Günter Nikodim

Wetmaps: I created a vertex map in field mode on the dinosaur and put the liquid mesher (with a bigger voxel size so it renders faster) into the field list. In the field settings, the mesher was set to volume, and I put a decay field on top with 100% so it stays wet where it has been touched by the mesher. I rendered the vertex map just as black/white mask, so I could use it in comp to darken the original render. Additionally, I rendered a seperate, very reflective pass of the dinosaur, that I added on top in comp, again using the wetmap as mask. trex_water_v007_Wetmap.mp4 I then comped everything together in Nuke. I tried to make a smooth transition between the simulated area around the dinosaur, and the rest of the lake, but that was a bit of a challenge and didn't work out to my full satisfaction...it's still noticeable that the simulation doesn't spread out to the rest of the lake. Some final thoughts: AFAIK, our liquids solver is an SPH solver, which is generally designed for small-scale fluids. FLIP would probably be better suited for fluids of that scale. I find it even more impressive that the new solver is capable of pulling this off. It's also very speedy and very stable. Fully art-directable as it works with all the forces, all other simulation types and all the particle modifiers. I enjoyed that little project.

Here are some insights from that little project: Easy things first. I reused the animation from my T. rex museum breakout: The environment is an HDRI from the asset browser. The lake (or river) is just a reflective plane with animated noise, and a gradient in the opacity channel to fade it softly into the backplate HDRI. The trickier part was the simulation, naturally. I wanted to simulate only the top layer of water, and I needed to figure out a way to only simulate particles in the area around the dinosaur. A whole lake would have been way too many particles of course. At first I tried to simulate it only in a "pool" that was a bit bigger than the T. rex, using the liquid fill emitter. It worked, but I had visible splashes at the border of the pool - unfortunate if you want the transition between simulation and the rest of the lake as seamless as possible. Instead of walls, I then gave the pool a slightly rising floor towards the edges. That helped mitigating the splashes at the border. However, I still had the problem that the liquid spreaded out pretty fast and drifted away from the dinosaur, where I actually wanted it to be. The beauty of the new liquids is that they work so seamlessly with the rest of the particles system: I could just use an attractor force in the center of the dinosaur. With the right settings, it helped keeping the particles near the T. rex, while still allowing it to act like a liquid. Here's the cached particles with the collider geo: trex_water_v007_Viewport.mp4 The sim contains 3.7 million particles. I used the liquid fill emitter with a radius of 0,7 cm. Sim time was IMHO very reasonable on my RTX 4090, around 20 mins for 150 frames. I would have liked to simulate even more particles, but when I lowered the radius to 0,6 cm (= 6 million particles), I didn't even see the first frame after waiting for several minutes, so I quit that. The cache is 27 GB. I only cached velocity, color and radius, which helped to reduce the size. I then used the liquid mesher with pretty much the default settings, but smaller influence scale, a lot more smoothing and the droplet size set to 10% - that really helps getting rid of those huge blobs that often appear. Here's a clay render, liquid mesh only: trex_water_v007_Clay.mp4 The particle color is mapped to the velocity, which was very handy to fake whitewater. I just remapped the particle color with a ramp to diffuse and reflection strength. The faster the particles, the brighter they are. I rendered them as a separate pass and added it on top in comp. Those particles render crazy fast! A few seconds per frame for 3,7 million particles with motion blur... that was a real joy. trex_water_v007_Whitewater.mp4

Here's a little project I made while I was betatesting the new liquids. trex_water_v006_high.mp4

Characters and rigging are also the areas that I'm mostly interested in. So I fully agree with all your points... oh how I wish Xpresso was faster and would use the new node UI... Anyway, even though these areas seemingly weren't touched, with character rigs of medium complexity I roughly got a speedup of 30%. That's what I meant with almost everyone will benefit from these core speedups. I still wish for more optimization and performance, but it's certainly a lot better than it used to be.

I am neither offended, nor do I think Maxon should get "anything less than full blown adoration". Quite the contrary, I've been criticizing Maxon on countless occasions. I just think this is a quite extensive release that offers substantial improvements in many areas. Just my personal opinion, of course.

These are the kind of comments that I honestly don't understand. Have you even tried the release for a minute? The performance gain is something the whole community has been asking for for the last 10 years and now almost every use case benefits from it. Finally the new core shines through... and your conclusion is "nothing groundbreaking"? I also don't know how one could come to the conclusion that Maxon concentrates more on ZBrush than Cinema. C4D just got a massive update in this fall release, ZBrush didn't.

Here is the most extreme performance gain that I found during testing: 1 fps in 2023 vs. 49 fps in 2024! This is not the usual performance boost and certainly a crazy example, but there is a noticable speed increase in almost every scene. 2024performance.mp4