FX Blog - Smoke Plume

Smoke is quite a general element that can be classed under different categories. For example there are: dust clouds, which look like smoke but are affected by gravity; linear smoke - things like cigarette trails; steam, where the water vapour heating results in smoke-like properties; explosion smoke; and finally billowing smoke, something you’d see as a result of heat on a medium to large scale. To simplify things, I will be focusing just on that last one for this section of my project.

Gilland (2009, p. 161) says that for smoke, animating it is “far more about understanding the energy that is in play than moving pretty shapes around in random ways”. That second part is pretty much what I ended up doing for my first attempt, and as the experiment showed it was unsuccessful, therefore to animate smoke properly I must know how it works first.

Smoke essentially is just “the waste by-product of a material reaching its combustion point” Gilland (2009, p. 196), that reacts to air currents such as wind, movement of the object the smoke is coming from and the intensity of the energy source - how hot the fire is. So for a billowing smoke caused by a large fire, the hot air pushes the smoke upwards quickly and the cool air above it slows it down, acting like a barrier due to the friction. This causes the smoke to billow outwards and forces it to roll over itself. The continuous rolling effect you see in large smoke clouds is caused by the hot air sucking the rolled smoke back into itself, David Tidgwell (no date) and Gilland (2009).

Upon dissipation, smoke breaks down into smaller shapes rather than fades out with opacity, Gilland (2009, p. 201). A common problem when animating the dissipation is making the particles too similar in shape, size and spacing, which makes it look unnatural. Gilland suggests changing the path of action - which way the general movement is. David Tidgwell recommends dissipating the particles by having “one area ‘bite’ into the shape faster than the rest of the reduction”. This gives the smoke quite a dynamic look. For bigger plumes, adding more holes is needed to properly suggest its size.

Both Tidgwell (no date) and Gilland (2009) suggest that you start off the animation very rough, using spheres to describe the form and concentrating on the animation before going into the details. When cleaning up the animation it is important to add a degree of shakiness to the circles as smoke particles are affected by external forces, they’re not just rising spheres.

David Tidgwell (no date) says the “best way to start an effect you have never animated before is to first study any live action reference you can get a hold of”, then to look at animation to “see how others have interpreted reality”. This is exactly what I’m going to do next: 

Reference video by ActionVFX 

I first studied this reference of a smoke plume from ActionVFX. You can really see how energy affects the movement of the smoke with the varying speeds at which the smoke rises (very quickly at the start, and then much slower as it gets higher) as well as the rotational movement as the colder air forces the smoke outwards. I think convincingly animating this rotational movement will provide a challenge, as well as how to merge the faster, lower smoke with the slower, higher section.

Matt Luck (no date)

RT-FX (2017)

Finding the right type of smoke (billowing smoke plume) was surprisingly difficult, with most examples being from dust clouds or explosions. This was disappointing as I wanted to really study the way the smoke merges and rotates, however I found these two examples that show this together. Matt Luck’s (no date) animation is a great example of billowing smoke on a slightly smaller scale than what I intend, however it really demonstrates the different speeds of the different sections of smoke. I think because it is a rough animation, the forms get quite confusing when they merge, and there is no clear rotation seen. RT-FX’s (2017) animation fills this gap by very effectively animating that rotational movement. This is done by having the outer edge of the smoke rise quickly while having the shadow line remain at roughly the same place/slightly downwards. I also really like the style of cleanup done, with two-tone lineless rendering. 

Before animating the final version of smoke, I warmed up by tracing Gilland’s (2009) smoke formula in his book, whilst also adding inbetweens. I did this to get a feel for animating that floaty quality of smoke and to help wrap my head around it.

I’m not sure how convincing this is, but it did help me with the use of spheres to represent floating forms. For my actual animation, I think I will have less of a side to side movement.

The structure I took with this animation was to draw the first frame as the starting point, and then animate straight ahead from this. This meant I spent a lot of time on that first drawing making sure there were clear forms, and that it was an appealing drawing. I also wanted to clearly separate the slower, meandering section from the faster one, therefore I drew the former with larger spheres. This also gave the smoke a good sense of scale. From there, I started animating the different forms rising, making sure to have the spacing much farther apart for the lower ones, and the higher ones would barely move.

The main problem I encountered was how exactly to do the rotational movement of the higher section. Initially I had the lines move downwards, however that didn’t look right and I wouldn’t know how to end that. So instead I just focused on conveying different speeds, and had the spacing for the higher section be very close together throughout. This worked a lot better but I don't think it effectively shows the rotation talked about in the theory of billowing smoke. I also started to show how a lower section would merge into the larger section and overcome that, however due to time I didn’t get to properly show that.

Another problem I had was the further I went from the starting drawing, the less appealing the forms were getting. I think this was as I was so focused on what moves exactly where that I focused less on the design principles talked about earlier in this project.

I think to make this effect more appealing - in addition to focusing more on the later shape designs - I could add more in betweens so it’s animated on 1’s. Smoke naturally has a very flowy quality, and a smoother animation would help sell this.

Comparing this animation to my original smoke animation you can clearly see how much the theory and studying benefitted me in creating a more appealing and convincing smoke animation. This animation has a much better sense of form, which describes it in 3D space more effectively and also renders it much easier to look at, as there are no randomly disappearing shapes. In addition to that the overall silhouette of the smoke is considerably more interesting than the almost parallel line of my original. Finally, the original had no varying velocities like this animation has, thus it had no overlap and no contrast - it was too linear.