In part 2 of "binge-watching all the Blender Conference 2024 videos" let's have a look at videos 11 to 26. I found a lot of the videos kind of dull, but that's not because they were, it's because the topics were not for me. I'm impressed by all the smart people giving these talks.

However there were a couple of ones that stood out (in a good way!). For the most interesting talks, I've put a video thumbnail in this list, so you know which ones I personally would recommend watching. Again, this is purely subjective.

You can find the entire playlist on the Blender YouTube channel.

Note to Gmail users: If the e-mail gets clipped at some point, just click on "View entire message" at the bottom, to load the rest of the content.

Marion Stalke talks about the differences in how humans and quadrupeds stand on their legs, which relates to rigging. Some animals stand on their nails (horses), some on the tips of their fingers (dogs, cats), so it's important to keep that in mind when rigging and animating animals. Check reference! She also discusses the differences in physical aspects of male and female animals. Rigging birds: Everything doesn't need to be connected (in terms of bones). Weight painting is very important! Good mesh flow is important (unless you hate your rigger).

They do architecture renders of historic buildings, and they were missing oblique projection in Blender, so they added it. As a Blender user, you probably know the difference between Perspective and Orthographic cameras. Do you know what an Oblique camera is? Like orthographic projection, it's a type of parallel projection.

They've made an addon that lets you easily select any of the (apparently many) different kinds of parallel projections, so maybe it's interesting to experiment with if you do this kind of architectural renders or need to visualize floor plans in an interesting way.

The add-on can be found here: https://extensions.blender.org/add-ons/pohlke/

This team created some very detailed renderings of buildings in Brussels that have historic value, and in some cases no longer exist.

A fast-paced  talk about color spaces and the workflow between Blender and Davinci Resolve. The speaker also has a course coming out about this topic, in the context of integrating VFX into live action plates. He has invested  a lot of time learning and researching this topic. His YouTube channel has well produced content about these topics.

A talk by three guys from an agency that focuses primarily on high end car renders. https://mondlicht-studios.de/ Really impressive work. They highlight some add-ons they use, and the importance of their community and some other tips. 

Apparently they also sell their own add-on for car animation in Blender, called Launch Control. https://danielvesterbaek.gumroad.com/l/orzsz There's a few car animation add-ons for Blender on the market now, so it's a good time to try and make some car renders!

Rendering tips for improved results. They have beautiful glass renders, how? Just the simplest glass node material. It’s all about lights, and light with gradient textures. Also render different passes for the different types of lights in your scene, combined together looks much cleaner and renders faster. 

Render complex scenes out in layers, to make it easier on the VRAM. He also talks about how they use AI, something called Magnific. He uses AI not to generate a final image. But just to make suggestions based on a render, which he then implements with modeling, etc to create an improved render. So kind of like an AI art director, giving some suggestions for improving the scene. Kind of like using AI-driven reference. 

I think the main takeaway is, that it’s not about tweaking very specific render settings. But it’s about art directing the scene, good lighting, materials, and splitting it up into different light passes, different layers that are then combined again later. 

This is kind of two talks in one, the first is about math and geometry nodes, and to be honest went over my head. Maybe it’s about magnetic fields around the earth, and how to visualize them with geometry nodes. But I could be wrong! The second part is about 3D printing the moon.

In the moon talk, he discusses some of the obstacles in making a nice moon model from a height texture, like choosing the right color space (non-color) and avoiding the pinching at the top (which I’ve also made a video about- https://youtu.be/iNL98QwGEmQ?si=-qsDj9tXmpQAEBnN) He uses a different method than the one I show in my video, using a Python script to change the UV mapping of the Icosphere.

Special appearance by a little spider at 26:30 (going on its merry way on the white wall behind the speaker). 

A talk about open source stuff, I have to admit this was a bit over my head. I'll be honest, I didn't finish watching this. You'll probably know it from the title, if this talk is for you!

Some theory about denoising of images. For me, this is one of those things that just have to work, if you check a box, I don't really need to know exactly how it works. But I'm sure there are people who will find this interesting. I watched this at 2x speed to see if there were any juicy "tips", but no such luck. However, much respect and gratitude for this person trying to make Cycles better.

A talk about the advantages and reasons behind Blender switching to the Extensions platform instead of having lots of add-ons shipped with Blender. The speaker is Nika Kutsniashvili, who if I'm not mistaken is the creator of Bool Tool and some other great add-ons that I use very often. 

He explains how stressful it was previously to have to update add-ons according to the official release schedule of Blender, and how the Extensions platform is very liberating for add-on developers.

He appeals to educators and youtube tutorial guys like yours truly, to spread the word about add-ons. I'm not sure if it's true that we don't make enough videos about add-ons, but I guess there are many, many add-ons that are never talked about.

A talk about physically based rendering, very technical. Do you want to create your own rendering engine? Well, in that case I know just the talk for you to watch…

Another pipeline talk, from an NPR animation production studio. Pretty standard stuff. Complicated nodes, linked assets and library overrides. They talk about some of the challenges and benefits of using Eevee. They use Geometry Nodes to fake render passes for the materials.

They flatten the characters sometimes from the viewpoint of the camera using geometry nodes, to control better the 2D look.

A sponsored talk about FAB, the Epic-owned marketplace that replaces Sketchfab and combines it with Quixel,  Artstation etc.

She recommends focusing on one thing ("become the person known for …") which I completely agree with and recommend as well. As well as some other general tips for publishing your assets online.

"Set up social media channels to advertise your products" is one of the tips. So very general, basic stuff. I recommend that you set up a social media channel to advertise your products, and retire to a tropical island living off your passive income. What are you waiting for?

If I may give a tip to people thinking of selling their 3D scans or models on these platforms. Please, make them actually useable. I've purchased many 3D models on these channels (especially Sketchfab) and most times, you have to spend about an hour to reconstruct the messed up materials, relink missing textures, fix issues with the model etc. This is why I actually stopped using these platforms. The models are cheap, but my time is not (says the guy who intends to binge watch 86 BCon videos).

Jacques Lucke talks about nodes. Interestingly, this talk has an interactive component using https://polli.live/ - I thought that was nice, and it seemed to function well.

The official description of this talk is "A technical deep dive into various aspects of how Geometry Nodes is evaluated". So, only for those of you who are into that sort of thing.

So this is more a talk about machine learning than anything Blender related. However to train their software to quickly recognise soda cans in the supermarket, instead of training it on actual photographs, they use renders of those cans made in Blender. 

That allows them to quickly include for example a transparent screen in front of those cans, that might be present in some store displays, without having to go out to those stores and take photographs.

You won't learn anything about Blender from this, but I guess if you're working with visual machine learning, this could be interesting?

A live geometry nodes tutorial about creating a visualisation of points that react to music. Not much more to say about that, other than that the art with all the cut out images is intruiging. 

A talk about a previz / storyboarding process using Grease Pencil. She mentions CG Boost’s environment course, and Gaku Deep Paint. 

She starts with the script, makes a floor plan to model in Blender, and records voices with “Casting Call Club”. As she puts it: Yes, you can use AI to do voice overs now. But we create things in order to connect with others. And there are many voice actors waiting to work with you.

Edit the sound in Blender, so you can change the pacing of the dialogue easily. She recommends against using images on planes for previz because it’s too slow. Just draw in 3D with Grease Pencil with the default brush.

She has a more detailed tutorial on her YouTube page: https://youtu.be/vBbLZGsfLYE?si=AL8WqB0kTfRTHw-b

And in the comments of this talk on YouTube she also links to a bunch of other useful videos.