it's really fun. So I tried to um tried out some things to just to get myself uh more uh yeah to know what I was doing. Um, and in the end we we came up with a couple of different uh of hairstyles and beard styles um because uh they wanted to the visitor to be able to to to relive this um this uh scientific process themselves. So they would go through the through the exhibition and then test out different hairstyles and uh eye colors and stuff like that. So these are some of the variations and because we had to do this for an app, we had to uh render out all possible combinations of that. So long hair, blonde, then blue eyes and you know um but I was familiar with batch rendering of different combinations. So it was quick to write a a Python script in Blender to just enable or disable all the different collections with the different hairstyles. Um and so I got like thousands of combinations out of that. And we put that then into Unity where we uh created an app for this exhibition. And now I have to speak really fast because I have just one minute left. So this is the exhibition and the visitor goes in there and um plays with an app called Antrolab and um this is the uh the interface which we created in Blender obviously of course as well. Um also we did we used Blender um as a storyboarding tool for the uh for the app. Um, so the visitor goes in and fills in the the results that he's learning in the in the exhibition and we put this in the postcard scanner which is a tool we de developed ourselves. Also, of course, the printing 3D printing was done uh created with Blender. So, this thing will then read the results of the visitors and put this into an app and I really crossing all my fingers that this will now play in the last 30 seconds. Okay, please Thomas give me one minute of overtime. We have some great music for that. Um, so this is like the the walkthrough through the exhibition. Um, people sit there uh look at the uh uh exhibits and then put in the results or the stuff they learn into the postcard and then they approach this uh station where you have the the skull and then you will start the app by putting in your postcard and that will then start uh scanning of the skull that you see on the on the right side. Uh this will be scanned in and then you can uh play around with the with the scanned uh skull. Put in the the missing uh parts of the skull. I cannot show everything here but in the end um you will have um the skull uh with a reconstruction and then you can enter uh your um values from the postcard and you will see the reconstructed face. So here you can see now this is apparently like was a like a monk with brown hair and green eyes. But if you don't like that you can change that and um see some other variations of that. And coincidentally this exhibition was opened today at 11:00. So uh I hope it runs fine there. Thank you. >> All right. Thank you so much, Sebastian. A little bit over time, but it was definitely worth it. And the music was worth it, too. You know, some classic vibes. If you're more into music, we also have live music today uh in the studio at like, I think, 7:30 or so. So, don't miss out. All right. Next we have Robin talking about USD. >> Awesome. Thank you very much. Uh so hi, this is a USD file and is the USD file is in Blender and here it is in Unity and here it is in Unreal. Here it is in Houdini. Here it is in 3ds Max. Here it is in Maya. Here it is in GDAU. >> Yeah, GDAU. Uh here it is in Da Vinci Resolve. Yes, even Da Vinci Resolve has the ability to import USD. So USD is uni universal scene description. It's an open- source file format and it's literally the exact same USD files open in every editor that I just showed. Like literally the exact same file open in each one, right? And the thing is USD feels complicated, right? But I'm here to tell you don't worry about it. Don't worry about it. Like for example, let me explain USD to you in human terms. So open USD is an API. Don't worry about it. USD is a file format. For 90% of people, you're happy. You're good. It's a file format. Don't worry about it. Uh USD uses a system of layers. Don't worry about it. Layer is a file. USD layers are just USD files. It's real simple. Uh USD layers are displayed on a stage. Don't worry about it. USD files load as an asset. It just you're loading assets. It's simple. Uh USD has composition arcs. Don't worry about it. USD has references and some references are special. It's real simple. Uh USD layers have opinions on prim. Don't worry about it. USD files have objects with properties and you can override. You can override properties. It's easy. Uh USD has composition arcs to combine layers and scene descriptions into unified stage. Don't worry about it. UsD files reference each other to load together. It's simple. Let me show you how is a USD used. Let's say we have a wheel geometry file and then we have a material file. We combine those into a file that is has a material on a geometry. Then we have for example a car file and then we add that the material to there and now we have a car with material on it and then we combine those together and we have a car. We just reference the wheel four times. And so the interesting to know here is that each box you see here is a separate USD file. It's a separate USD layer, separate USD file. Which means if I change, for example, the wheel all the way at the beginning, that changes the final result. So if I change that wheel, it changes the final result. It's non-destructive editing. That's USD. It's just it's non-destructive editing. It's real simple. So why not use glTF or FBX? People ask that a lot. Fine, because USD has non-destructive editing between references. That's