I made this fully rigged Mercedes AMG 5.5L V8 BiTurbo engine. All the motion is controlled by drivers and constraints. The speed of the engine is controlled by the position of en empty, which can be animated using keyframes.
Because the fire and smoke took a loooong time to render, I only used real volume for the second shot - In all the other shots the fire and smoke effect was made using a couple of alpha-planes with an image-sequence on them.
I am not completely satisfied with color grading, but I haven’t managed to find a better alternative to the blue look - sadly. To be honest, looking at the project now I see so many mistakes and things I should have changed… Buuuut… I am posting it anyway. Time to move on.
Oh, I didn’t know! Thank you for the heads up guys.
Funny… When you look at other animations online almost all of them show a orange-ish combustion.
Yeah, sorry about the noise! Rendering these shots took forever. - And I only used 256 samples.
Yeah, complementary colors always create a good feel. Right as the fuel ignites it is a bit blue-ish, but for a very short moment. It definately needs some fine-tuning.
Thanks! Yeah, I did all of it myself (except the music) - for a personal project. Some of the materials are from my old projects though. Thank you for commenting.
Yeah, I used the simulator and volumetrics to create it, but it took forever to render. So I decided to fake it by rendering the smoke simulation onto a couple of planes with alpha-images. What you see in most of the shots are just planes with emission-shaders and image-sequences mapped on them. Thanks!
Thanks man! Really glad you like it! I didn’t make the music myself, but I found that this track worked really well for the animation. It’s a copyright free track from FreeMusicArchive.org
Nice, as a former engine builder, it looks believable, no reason to knit-pic on the engine, the video is very nice. I personally would be interested in how accurately the valve train timing/phasing is, also how your rig set-up looks like. I have done a bit of this and I know how challenging it was to get right, over all I like it - great job.
@ajcdfin: Really glad to hear that! The timing should be pretty accurate, but I made a mistake on one of the camshafts (which is the first one you see in the render). By mistake I rotated the camshafts a bit too much, so it does not hit then valve-things. I read a bit about different camshafts having different shapes so they open and close the valves at different angles (Don’t know it that makes sense) - But I don’t know the correct profile of the cams for this engine. I just modeled them from reference images.
Don’t have much time atm, but I will post some rig and breakdown images asap. Thank you for the feedback!
So here is the breakdown I promised. I decided to make a short (and quite horrible) video instead because the project is all about animations and movements. Hopefully, some of this makes sense. Feel free to ask in depth questions I am only happy to answer them.
@marckiener: Oh, I think that’s because piston 4 combust at the exact moment piston 2 is at top position. The combustion is not visible and I might have skipped it completely to save memory. Anyway, thank you man!
Gotta say though, the whole empty/sine setup is overly complicated, and i suspect very fragile, at least based on my experience. I did this sort of thing originally when doing mechanical stuff, and if this was on a prop that had to move around, let alone rotate as a “solid” assembly, i suspect this is very easy to break.
After a while i just replicated all that behavior with a single skeleton rig. This is an old and out dated “it’s 3am fuck it” rig demo for my PRR T1. And everything is done with bones. The only drivers in there are to control the equalizer bars (thanks clockmender) and the wheel rotation (Y movement of control bone multiplied by the circumference of each wheel in the same unit), everything else is just IK, track to and other basic constraints. The crank/piston behaviour (or in this case, crank pin and crosshead) is simply an ik handle on the crank, and an Ik chain going from the cocentric center of the rod eyelet to the back of the piston, with another bone in the chain behind that, slightly longer than the maximum stroke set to stretch, and the crosshead/piston bone allowed only to move on the local Y axis. Much simpler than an array of empties and object constraints, and also entirely self contained in the rig object, while still being technically identical in behavior and execution.
So far i’ve yet to break anything on this style of rig. You can flip it upside down, you can roll it over and spin it around and pull bits apart and so long as the transform space on each constraint is correct it’ll never break. And i find it’s simply far simpler to create and work with than empties and object parenting. And all you have to do to animate it is drag it forward, so it’s all very natural. Which for cars is very helpful.
I don’t mean to be a pedant or anything, just as a quality of life and reliability of end result sort of thing i’d suggest doing it this way. Though i don’t know how well it’ll interact with the “Cammed” valve/spring animations, for the rest i can guarantee it’s a better way to go in every aspect.
Everything else is utterly perfect besides though, bravo. I’d really like to know more about how you did the valve timing setup, i’ve yet to make anything work well, and that setup looks sick.