The Pixel Crush

-------------------------------------------|Digital Animation & Game Criticism|-------------------------------------------

Monday, 28 February 2011


I feel I have left the Pixel Crush slightly too long between updates as I'm now struggling to hold all these things I've been working on in my head at once, which means I will forget half of them as I write, and this blog post will be shorter than it deserves. The way I said that makes it sound like I'm working on things in my head...
Lets move on before this post falls apart completely, brace yourselves again, it's even longer than the last.

The Last Trophy stuff is getting pretty intense, pressure is on from our Kingpin to commence the administration of renderine into the the first scenes so everyone's been hard at work rushing to get a scene ready for final render, unfortunately we couldn't quite catch the Friday that'd been chosen to hit render but we did anyway for the sake of testing. I've been trying to learn two new and illusive techniques which could be very helpful in a rendering pipline but which, so far, have both got the better of me, perhaps I'm not the light after all?
Blimp Before Blur
The first is connected to the second and is motion blur in post. Its such a simple thing to do for real but to fake its a complete pain, requiring far too many compositing nodes and causing all kinds of problems its only the speedy and smooth results that make it worth investigating. I wont go on about it here suffice to say there are issues with creating alpha channels to separate the various elements of the scene that need to be blurred over each other.
Too Many Nodes For A Simple Blur

This is where the second technique comes in and would be the perfect solution if I could get it to work. Contribution maps are what you wish render layers were, you can select parts of a scene and assign them to their own contribution map, here they are rendered to a separate image with their own passes and yet they still have the shadows, indirect lighting and reflections from the entire scene, this is perfect for isolating elements that will need their own alpha channel for motion blur without losing the coherence of a scene's lighting setup. Unfortunately Maya hasn't performed for me as it did in the tutorial I watched thus far but I still have hope, probably just because this would be so bloody useful and I don't want to admit they aren't working.

I also did a bit of modelling and found the use of non-linear deformers most useful when it comes to making bent looking things, top tip courtesy of Alan.

Post production continues to taunt me and Hugh with bad weather for shooting, shoddy tracking and a host of complications that only help to reinforce the fact that we may have over reached ourselves again, but its way I learn best so I refuse to see the error of my ways. Here are some test, failures and a small amount of success. Just not in the right places.

I have been working to put a showreel together as I'm starting to get frustrated by seeing opportunities for internships, placements, jobs come and go without me having anything to apply with and show people. So the last few days have been a narcissistic retrospective of the last years high moments, and turntables, lots of turntables. Hopefully I'll have something to show by the time I'm next writing here.

I feel I should make a name for this section of the blog where I share stuff that has interested me, made me think/feel enough for me to want to share it. I'll have to come up with something...

Pixel Propaganda
Who doesn't love alliteration? This first article is an interview with the game director for Batman: Arkham City, and he says some incredibly reassuring things about the game in terms of its commitment to single player, and the bravery of making choices to better the game, instead of bending to the requests of the audience, sounds like a proper artist to me.

At a loss of things to play in between the mammoth release schedule for games this year I first checked out the Killzone 3 demo. I can tell you its as loud and obnoxious as the 2nd game but prettier and "more varied" in that you shoot people with red glowing eyes in a number of different locations. Then I turned to God of War III, as I hadn't touched it since I completed it nearly a year ago, the intro to that game is so ridiculously massive, in every sense, that it's similarly loud and obnoxious brand of storytelling and gameplay felt very abrasive very quickly.

More recently I checked out the Undergarden demo, which massively contrasted in terms of atmosphere to both the aforementioned games. But it still wasn't what I was looking for, the 2.5D aesthetic and controls where reminiscent the fantastic flOw, and the sound design soothed and delighted, but it was too far down the other end of scale. This lead me to thinking on genre within games, and how little contrast there is to be found within one game, its either a full on first person shooter overdosed on adrenalin, or a chill out game where your heart rate never rises above the speed of  a sedated funeral march. I want contrast in my moment to moment gameplay, contrast breeds drama. I've learned to cherish the quiet moments in Vanquish when the techno music fades out, the explosions cease, and there isn't a grizzled stereotype whining in my ear about how he needs a bigger gun. They are so rare that the contrast is quite a relief, and then its back to business as usual. Such a shiny, Japanese, well designed piece shitty entertainment. I love it.

One of my favourite writers when it comes to game critique is Tom Bissell, author of Extra Lives.
Here he has a showdown with a fellow critic who holds some opposing views. He attempts to classify their two different perspectives as that of a narratologist, and a ludologist. One holds the view that game's potential lies in their ability to craft interactive stories, and the other that game's potential lies in the procedural manner in which a games engine creates unique experiences by design through its systems and mechanics. Both valid and interesting aspects, though I find myself falling in line with the former.

If you saw Black Swan then you'll know that the understated effects were vital to the films themes of mental instability, here is how they were done.

One a totally irrelevant note, everyone should watch the Zeitgeist films, this is the 3rd and most recent. I have struggled with the opening to this one as it shifts its focus from conspiracy and capitalism to genetics, but its essential stuff none the less.

Zeitgeist: Moving Forward

Sunday, 13 February 2011

The Post-Men

Post production looks to be the next exciting project and I've started mudboxing Hugh's island model, so effectively he worked from his design and then I took the base mesh into Mudbox to add cracks, stalactite type details and general rockiness. Amazingly Maya loves working with Mudbox's displacement maps, it exports them in 32bit Open Exr by working out the difference between the base mesh and the sculpting I added through a process of either "raycasting", or "subdivision". Don't ask how either of those work, because I only understand the principles of both.

