I have finally got this liquid sim looking somewhat realistic, except, of course, for the square of liquid at the bottom and the extra droplets bleeding through the glass.
Is there away to erase this post bake? I saw something like it in an Andrew Price video, but can’t for the life of me remember which one it was.
(Also, if you could give me some advice on texturing in cycles I would appreciate it a lot. I have watched some tutorials on it, but, again, cannot remember which of them went over cycles texturing.)
To prevent the fluid from going through obstacles, you just need thicker obstacles.
Take a copy of the inside of your glass, extrude to make a big block around it and use this new object as obstacle instead of the glass. When the baking is done. Hide the obstacle, show the glass and… admire! (Don’t forget to deactivate the fluid properties of the glass. Less calculations for Blender…)
And if you see weird artifacts where the fluid meet the glass, you can always use the thick obstacle to cut out the fluid domain thanks to a Boolean modifier.
Your proxy doesn’t need to be huge, but it does help if it’s well subdivided. Also, if it’s “huge” make sure it’s not bigger than the domain. If it’s still not working, post a .blend for us to diagnose.
Also, isn’t this like the third new fluid-going-through-a-wineglass-obstacle thread in this forum in as many days? Don’t people read through other threads before posting?
Most of them were me xD It’s my first project, and I’ve found that a lot of my problems aren’t applicable to the problems that other people are having. I will try to perform better forum etiquette in the future though sorry for the trouble.
Haha, I didn’t even look to see who they were. To be fair, I think only two of them are you, counting this one. It’s no trouble, I just thought it was funny. You should probably stick to one thread if they’re all related to the same basic issue though. But since we’re all here in this thread now, can you post a current version of your .blend so we can see what the story is with this?
I can, but unfortunately I forgot to save so the issue that I’m having now is getting it back up to date. I will post when I figure out what I need to do again. Thanks for helping me again. Good people on this forum.
This is my current file. I have made multiple upon multiple proxys for my glass and bottle obstacles, but continue to have the same exact problem. I am obviously missing some important fundamental point on how to avoid this problem. Anyways… the file is attached now (hopefully it is the right one).
Yes, I know that the proxy’s in the file are probably way too big and whatnot, but I can’t seem to understand why the liquid is going through the obstacle and it is getting on my nerves, so I just made the biggest thing I could think of.
Once I figure this out I still have to do the floor and wall texturing! Which will likely lead me to more beginner problems that I may torment you with! Muahaha.
I don’t have time to play with it before work today, but at a quick glance I see two things you can change that will probably make a difference:
The visible wineglass and bottle don’t need to be obstacle objects. Just their proxies on layer 3.
The inner wall of the proxy should be inside the visible object’s inner wall. So the wineglass proxy should have a smaller bowl than the real wineglass. That way a small amount of penetration through the proxy’s wall won’t matter; it’ll still be within the bounds of the rendered wineglass.
That probably won’t fix everything, but it’s a start. Once you get everything in place, try applying those solidify modifiers before running the simulation too. If that doesn’t do it, I’ll play with it more later.
Wait, no, I lied. You don’t want to have that solidify modifier on at all. It’s just making the thing hollow. I thought you were using the solidify modifier to generate the thickness of the mesh, but you’re just making a hole inside the mesh. If you’re going to use a solidify modifier on this at all, it should be because you only modeled either the inner or outer wall of the obstacle. What you’ve done is a good method for modeling a Bodom double-walled tumbler, but very little else.
You also have some messed up geometry on the bottom of the proxy. The convergence point at the very bottom still has a whole edge loop’s worth of vertices sitting in the same space. They should be merged. Then you’ve got an edge running from that up to the same point on the inner wall, which has the same problem. They’re just far enough apart that Remove Doubles will miss it, probably also leaving just enough of a hole to confuse the fluid simulation.
Both proxy obstacles are off center from the object they’re meant to substitute. I’ve recentered the wineglass proxy and remade the bottle proxy altogether. Its geometry was a bit weird and its normals were all facing in the wrong directions, in addition to it being in the wrong place. The rendered wine bottle’s normals were also messed up.
Your domain was small enough that one proxy was partly outside it. I believe that can cause problems.
Finally, you had a bunch of things that didn’t have their scale applied, including both proxies and the domain. That will definitely cause problems. I’ve fixed all of those things, but I didn’t run a test simulation yet. Try this out and see if it at least improves matters a bit. winenewcorrections.blend (783 KB)
Dang! I was doing a lot wrong. So just a run down.
Should the normals be pointed inwards or outwards when I recalculate them?
and what scale are you talking about? Should everything be scaled to one? (That’s control-A right?)
And you’re saying I need to merge/remove doubles? I completely forgot to do that.
Normals: outward from the mesh surface. That means outward from the outside of the bottle and inward on the inside, right?
Scale: Hold your mouse over the 3d viewport and hit the N-key. There are a bunch of properties on that panel, including scale, location, rotation, dimensions and other stuff. If you scale something up or down in Object Mode, that changes its scale property. You need to apply that before running simulations or rigging or a bunch of other things. And yes, it’s control-A, or down in the Object menu.
Merge in this case just because you hadn’t scaled the edge loop down to 0, where merge doubles would work. But you also had a face loop connecting the inside and outside of the glass bowl, meaning that topographically speaking, it was actually a weirdly stretched torus. Merging (or scaling and removing doubles) results in a single edge running inside the mesh, which is never a good thing to have.
