Everything happens for a reason

 It's been a while since anything was posted on here, the main reason being we were off making Chicken Run :Dawn of the Nugget which was pretty much took all our focus.  This break has miraculously worked in our favour though, because in the past few years Unreal Engine has developed enough to allow us to actually animate directly in the engine itself.  

I've always believed that everything happens for a reason, so the old plan of animating the film in Maya and moving the whole thing across into Unreal has been scrapped and replaced with a new shiny plan of doing the whole lot in Unreal.  This sounds much easier and we foresee no stumbling blocks whatsoever going down this route!  

Stay tuned for updates on all the stumbling blocks we encountered...

Momentum

 The hardest thing when working on personal projects is keeping the momentum going. By nature we can only work on these things in our spare time or by grabbing a few hours here and there, so it's inevitable that sometimes they come to a grinding halt. This is now the situation with No Space! We're too busy doing our day jobs, looking after the kids, eating chocolate and all the other things that we get up to, but hopefully, after Christmas it will get going again. Once it does, and it gathers momentum again, we'll hopefully get a good chunk done. 

Getting the character rig into Unreal...attempt/fail 01

So as we've found over the course of trying to make this film, things never run that smoothly. We finally got one of the characters rigged so that it would work in the Unreal Engine and I exported a run cycle that I'd animated to see if it would work. I'd say that it almost made it but the sticking point as always is getting anything that's got visibility keys to work. I think on the next attempt we'll look at hooking the visibility up to the scale so that when anything is hidden, it actually scales down to 0.0 instead.  There were also issues with the squash and stretch on the rig not working but Nathan, our rigger seemed quite pleasantly surprised that it worked as well as it did for the first attempt. There's a screen shot below to see what happened and you can see the inflated head mesh (which should be hidden) is visible, and although you can't see it here, the normal head inside the visor is completely mangled! On the plus side the animation worked on the body apart from, as I said, the squash and stretch on the torso and head. 

 So hopefully when we get the rig tweaked and try it the second time it works a lot better. Now I'm old enough, and ugly enough to know that the chances of it working perfectly the next time are pretty slim but you never know. 

 After-all everything in 2020 has gone perfectly so far....ahem.

 

Made a Zoetrope for fun

 So this has nothing to do with the film other than it has the two main characters in it!  I've always fancied building a Zoetrope, and even though this is a virtual one, I'm still pretty chuffed it actually works.

Animatic teaser

So as mentioned in a previous post, the whole animatic will never go up here to avoid spoiling the phenomenally hilarious and spiritually moving plot......er...... but since it's almost finished we thought we'd post a small section as a bit of a teaser for the film.  Normally on a production the first animatic is done in pencil (or the digital equivalent) from the storyboards, but because this is a personal project and I find it easier to tweak timings and shot composition in 3D, we went straight into Maya.  This also gives you a kind of basic previz too where you can see if the camera moves etc work. 

You will notice that there is no sound on this, and we did the same with Remote, with the thought being, that if it predominantly worked and everything made sense as a film in silent pantomime, then it's pretty solid. The characters and props in this are basic proxy placeholders, but hopefully it's enough to wet your whistle for when the actual film is finished in May.  As in May-be next year, May-be the year after...or May-be the one after that!

New rig animation test

It's been a while since I posted anything on here and that is all down to me being too busy doing the day job. Throughout the Covid-19 lock-down I've been animating flat out but have managed to spend some time in the evenings and weekends animating this test for one of the main characters in the film. As is always the case this only needed to be a short exercise, but ended up being a 33 second beast, and I spent longer than I care to admit on choosing the song he tries to sing to. The plan is to get this shot into Unreal as soon as the engine friendly rig is finished, which hopefully wont be too long, and I will then post those renders here too! 

