Update edit: Been a few months, so I figured I'd put some updated visuals in the first post. Below is an animation in progress. I'm rendering a version without hair clipping right now, but this is my most recent output... 150 frames looped, each frame is roughly 5 minutes of render time on my 4-core comp (cpu-based render engine), so I'll set it to go overnight and it should be finished. Obviously, production and process is in the thread posts.
Original posts starts below.
First things first, I'm a graphic designer. I've never worked in 3D professionally, but I know my way around graphic programs.
So, initially I stuck with Daz and got decent enough at posing and all that to start work on what I was going to make into a game. Pretty normal. But I'm sort of a perfectionist... and I wanted to have some animated story elements added in. Just quick 10 or 15 second cutscenes to introduce characters or transitions. I saw that Daz had animation tools, so I went to work on that pretty much right away. My initial attempt actually went really well (I think). You can check out a quick render video below with two different light setups. Obviously I intended to combine them in post, but, well... we'll get to that in a bit.
So yeah, I can find a ton of flaws, but whatever. The biggest problem is that that animation took TWO FUCKING WEEKS. One day to get a fifteen second animation down... simple stuff, just twirling and rising into the air, then she senses something, scowls, and falls to the ground in a power pose. Cool. Then I added dForce for movement. Cool so far, EXCEPT... everything fucking collapsed. The clothes had to be simulated one piece at a time, and not simple sims, but 120 fps 20 collision sims that took 6 hours per piece of clothing or they would stick against something and explode. And I had to do that multiple times until I got the wind and gravity correct. Then came the hair. The fucking hair. The wind and gravity on that sucker was so many iterations that a full week was dedicated JUST to getting it to spin correctly, since it took two hours before I knew that whatever wind node or gravity set I used was either too much or too little. Daz is shit for simulation.
Quick tip: Wind isn't actually wind, it's a push force. If you want to blow wind up from below, you can't put a wind node under your character and expect it to act on the hair, because the wind won't flow around the character like actual wind even if you set it to blow at 500 mph, it will simply push the immovable object and then stop moving. To blow hair up, you need to put nodes at the base of the neck.
So I moved on, and then I came upon the first large story scene (the catalyst of the main events) and I wanted to animate that. Great. Here's a still from that animation for the perverts:
So I have her writhing around (of course). but she's slipping and sliding all over the damn place. No amount of pinning is even remotely helpful. Then I realized something... Daz is shit for animation, too.
Enter Maya. I have a student license that goes along with my fees for the local community college (where I fence), so why not. Super complicated, but also really rewarding and SO MUCH BETTER AT SIMULATION. Like... I'm simulating a video that I'll post in a few minutes under this sentence, and I've done it six times trying to get the hair physics right since I started writing this post. And it's twice as long as the Daz animations I did and a dozen times more complex.
Anyway, there are some Maya pipeline tools that let you import characters from Daz, but it's not terribly intuitive. I've gone ahead and used a combination of default rigs, imported rigs, and manual facial morphs to get the below test video. This is a playblast video (so a viewport video, essentially), not a full render. So there's no lighting or subsurface skin stuff, just a simple dynamic hair test (so I can decide how much shape retention the model has in the hair). She was imported from Daz (a morphed Girl 8), then further sculpted inside of Maya, and the hair was created with the default xGen hair utilities. I also have a softbody version for dynamic breast and body deformation (I'll be posting another video of that system shortly).
Pros of Maya:
- Animation is super. Human IK controls lock parts of the body in space like magic, and simple movements are a breeze. Tweaking keyed poses is also really simple once you get used to animation curves.
- I'm rendering with Redshift, because my CPU isn't the greatest, but Maya gives you options, unlike Daz. Also, render view is actually useful and refreshes in seconds instead of minutes. Tweaking lighting can be done in realtime.
- Simulation is not even comparable. Dforce cloth is nice for very simple interations, but this is literally right before my eyes. I can change settings during the actual simulation to correct flaws and then cache the full sim into a static animation in less time than even one prelim sim in Daz.
Cons of Maya:
- Learning curve is steep. There are a lot of tutorials around, but you still need to be really savvy and have a lot of free time.
- There are very few baked in assets. You want hair? You make the hair. You want to move the hair onto a different character...? Well... you can make the hair onto a scalp segment, and then parent that scalp onto the bald head of someone else, I guess. The same goes for clothing.... my characters are currently naked for a reason.
- The exception to the above is material shaders, actually... since the simulation aspect of professional software is so strong, materials are available right out of the box, and third parties also supply a vast number of textures and shaders for pretty much anything.
Ultimate thoughts on this after a month of Maya:
If you really want to use the dynamic aspects of dForce but were really really really disappointed with the constraints (like blue jeans falling to the ground in a puddle like they were made of silk), then you may want to look into Maya. There's a plug-in that is sort of mythical that's been posted about every once in a while on the Daz forums that allows Maya to natively read *.duf files... if that ever sees the light of day, then we'll see a lot more people around here posting renders using Arnold, RenderMan, and Redshift. (it's currently in development with the help of the Daz team, so I'm hopeful). Right now, the time commitment keeps it from being actively used in most game pipelines.