I made both my bottle and my glass with the spin tool, but was having a lot of trouble with it, so I assume that that is where a lot of my problems are coming from. Another user told me the spin tool was a better option, so I probably just need more practice with it.
Ah, that makes perfect sense. So the vertex that should have been at the very bottom of the bowl was probably just a hair’s width away from the z axis, and the spin tool just duplicated it all around. You have to zoom in quite a bit to spot the problem. Which reminds me that I didn’t check the bottle for the same problem, though it probably doesn’t really matter much since the wine won’t be flowing uphill.
The other user was right though, the spin tool is the best way to model something like this. You just have to watch out.
Just to make sure it’s clear, all of those things I listed as wrong in post #10 above were fixed in the attached file. So you don’t need to recalculate any normals or anything unless you want to start from your file and modify it yourself to match mine.
Now come on and run that simulation! I want to know if it worked. :eyebrowlift:
well… it came out TONS better than mine, but the liquid is still moving through the glass… I was thinking about maybe applying the liquid modifier to the wine and then moving the outside vertices inside of the glass? I don’t know how professional that is. Also, the wine isn’t all in one consistent flow (I literally just noticed I hadn’t changed the time settings in the file, so this might fix that? maybe?), but I want the liquid to look as realistic as possible. Everything else looks good though. The proxy’s you made worked way better than mine.
I will post the file so you can see what I am talking about with the liquid. This project is teaching me way more than I ever thought it would haha.
I was toying with this last night too, and I have a few more suggestions that might help.
It’s not moving through the glass so much as it is spraying all over the place. Hit Z and advance through the first 40 frames one at a time to see what I mean. Cut way down on the particle settings for a start. In fact you might just turn off fluid particles entirely, at least until you get the basic flow right.
A lot of that problem is actually caused by it splashing around in the bottle before it every pours out. This is because your liquid object is a floating egg that has to fall down in the bottle before it is poured. You’d be better off to make a cylinder that is fitted to the inside of the bottle, make sure the end caps are quads or triangles, not n-gons, W>Subdivide it once or twice, then use that as your fluid object instead of the squashed sphere you’re currently using.
How would I go about turning of fluid particles? The only thing I did was press the fluid option in liquids. That’s really the extent of my knowledge. I know how to use an inflow and whatnot, but I feel like that would just cause more problems. I will make a fitted cylinder and see what happens. I’ll probably be able to upload another file by tonight, hopefully before.
Way down at the bottom of the domain’s fluid setup there’s a panel for Fluid Particles. Try turning it down to a lower number at least. That should cut down on the spraying a bit.
I’m attaching a new version where I did that for you and also made a new fluid object (it’s on Layer 2 now). Of course now there’s a lot more wine in the bottle, so it will eventually overflow the glass if you don’t tip the bottle back up first!
You might eventually want to turn particles back on to some lower degree, depending on how graceful a wine pour you want this to be. I’m inclined to think leaving it at zero will probably be fine though. But it’s a setting you should learn how to play with. winenewcorrections2.blend (793 KB)
Edit: Oh, I never noticed a few posts back where you were considering applying the simulation and manually moving stuff. I’m gonna go with “No” on that one. Not so much because it’s unprofessional, but because you can make the simulation work right if you tinker enough. And doing it manually like that will be an enormous pain that will never look as real as just working through the simulation’s issues.
OK, damn you, now I can’t stop playing with your fluid simulation instead of getting back to my own project.
I don’t know how I missed it, but there are also unapplied rotations. Same problem as the unapplied scales, but slightly less disastrous to the final result. Still, you should look into it.
I’ve also made a few other changes to the timing of the simulation (also under the Domain’s settings). I’m still fiddling to get that right, but you might try adjusting that yourself if you’re working on this over the weekend. I am running a sim overnight, but then I’m going away for the weekend. If the results are favorable I’ll post another file in the morning before I leave. Either way, you should try the following on your own:
Move and rotate the bottle/proxy/fluid objects so the wine is pouring down more than across. Part of the problem, now that I’ve gotten it into a coherent stream coming out of the neck of the bottle, seems to be that the fluid is splashing off the rim of the glass and going all over the place. Your aim is off.
Make your domain a little bigger. Get that back face away from the glass in case there is a little bit of splashing.
Hit the N-key in the 3D view to bring up the Properties panel. Take a look to make sure all your scale values are 1 and your rotations are 0. Now look at the dimensions and think of those numbers as measurements in meters. It’s huge. I am honestly not sure whether or not this affects the simulation much, but I think it does. Even if it doesn’t, it’s usually good practice to stick to a 1 Blender Unit=1 meter scale. I scaled everything way down to something closer to real world sizes for these objects (and applied the scale!), and it seemed to cut down on the splashing effect. That’s how I spotted the bad aim of the bottle.
Edit: Yep, it’s working really well now. The fluid is still penetrating the walls of the bottle just a little bit, and now it’s moving a bit too fast because of the way I set the start and end times for the Domain, but I think this is close enough for you to tinker away the remaining problems. winenewcorrections3.blend (792 KB)
(Don’t forget to deactivate the fluid properties of the glass. Less calculations for Blender…)
Thanks!
To be fair, I think only two of them are you, counting this one. It’s no trouble, I just thought it was funny. You should probably stick to one thread if they’re all related to the same basic issue though. But since we’re all here in this thread now, can you post a current version of your .blend so we can see what the story is with this?