 I have a few versions of the shot to show from blocking to final just to show the process, but it has to be said that you wouldn't always work like this on a live production where you change parts of the shot significantly in the latter stages.  You'd normally try to nail this down in the blocking stage, but it is fun to tweak things as many times as it needs when it's a personal project to make sure it is as strong as possible. It also highlights how getting feedback is vital to make sure everything reads as intended because the animation process is such a slow one where you spend weeks looking at the same shot, you often can't see the wood for the trees, and need an extra pair of eyes is essential to point out glaring problems which you simply can't see anymore. These issues could be to do with timing, shot clarity or something that you thought is funny...just isn't. My go to man here is Andy Lavery, a colleague who I've done several projects with in the past, and has a very polite way of telling me if stuff is looking crap. 

So here is a block version of the shot which contains all the main key poses that need to be there as well as any breakdown poses that help sell the action. The test had to put the rig through its paces and I tried to push the mouth rig as far as it needed to go. So So with the first pass block the first 13 or so seconds was fine (apart from needing to shift the Karaoke machine slightly to the left to create some negative space between it and the character) but we felt he got happy a bit too quickly as if he already knows what the machine is. It made more sense for him to have a bit of an incling but be slightly unsure until he turns it on. Adding a little tap of the mic confirms what he hoped it would be, and then he gets really excited! This thought process felt a lot more believable than what was there in the first pass. He then tosses the mic from hand to hand before going into the song because he's got his confidence up and nothing it going to spoil his day now. So here's the second pass block... It It all hopefully works better with a more clear thought process from the character. The only part that felt a bit weak was the ending. It was a shame to do all this work and not finish with a punchline so I polished up the shot and added the gun misfire and reaction. Next problem was that the gun firing was lost in his movement and his reaction at the end felt really weak. So after some more tweaking we got this... With there being a bit more of a delay before the gun goes off, it now reads much better and to add to his sense of panic, because lets face it if your head ballooned up to that size you'd be slightly concerned too, he now frantically tries in vain to close the visor which can't get past his huge melon. His reaction at the end though feels a bit weak and makes no sense considering he was panicking a few seconds ago. So it was revised again to get this... This ending felt much stronger, and more in keeping with the character too. So happy days. The sound also needs to be added which we'll get to at a later date but the good news is it reads without it. 

Obviously it all needs rendering, so the next step, once the Unreal Engine compatible rig is finished, is to render the sucker in real-time and see how it all comes together. What could possible go wrong.

Early animation tests

As mentioned in the previous post, the design of the characters has gone through several iterations throughout the production. Below are some of the tests I animated with the character designs that got as far as some sort of rigging. Some had a face rig and some didn't, but we just wanted to see what we could do and play around with a certain style. We also decided to do the animation on 2's, which means that each pose is held for 2 frames, so you have to deal with 12 frames in a second instead of 24. Fast moves still need to be animated on 1's though so that the animation still reads. There are two reasons for this decision, firstly, Remote was animated on singles, and it requires a lot of polishing time to get the animation to look as nice as possible which is very time consuming and the character needs to be kept alive at all times otherwise it feels dead. However when animating on 2's, it is a lot more forgiving and it's completely fine to hold the character still without it looking dead. Secondly, I like animating on 2's! Having worked at Aardman for over 11 years, I've had to animate in all kinds of styles, and nowadays I prefer the look and feel of animation on 2's. It is also quicker than animating on 1's so there's a good chance we can get this film finished before I retire! So here are a few animation tests of the main characters that were done early on to play with the character rigs...







And this was a short test with the re-designed, simpler character. The staging is quite wide on this and there is no explanation in the shot as to why he looks up, gets scared and runs back into his spaceship. We actually made a short AR experience with this where he landed on a tabletop (or wherever you wanted actually) looked up and saw the viewer, then scarpered. The character was only about 150mm tall and it worked really well to be honest and Andrew Lavery did a great job of building the whole thing in Unreal. We might even get back to finishing it one day.



We are currently in the process of rigging the final versions of the main characters, so once that's done I will post the animation tests for those too!

We will leave you with one final animation test done with a secondary character that was rigger brilliantly by Duncan Rudd. The only way we could figure out how to get the extreme mouth shape was to have 2 separate rigs, and swap them out on over the transition, another advantage in animating on 2's is that it's easier to hide the join.

Monster_animTest_01 from Mathew Rees on Vimeo.