If anyone else around here has looked into this for similar purposes, I'd appreciate feedback or helpful tips.
Original posts starts below.
First things first, I'm a graphic designer. I've never worked in 3D professionally, but I know my way around graphic programs.
So, initially I stuck with Daz and got decent enough at posing and all that to start work on what I was going to make into a game. Pretty normal. But I'm sort of a perfectionist... and I wanted to have some animated story elements added in. Just quick 10 or 15 second cutscenes to introduce characters or transitions. I saw that Daz had animation tools, so I went to work on that pretty much right away. My initial attempt actually went really well (I think). You can check out a quick render video below with two different light setups. Obviously I intended to combine them in post, but, well... we'll get to that in a bit.
So yeah, I can find a ton of flaws, but whatever. The biggest problem is that that animation took TWO FUCKING WEEKS. One day to get a fifteen second animation down... simple stuff, just twirling and rising into the air, then she senses something, scowls, and falls to the ground in a power pose. Cool. Then I added dForce for movement. Cool so far, EXCEPT... everything fucking collapsed. The clothes had to be simulated one piece at a time, and not simple sims, but 120 fps 20 collision sims that took 6 hours per piece of clothing or they would stick against something and explode. And I had to do that multiple times until I got the wind and gravity correct. Then came the hair. The fucking hair. The wind and gravity on that sucker was so many iterations that a full week was dedicated JUST to getting it to spin correctly, since it took two hours before I knew that whatever wind node or gravity set I used was either too much or too little. Daz is shit for simulation.
Quick tip: Wind isn't actually wind, it's a push force. If you want to blow wind up from below, you can't put a wind node under your character and expect it to act on the hair, because the wind won't flow around the character like actual wind even if you set it to blow at 500 mph, it will simply push the immovable object and then stop moving. To blow hair up, you need to put nodes at the base of the neck.
So I moved on, and then I came upon the first large story scene (the catalyst of the main events) and I wanted to animate that. Great. Here's a still from that animation for the perverts:
So I have her writhing around (of course). but she's slipping and sliding all over the damn place. No amount of pinning is even remotely helpful. Then I realized something... Daz is shit for animation, too.
Enter Maya. I have a student license that goes along with my fees for the local community college (where I fence), so why not. Super complicated, but also really rewarding and SO MUCH BETTER AT SIMULATION. Like... I'm simulating a video that I'll post in a few minutes under this sentence, and I've done it six times trying to get the hair physics right since I started writing this post. And it's twice as long as the Daz animations I did and a dozen times more complex.
Anyway, there are some Maya pipeline tools that let you import characters from Daz, but it's not terribly intuitive. I've gone ahead and used a combination of default rigs, imported rigs, and manual facial morphs to get the below test video. This is a playblast video (so a viewport video, essentially), not a full render. So there's no lighting or subsurface skin stuff, just a simple dynamic hair test (so I can decide how much shape retention the model has in the hair). She was imported from Daz (a morphed Girl 8), then further sculpted inside of Maya, and the hair was created with the default xGen hair utilities. I also have a softbody version for dynamic breast and body deformation (I'll be posting another video of that system shortly).
Pros of Maya:
- Animation is super. Human IK controls lock parts of the body in space like magic, and simple movements are a breeze. Tweaking keyed poses is also really simple once you get used to animation curves.
- I'm rendering with Redshift, because my CPU isn't the greatest, but Maya gives you options, unlike Daz. Also, render view is actually useful and refreshes in seconds instead of minutes. Tweaking lighting can be done in realtime.
- Simulation is not even comparable. Dforce cloth is nice for very simple interations, but this is literally right before my eyes. I can change settings during the actual simulation to correct flaws and then cache the full sim into a static animation in less time than even one prelim sim in Daz.
Cons of Maya:
- Learning curve is steep. There are a lot of tutorials around, but you still need to be really savvy and have a lot of free time.
- There are very few baked in assets. You want hair? You make the hair. You want to move the hair onto a different character...? Well... you can make the hair onto a scalp segment, and then parent that scalp onto the bald head of someone else, I guess. The same goes for clothing.... my characters are currently naked for a reason.
- The exception to the above is material shaders, actually... since the simulation aspect of professional software is so strong, materials are available right out of the box, and third parties also supply a vast number of textures and shaders for pretty much anything.
Ultimate thoughts on this after a month of Maya:
If you really want to use the dynamic aspects of dForce but were really really really disappointed with the constraints (like blue jeans falling to the ground in a puddle like they were made of silk), then you may want to look into Maya. There's a plug-in that is sort of mythical that's been posted about every once in a while on the Daz forums that allows Maya to natively read *.duf files... if that ever sees the light of day, then we'll see a lot more people around here posting renders using Arnold, RenderMan, and Redshift. (it's currently in development with the help of the Daz team, so I'm hopeful). Right now, the time commitment keeps it from being actively used in most game pipelines.
If anyone else around here has looked into this for similar purposes, I'd appreciate feedback or helpful tips.
Last edited:
