Prolonging our introduction to Bifrost, we created another water simulation in this lesson, this time by following a powerpoint. I followed the steps to create a basic water crashing sim and then experimented with the Bifrost settings and parameters to make mine unique to my idea.
The first step was to set up the scene. After seeing the example clip on the board and the one shown in the PowerPoint, I understood that I needed to build a corridor and started out by creating an ‘L’ shape. In terms of size, I kept the tip we were given in mind and made it fit inside the grid to ensure that the scale of the simulation could be as realistic as possible.
! Remember Bifrost recognises each grid unit to be 1 meter even though Maya recognises each grid space to be 1cm !
After this, I inserted edge loops along each side and selected the thin faces, extruding them outwards to make walls. It was important that they weren’t too thick or short in order for the simulation to work and be viewed correctly.
Next, I had to create the emitters in order for there to actually be water spilling into the scene. I generated a poly plane and rotated it 90°, reducing the subdivisions to 4 x 4. This part was simple enough, but then we had to use a tool that I had either never used, or forgotten about if I had: the target weld tool. After Googling how to actually get to it in Maya (since I couldn’t find it), I then attempted to create a star shape from my plane. I dragged and clicked and tried different combinations and vertices but I couldn’t get those sharp edges that were needed until the tutor showed me how to do it.
Now that I had the shape of the first part of the emitter ready, I need to give it some bulk. I did so by extruding the front face, after which I duplicated the object and positioned both in between the walls at the entrance to the corridor. I also added a cube, stretched it out and placed it in front of the stars. This one didn’t need any subdivisions, so all that was left was to combine everything and finish them off as one big emitter. I wasn’t sure exactly why we created it in this manner, but I could guess that it was necessary for a big wave.
Like last time, after all the shapes were ready it was time to actually create the simulation. With the combined shapes selected, I changed Maya’s format from ‘Modeling’ to ‘FX’ and went to the Bifrost drop down to select ‘Bifrost Liquid’. In order to make it more of a wave, continuous emission was also needed, so I opened up the outliner. There, I selected ‘bifrostEmitterProps1’ and jumped into the attribute editor to tick the ‘Continuous Emission’ box under ‘properties’.
Before I could play the animation, however, I also needed to add a motion field in order to “accurately simulate objects moving through a fluid”, which is their purpose. They apply force to their parent object so that the object then appears to move or push the fluid, in a similar way to how a boat’s propeller forms eddies, vortices, and slipstreams in the water. At this point, I created a collider, as done in the first lesson so that the water didn’t fall straight through the walls. THen, with both the emitter and its bounding box selected, I went back into the Bifrost menu and selected ‘Motion Field’, thus allowing me to ‘BifrostMotionField1’ in the outliner. I struggled somewhat with the next step so I made a quick playblast here as a progress checkpoint.
As you can see, the water falls from the back, which I didn’t want. It made me jump ahead and solve this issue, but through extrusion and not through the method in the powerpoint. I simply inserted another edge loop at the end and extruded one of the faces so that a wall could block the back.
I then selected the motion field and looked in the field direction tab in the attribute editor, as per instructions. There I could set the direction I wanted my water to flow in, and also change more familiar settings such as turbulence and noise. However, Maya was having a really difficult time responding to my changes and every time I tried to change one of the settings with a slider, it crashed, hence why I got the playblast previously. The direction was fine since I wanted the water to flow to the right, and it was already set to the X-axis so I made do with just decreasing some of the settings manually (not with a slider) and playblasting once again.
I was very pleased with my simulation at this point since it was doing everything it was supposed to and there was a nice impact where the particles made contact with the wall. The following step was the one I would have done next, had I not jumped ahead. Wherever the water exits or the particles appear to be escaping, a kill plane could have been made as an invisible blockage. They are also apparently infinite so their size doesn’t matter (I wish I could have tried them out!). Since I didn’t need any, I went ahead to the final part of the PowerPoint, which involved changing the appearance of the liquid and experimenting. This is honestly my favourite part of the process as I have said before, which makes it very unfortunate that I seem to have problems with it especially. This time, my idea was to recreate the elevator scene from The Shining where the wave of blood comes pouring out into the corridor, which is what I immediately thought of when I heard what we would be doing for this lesson.
In order to play around with the appearance, you select the liquidShape1 tab in the attribute editor and change the following settings:
- Display – Particles, Voxels or both, here you can also turn off the display bounding box
- Particle display – Here you can change the size of the particles. You can also change particles from displaying points to spheres if you wish
- Opacity Channel Remap – In this menu you will have the option to tick either Numeric or Vector. Numeric gives you numerical data as water moves whereas the vector displays lines indicating the different directions of all the particles.
- Colour Channel Remap – This will allow you to change the gradient colour of the water, you can also use the point graph to change how the colours are displayed just below the gradient.
I wish I had done this first, but at the time I went off and did my own thing before coming back to these settings. I can’t explain my reasoning behind straying from the PowerPoint this time, but I can guess that I was overly excited to use presets again and applied an ‘aistandardsurface’ in order to do so. I also remembered seeing ‘blood’ as one of the options, which is what I chose. I was rather disappointed to find that the preset wasn’t all that realistic. My mesh was, all of a sudden, thick and slime-like, and I was thrown into an intense period of tweaks and adjustments in an attempt to make it more realistic, circling back to liquidShape1 settings in the process.
I had to switch between the model in Maya and the render box quite a lot since certain options couldn’t be changed if Arnold was up. I tried turning down the particle and voxel display, as well as altering the colour, metalness, velocity and other settings to do with both the liquid & mesh and the preset applied. Nothing seemed to get it close enough to the consistency that I was looking for.
I didn’t think of this in time, but I developed a theory towards the end of the lesson that the reason this was happening was because of the material itself. I should have applied something naturally thinner (such as milk or water) and then simply changed the colour to red.
I also began seeing some problems with my simulation. I think the number of changes I was making was messing with the emitter, to the point where I could no longer even see the liquid, up until the very last frame. By the time I realised this, there wasn’t enough time to actually attempt anything to fix it, since the final playblast was underway. Unfortunately, I didn’t manage to get my Shining-inspired scene.
It can be really discouraging to fail at the final and most exciting part of the process continually. I don’t think that I’ll ever have that ‘knack’ for Maya that comes naturally to others, since even following clear, laid out instructions, I managed to mess up. In order to improve enough to know what to do in situations like this, I would need to watch a lot more tutorials and possibly take online courses on the side, for which I don’t currently have the time or motivation to do, especially since I am not planning on continuing with 3D. Overall though, the simulation introductory lessons have been enjoyable, especially when seeing the realistic way that the particles behave before any mesh or material is even applied. With a simple scene like this opportunities open up, to create fluids of any kind crashing through buildings (milk wave?), or even planning for water pipes, fountains and other constructions that require water to pour or run through.