it. That's really it. That's why people are using it. Now, composition composition arcs use liver PS to resolve opinions. Don't worry about it. Some references override other references. That's all you got to know. It's real simple. The USD also has different file types. USD C, USDA, and USDZ. Don't worry about it. It's a small file size. It's easy to debug or it's one zip file with all the files in it. That's the difference between them. It's real simple. Uh USD is made for movies. Don't worry about it. For example, this is folks from Embark and they gave a really really good talk at uh GDC and also at everything procedural conference about using Houdini and USD and Unreal and their entire character art pipeline is done in USD and it works great for them. Now USD isn't made to be imported and exported. That's weird, but don't worry about it because for example, usually you have like a dcc like Blender, Da Vinci, Houdini, and whatever your engine is, and usually you have like a file, whatever file it is, right? And then you export it and import it with a little bit of data loss, and then you do it again. Little bit of data loss. You do it again, a little bit of data loss. This is bad. Don't do this with USD. It's not how USD is supposed to be used. The way you use USD is you have a USD file for each dcc for different people working on different things and then you just reference them all together in whatever your final result needs to be. So you're never importing and exporting. It's references. It's non-destructive editing. There's no data loss. It's great. Now building U a USD import export in Blender is building the wrong USD workflow. This is why Houdini has its own USD environment called Solaris which is purely USD. There's no importing exporting. you're just referencing USD files. So I am not saying Blender has to build a unique USD environment unless you want to. I'm also saying not not saying you have to use USD. If you want to, great. If you don't, that's also fine. I get it. But I am saying USD is easy to understand. You know, don't worry about it. Now, if you want to learn more about USD, you can join the USD community. Biggest one being the Academy Software Foundation one. So you go to ASWF Slack. whole bunch of people talking about USD. It's a great time from all kinds of companies. Now, do you want to try out USD? Cool. Don't worry about it. Here, I I released some CC0 licensed free USD vehicle example assets. It's one megabyte large. It's really simple. This is the time to take out your your phones and take a picture of this. If you want to try use do a little dance in the meantime, Aras doing the dance for you. Yes. Thank you. Thank you. Thank you. Okay. Also, shout out to Kenny. Uh, I used some assets from him like as OBJs and I just made them into a USD. It's real simple. Uh, and that's it. Thank you very much. Thank you. All right. Thank you. Don't worry, we'll still have five presentations, so don't worry about it. We'll still have a lot of fun here. I like the spirit in this room. It's really fun. So, next we have Evan. >> And here we go. >> All right. That That is a tough act to follow. Um, you're stuck with me now for five minutes. I'm going to go fast. That's me and my brother. This is our new studio. It's a third studio we've started. The first one was called Jib Jab back in 99 when the internet was new. Thank you if you know what that is. Um the next one we used the success from that we had kids so we started a kids thing. Um we self- financed this with the money we made at Jib Jab. We licensed it to Netflix. Ultimately it won a bunch of awards and Netflix bought the studio from us. We went to work there for four years and when we got out we said what's our next project? If you want to see it you just spybros.com storybots. That's a real I don't have time to play it now. Um but this is our new project. We're making an independent animated feature film. Uh, and it's based on the music of our friend Perry Grip. Perry um has hundreds and hundreds of songs and they're all character-based and they're hysterical and they're earworms. Um, he has billions of streams. A lot of the library was written back there and you can see it just keeps growing year over year. Um, and this is uh this is me when I left Netflix. I said, I'm gonna learn Blender. So, I started teaching myself Blender. I was 49 years old and I was like, I could do this. So, I'm going to play this. This is just a sample of what I could do learning Blender with Perry's music. >> He's mixing up the batter and he's laying it down. I am a banana. I am a banana. Banana llama lama. >> Black hamster misunderstood now. Black hamster fighting for the good now. Soaring through the night in his rocket power der with his sidekick the knee on jug. >> So that gives you a sense of Harry's music. He's amazing. Um thank you. Our film is based on one of his most popular songs. is called Space Unicorn. I all my friends are way more talented than me. So like I'm brought my friends in. We're drawing. We're sculpting. We're doing all this stuff um the traditional way. But um here's just Vizdev. We've been creating. But this is what I want to talk about. So we're excited about pipelines. Um a lot of people in this room probably hate AI. And I will agree with everything that you hate about AI is really bad in a lot of ways. Our feeling is just because something is terrible doesn't mean there's not good stuff we can extract from it. So our film is going to be handdrawn and hand animated in Blender. Um here's a sample of our pipeline. This is an animatic. This is a waffle. It turns runs. And then this is what my friend Kendall animated. grease pencil. So now we're we're