Hello, im testing the viscous fluid for an upcoming project. I have this stepping issue as it emits, which idk which way to tackle. Didnt fix with upping substeps by a lot, though it was a bit better. I attach video and scene. Any thoughts? thanks a lot!
ps. btw gpu mode is really buggy unfortunately, been testing it in many different pcs with different cards and setups and still get crashes very easy. For example in this scene i couldnt get substeps to 8-16 without insta crash.
scene: https://www.dropbox.com/s/ryr09ft6jk8ui ... t.c4d?dl=0
video: https://www.dropbox.com/s/hs6kweqml5qrf ... t.mp4?dl=0
Stepping in emitter
-
Thomas Schlick
- Posts: 178
- Joined: Tue Aug 29, 2017 12:35 pm
Re: Stepping in emitter
The stepping issue is indeed annoying, but there are some methods to suppress it:
- Set a higher emitter speed (start with 8-12) and counteract this high speed with a high drag force (4-10).
- Increase the emitter's H/V Randomness to displace particles at creation time.
- Try to work with lower gravity (400-600).
The best results can be achieved when you combine the above tips with higher FPS rates, e.g. 50 fps instead of 25 - if possible. Then you can also make your emitter thinner, decrease resolution, and simulate with 4/6 substeps, for example. With 50 fps it's also possible to work with a standard gravity of 981 again.
The crashes you observe on the GPU might be related to your settings. High substeps and iteration values can lead to instabilities. Of course the scene shouldn't crash, but it might be the reason, esp. since you're saying that you're not able to work at higher substeps. And viscosity shouldn't be 0 when you want to simulate a viscous material
- Set a higher emitter speed (start with 8-12) and counteract this high speed with a high drag force (4-10).
- Increase the emitter's H/V Randomness to displace particles at creation time.
- Try to work with lower gravity (400-600).
The best results can be achieved when you combine the above tips with higher FPS rates, e.g. 50 fps instead of 25 - if possible. Then you can also make your emitter thinner, decrease resolution, and simulate with 4/6 substeps, for example. With 50 fps it's also possible to work with a standard gravity of 981 again.
The crashes you observe on the GPU might be related to your settings. High substeps and iteration values can lead to instabilities. Of course the scene shouldn't crash, but it might be the reason, esp. since you're saying that you're not able to work at higher substeps. And viscosity shouldn't be 0 when you want to simulate a viscous material
-
Thanos Kagkalos
- Posts: 34
- Joined: Tue Nov 28, 2017 11:09 pm
Re: Stepping in emitter
Great tips Thomas thank u !! With some more tests, i noticed that gravity is doing this issue more than anything else but i didnt wanna mess with it. So 50 fps gravity behaves proper? Also scene scale matters? I ve noticed that if i am at around few meters, everything behaves better. With real scales i was having issues of things going fast! Sorry my feedback is scattered here and there cause i dont use RF daily.
This thing with gpus though, i get unreasonable amount of crashes and fortunately for most of em, support reproduces the bugs. In weekend i ll spam support with all the problems i get so to get fixed once and for all cause right now, my gpu is not useable unfortunately and that removes me a lot of speed!
Thanks again!
ps: a vital question, if i go 50 fps and sim there, can i playback at 25? thank u!
This thing with gpus though, i get unreasonable amount of crashes and fortunately for most of em, support reproduces the bugs. In weekend i ll spam support with all the problems i get so to get fixed once and for all cause right now, my gpu is not useable unfortunately and that removes me a lot of speed!
Thanks again!
ps: a vital question, if i go 50 fps and sim there, can i playback at 25? thank u!
-
Thomas Schlick
- Posts: 178
- Joined: Tue Aug 29, 2017 12:35 pm
Re: Stepping in emitter
You're welcome Thanos. Glad it's helpful.
This stepping issue is something that bothers me too and the bad news is that it occurs with any fluid type: sph, pbd, visocous... And it can be really difficult to find a workaround. I also have scenes where I've failed to get rid of this chunky and clumpy look.
In your test scene I was able to remove these chunks almost completely - even with a gravity of 981, but 50 @ fps. Scale always plays a role and many RF artists say that working at real world scales is essential. My approach is slightly different here: scale makes sense, of course, but if a simulation looks good and real in the end no one will ask you if scalewas set correctly. But that's just my opinion and the "true scale" supporters make a valid point.
The crashes you observe are related to Octane. I've mentioned it in the first version of my post here, but one of our developers gave me a hint on issues with substeps. So I've edited my answer. We thought we've fixed these Octance crashes and spent a lot of time on improving this bridge, but obviously there's still a GPU conflict.
Unfortunately, there's no way of playing back a 50 fps simulation at 25 fps in C4D. I normally do a retiming in RF standalone or in composting. But let me take another look, maybe I'll find a method...
Thanks for all your help and patience. I appreciate that and your bug reports help us alot.
P.S.: I'll upload the test scene later that day. It's on a different computer. Maybe it's a starting point for you to get the scene done with 25 fps.
This stepping issue is something that bothers me too and the bad news is that it occurs with any fluid type: sph, pbd, visocous... And it can be really difficult to find a workaround. I also have scenes where I've failed to get rid of this chunky and clumpy look.
In your test scene I was able to remove these chunks almost completely - even with a gravity of 981, but 50 @ fps. Scale always plays a role and many RF artists say that working at real world scales is essential. My approach is slightly different here: scale makes sense, of course, but if a simulation looks good and real in the end no one will ask you if scalewas set correctly. But that's just my opinion and the "true scale" supporters make a valid point.
The crashes you observe are related to Octane. I've mentioned it in the first version of my post here, but one of our developers gave me a hint on issues with substeps. So I've edited my answer. We thought we've fixed these Octance crashes and spent a lot of time on improving this bridge, but obviously there's still a GPU conflict.
Unfortunately, there's no way of playing back a 50 fps simulation at 25 fps in C4D. I normally do a retiming in RF standalone or in composting. But let me take another look, maybe I'll find a method...
Thanks for all your help and patience. I appreciate that and your bug reports help us alot.
P.S.: I'll upload the test scene later that day. It's on a different computer. Maybe it's a starting point for you to get the scene done with 25 fps.
-
Thomas Schlick
- Posts: 178
- Joined: Tue Aug 29, 2017 12:35 pm
-
Thanos Kagkalos
- Posts: 34
- Joined: Tue Nov 28, 2017 11:09 pm
Re: Stepping in emitter
Thanks so much Thomas. U all been very heplful and patient with my feedback as well. Very nice customer support which i dont get from other products 
Thanks for the file as well!!! Looking forward to do more and more stuff in RF!
Thanks for the file as well!!! Looking forward to do more and more stuff in RF!