drawing, right? And then what is then we used Nano Banana just to help us paint. It didn't make it for us. It took us like 50 or 60 shots to get there and we had to composite it together. But once we got that, we were able to do this. We're using it like a paint bucket. It's really good at textures. This is not stealing people's art. This is an artist animating and us using AI for what it's good for. Um, here's a better example. Let's see if it plays. Last week, I even got one of those little refrigerators makes ice. I ain't going nowhere. Last week, I even got one of those little refrigerators makes ice. I ain't going nowhere. Last week, I even got one of those little refrigerators makes ice. I ain't going nowhere. >> So, we're trying. We want to bring handdrawn animators back to the art of feature film and we're putting this in Blender, right? So, here's a very simple example. How much? I have one minute. Very simple example. Background built in Blender testing different light sources, soft light, spotlights. Then it goes to an animator. An animator can draw on top of that. Then we do what we call our synth 3D pass. We'll we'll use the robots to help us paint. And then we do another pass. This is an interesting. We're actually using AI to convert the drawings frame by frame to models, not to use, but to then use as reference to say, how does that light work? And then we can bring it all back together and composite it. I have lots of ways that we think we can use these tools to empower artists. Um, we don't need it to make art. Like, we really, it doesn't. Everything starts with an artist making something. And uh and we're excited about what this can offer. Um and just here's a quick example of Oh, that I went too fast. Anyway, I think my time's up, but this is uh let's see if it'll play. Nope, that's me. I'm at zero minutes. Thank you everybody. Now I want waffles, but we have four people four more people to go. So, all right. Uh, next we have Mark talking about area lights. >> All right. Um, I hope you're all ready to do some math. Of course not. Uh, so, hi, I'm Mark. I'm one of the Eevee devs. Uh, and I spent about three months being stuck on a very annoying little problem. I figured I would like to share in the confusion with all of you. uh don't worry there will be no math you don't have to remember anything it's mostly about a lesson to take away from all of this right uh so Eevee supports area lights you know these things and cycles kind of has it easy here because it can sort of stochastically uh evaluate these things over time but EV we have to do all of this once a frame we have to resolve the entire thing we have to consider all the light coming in from that aerial light which is a lot of it we have to consider the material that's in between how it treats this light and then how that reflects back to the camera or some other point in your scene. And all of this has to happen in a single computation. No sampling, nothing. And so, uh, we have a solution for this. It's a giant hack. I'm going to explain it. You'll hate it. Uh, and for reflection and refraction, this actually looks pretty good. But I noticed something. If I look at these things from the other side, I'm getting this weird little, what do you call it? Leakage, bleeding, I don't know, uh, banding. Looks weird. And I was like, what's up with that? I'm going to try and fix that. Can't be that difficult. Uhhuh. Um, under high exposure, it looks terrible. And it turns out it's just been in EV all this time. Thanks for not reporting it, by the way. Um, and this this was my uh state and the state of Clemon, the EV lead for about 3 months. Um, and you know, it's you can see it's really bad. And so what's what's actually happening here is pretty low-level stuff. So, in order to compute these aerial lights, normally you have to do a lot of math tricks. and we have very complicated materials and we cannot really compute aerial lights in one go with them. The math for that just hasn't been solved yet. It's not even a solved problem. That's fun. For simpler diffused materials though, we can because well some mathematician in the uh 17th century effectively figured out a lot of cool work here because one trick we can do is we can say okay we have this more complicated material. Can we find some kind of way to warp it and reshape it so that under the hood it actually is just a diffuse and then we warp it right back and you wouldn't notice the difference. Turns out the math for that sort of just works. There's a lot of cool stuff that people figured out for this. And we can apply that sort of transformation to the aerial light where we say, okay, this little area light on the the the microfacet material at the left that would be actually the same if we took a diffuse material to the right and we just completely stretch the light uh instead. That is legal. That just works. Um and you get all kinds of weird stuff from that. That's uh Lumber from the 17th century uh doing all the work. Uh but there's one problem we have to take care of here. If a light is slightly below the horizon, we have to cut off the part that's not visible. Otherwise, it's just adding energy that shouldn't be visible. And already you should be maybe getting a little bit suspicious here because I was getting bleeding on the backs of objects, right? Um, so unfortunately clipping polygons below a horizon is on GPU is not very friendly. So we do a little trick. We replace the polygon with an actual sphere. Uh, not legal. And then we clip the sphere instead. And you might be saying, "Hold on, but polygons are not spheres." Well, you cannot tell the difference. We've been doing this for 5 years, and no one has noticed. Uh uh so don't worry about it. Uh but if we if we look at the the sphere and the polygon overlapping at this point, um you should be getting a little suspicious of this shape here that you're sort of getting of this intersection of the two. So it turns out clipping a sphere and clipping a clipping a polygon well um I didn't know this but they are not the same thing actually. Uh so if we take a horizon uh yeah that's where the dragons are and we we do this clipping. We cut off the part that's below the horizon. Well if they're slightly above the horizon it's actually fine. If I move it a little bit down though my polygon is clipped entirely but the sphere is well still above the horizon I I think. Uh, and if we rotate the polygon a little bit. Oh boy, now it's the other way around. That's That's no good. So, um, we could take these shapes, right? And we can just we might be able to recognize them from that little bleeding issue that's in EV, uh, that nobody reported. And, um, yeah, that turns out to actually be a pretty annoying little problem. I'm skipping so many steps on why this was honestly really hard to figure out what the source of this was. Um, and the lesson here is we actually had a filthy little hack in place to try and account for this to try and call it, but it didn't always work. It didn't catch every case. Um, the lesson I wanted to take away from this is it took me a lot of time to really understand this problem. And until I actually completely understood it, I could not actually do anything about it. In order to build a good solution to annoying, pesky little math and graphics problems, you do need to sit down and do the research sometimes. Uh, it can be fun, but maybe don't overdo it. Thanks. Thank you. >> All right. And next up, we have uh Yan Villain. >> Yeah. Um, portable projection mapping with Blender. Um, I'm Yumm. I do lots of things, including volunteering at a maker space near me. And essentially, I wanted to build something flashy to pull people in. And projection mapping sounded like a fun way to to achieve that. Um but when I uh tried the existing ways to do it, I ran into a few limitations like uh it's often fixed to a single spot or room. Uh it's hard to set up and the viewing angle is limited, right? You have a projector, it projects onto something and you pretty much have to look at the thing from the direction of the of the projector. So I wanted something better. I wanted something portable that we could easily move around depending on the event or or uh what what we were doing that day. Uh I wanted something that you could just turn on and it would work. There would be no setup and no calibration. And I wanted something that would be viewable from all angles. So why use Blender? Because I love it, but also because I know it. It's very powerful. Uh I could do the table design and the motion graphics. uh uh in the same program and Eevee opens doors for interactivity. So here's a proof of concept, super fancy, but it actually solves a lot of problems. Um because I project from the inside out, um you can walk around and because it's vertical, if I connect those two, uh uh it becomes portable. And once I calibrate it, it's done. I just have to turn it on. So, um, designing the table, this looks very clean and organized, but this is more how I ended up. The process to get there was a lot more complicated. If you want to know more, I can show you all the drawings and notes in here, and I have way more. So, this is not really uh representative of the truth. Um, building the table. Um, again looks very nice. The truth is more like the photograph in the middle. Uh, what was really nice uh uh because I use Blender. I could do a lot of things with geometry nodes. So I have a geometry node set up to planerize the shape. I have a geometry node set up to generate the uh files for the 3D printing. Uh I use geometry nodes to generate the cutting files for the material. Um so that worked very well. Um, and because Blender is awesome and Eevee is real time, I can stream from Blender to the projector table, which means I can alter this in real time. And I could even make it interactive. Uh, for example, with touch or I can make it audio reactive or maybe I can do something with media pipe or a camera where I actually use video from people that are looking at it. So this is uh the final result where I painted the top surface. Uh and this is actually a fluid sim uh that runs in 2D on top of the shape. Uh it's from a notepad by 3D Sing VFX. Uh, and it it practically runs real time, which means that even something like this, if I would make like uh a magic wand with an infrared light on the top, people could perhaps stir it themselves or spatter it and do other fun things with it. So, who am I? Well, I already shared my name. I'm Yulclar. I love Blender. I love problem solving. I love bridging disciplines which was necessary to create this. I love beautiful things. I would consider myself a recovering programmer. I I code when I have to but I prefer to do the other things and I did this lightning talk because I would love to exchange ideas. So if you think this is cool or you have ideas or things to share or would like to collaborate in some way, find me. Uh, I would love to talk. All right, we have two presentations uh remaining. So, next up is Kelvin. Hello again, everybody. I'm back. Um, I mentioned earlier in my back rooms talk that I uh also do short films and a couple people asked me. So, I quickly threw this together this morning. Um, it's a just a little piece of a short film that I've been working on for a couple of years. And the pitch the the pitch for the film, well, it started as a as an exercise in trying to learn character animation, being a better an animator, more rounded animator. Um, and so I thought, what better character rig than the Lego rig? Because the Lego arms just do this and the legs just swing. So, sorry, >> the empty parachute game. >> We're going to stop that. Um, the arms swing and the legs move and that's it, right? So, very simple. Um, so I put this together. The pitch to myself is like if Love, Death, and Robots did a Lego film. Um, so it's kind of like the Lego movie meets Wolfenstein, if you're familiar with the video game. Um, the the end product will be on YouTube. Um, so if you want to check that out, you can just look up my name, Calvin Roma, and follow me there. Um, the the film itself has kind of like adult R-rated animations, so I've I've clipped together something that I feel is appropriate to show here. Um, but I'd love to chat with you if you're interested in seeing the rest of the film. Can we start this over? I just bumped the button. Do you mind just >> Yeah, we can just uh >> the empty parachute gag. My favorite >> Harlem. But somehow it appeared to Texas gritty and cool and jaded like everyone else. Maxwell. >> Well, they didn't exactly say that last part, but I know that they think it. >> Whoa, a hotel. Hold on. Hello. Anyone there? Red. >> I hear goodness gracious. What the heck was that? It's okay. He's too big to follow us up here. We'll get to the top and try the radio again. Look, man. I know you're like a literal head on a giant hunk of metal science. >> If I had that cool mech suit, >> I'd use it to help people. >> But I want to make sure you're never able to hurt anyone else ever again. And thank you very much. >> Thank you. >> All right. And then we have Suban talking about the land of Blender. Or was it Blender land? I don't know. Yeah, I I'll let you do the talking. >> Okay, thank you so much. Uh, we don't have to translate this. >> Hello everybody. Uh, last month I was at a big conference in Germany and it was really welcoming and really awesome. and a lot of huggy people. And one thing they they did really well that made me feel envious is that they had a map of all kinds of restaurants in the vicinity and you could like select and what your uh preference is like maybe you want to eat vegetarian or vegan or kosher or gluten-free or whatnot and you could filter down on these restaurants and just see what's nice in the vicinity to eat, what other visitors there actually recommend. And so I got envious and like over last weekend or so uh I put together a little map did a call out to social media to colleagues, friends. Thanks for everybody who helped. Um and yeah, you can see like there in Nord there's uh the Blender Institute here we have Felix the Bali and all kinds of different restaurants and places and you can click on it and get some information about it. uh like best drop baffles are over here. Um and so this is something that's here now uh is on blender.land. I had the domain laying around for something cool and I figured it's kind of fitting. Um and yeah, I would love to have some more help with this. Uh there is a link over here. If you close the popup and you didn't find it anymore, my name is over there. you click on it, you have a contact form because I would love to have more info about, okay, these restaurants are cool, but the vegetarian or the vegan or the whatnot that they offer was a bit meh. So, it's on the menu, but it's meh, or this one is purely vegan and it's amazing. You you just stop caring what you eat because you're just blown away. Like, all that kind of info I would love to have. Uh, I have more plans on putting more filters because that's not there yet. Uh, etc. So if you see a restaurant on this map and you know oh that's not just vegan it's purely vegan or they offer vegan or it's none like all of these details really help in building such a system and I would love all your help with this. So thank you very much. All right. Thank you so much everyone. Um before we move on this evening, I want to uh say a a few words about uh the Susan um not the Suzan the lightning talks. Um can we get uh the screen back please? So um every year there's a little bit of confusion how to apply for the lightning talks, where to find information. So um for next year if you want to have a lightning talk just go to the conference website conferenceblender.org talk, go to the schedule and then um yeah, go to the lightning talks um entry uh which is usually on Thursday evening um and then you'll find all the information here. So it's just saying send me an email with a a title, a short description and the link to Google Slides so we can have it on one computer. Um it can also be work in progress, you know, just so I have all the information. I can already put it down in my spreadsheet. that makes my life easier and then we don't have to exchange five emails with information. So just send me one email with those free information bits in there. The slides can be work in progress and then I have everything. And uh also if you're wondering uh what's the order of the lightning talks on the evening then you can find it here. So whenever you have a question just go to the conference website and then you find everything. Also it says we're full. So if you're wondering is there still a free slot available or not you can just go there. it will tell you everything. So, um I wanted to say thank you everyone for the amazing presentations. It's always cool to see uh how people use Blender and what kind of projects they do with it. So, I really enjoyed it uh to be your host this evening. Thank you so much. And now it's time to reveal the winners of the Suzana Award Festival. So, thank you so much