Floating Island Mk.I

There are some minor issues with the UV's which, while they work, they are the default cube UV's and there are some lines appearing in the displacement where the seams are. I've re-unwrapped it with RoadKill (the wonderful UV unwrapping tool, it uses Blender's open source UV algorithms so its free!) and luckily Mudbox allows you to import new UV's onto the base mesh whilst hopefully maintaining all the sculpting I've already done.

Floating Island Mk.II, fixed UV's
Those hard working animators making the The Last Trophy have been slightly preoccupied with dissertation work recently but I managed to fix the problem with the sky and finally make a decent night lighting set-up in my Monday/Friday sessions. After endless tweaking to make sure the matte backgrounds that I plucked from Google weren't overexposing due to being included in final gather or receiving light from anywhere else I managed to get a simple sky which really completes the image.

Background now shows through but no image yet for night lighting.
The physical sky makes a nice caramel effect for some reason when no image is provided.

This one has the correct colouring, shadows, and background matte.
The shadows no longer fall across the trophies on the wall which I might have to tweak but otherwise this feels close to completion.

 On top of all this we had Double Negative give a talk to us about their role in the visual effects industry (which is pretty integral seeing as they have worked on all these films and are in talks with the various software vendors to get their tools to better meet their needs). Some of the stuff they showed I'd seen on the Inception bonus features but other bits were totally new and it was fascinating hearing it explained by their head of 3D Alex Wuttke, I got to ask why they use Renderman, and whether they've overcome UV's yet and both questions yielded informative and entertaining answers. We even got to chat to him in the studio briefly along with their recruitment manager, it was surreal and humbling.

Me: pointing a gun at something...

Nelson my beast PC has opened new doors into the gaming world and I've finally managed to complete Half Life 2, Episode II. For a four year old game, damn is it pretty, makes me realise what a big deal resolution is when it comes to showcasing a game's art, everything sparkles at 1080p. Also in terms of storytelling I just continue to fall more deeply in love with Half Life's style of unrestricted player movement during key plot sequences. It goes some way to diffusing the conflict between authorial intent and player expression. For example the player can interact with the environment in fairly limited ways in any game, most of all first person shooters, so when the player is constantly pointing a weapon at what's in front of him its hard to create emotional experiences when the protagonist is a mute who points his gun at everything, psychopathic right?

Me: pointing my gun at something else...
Valve attempt to address this by having Gordon Freeman lower his gun every time the cross hair passes in front of a friend or ally. But this doesn't really solve anything because the player can still fire it right through them with no consequence, effectively breaking the illusion of agency the game has worked to build during gameplay. So I found myself switching the to crowbar during the more dialogue heavy parts of the game in order to avoid firing a shot that should blow the characters head off, why is the player not equipped to express himself in non-violent ways besides puzzle solving? This renders all the emotive aspects of the story slightly redundant as they can only flow one way. While Half Life has the decency to be self aware of these things: "you don't talk much do you?" -Alyx Vance, I feel its the one thing missing from an otherwise exceptional interactive experience. Why make a character silent if you're not going to encourage the process of the player using the character's actions as a cypher for their own? Why not create a character as a developer if you're deny the player the right to do so.

My one find recently has been series of blog posts by the founders of Naughty Dog on the making of Crash Bandicoot. It's a great insight into the process of squishing a hugely ambitious game into restrictive hardware, and then knowing when to battle with the corporate machine, and when to submit to it.

Thursday, 10 February 2011

Pixel Crushed

This is all slightly out of date as the blog has been given a rest while I attended to other things. The Animated Exeter promotional trailer is finally done. It was all worth it in the end when the "client", Susannah came down to see what we'd accomplished, and she saw the Cathedral come on screen. I believe her words were: "oh wow". That's job satisfaction :)

It was a real struggle the last few days leading up to completion for me, failed render after failed render, mostly down to human error (my human error) and I felt additionally responsible for pushing the high production aesthetic and, while this obviously created extra work, it really feel it was worth the effort, and it shows in the final trailer. Extra credit to Jake who held it all together and was the only one still able to power on right til the end, apparently you've got to be that hardcore to produce a trailer plus two 3rd year animations all at the same time.. I'll detail the two hiccups we had with Maya so people can avoid similar issues in the future.

I had a displacement node on the grass geometry in order to create detailed blades of grass. Mental ray uses a technique to control displacement quality by judging how close something is the camera and adding to the mesh accordingly, this meant when the camera swooped in close massive spikes of grass appeared and when it drew back and grass in the distance came into view there was little to no displacement. While this is really clever it unfortunately caused a rippling seizure effect in the grass so it was distracting and glitchy looking, apparently this can be fixed by using better sample of the displacement texture but I'd have to do a lot more experimentation to figure it out properly. Hurray for re-rendering. This meant using six computers to render out the fixed grass at 13 minutes a frame, which is pretty respectable considering the fancy final gather lighting, 4k and 2k textures and stained glass window shaders.
 Gorgeous I know, but as a moving sequence, problematic to say the least.

The second problem was with displaying a texture that was an image sequence on the robot's chest, this came out blank on the first render, fortunately we formulated a more elegant solution to this by rendering out the chest separately in the Maya software renderer and using surrounding models to mask the parts of the chest that were supposed to be obscured by using setting them as an alpha channel in after effects.

In my continued exploration of the mental ray render settings I uncovered the motion blur settings, one in particular caught my eye. It controlled the amount you blur something, I convinced myself it was necessary to crank it up to 1.5 in order to help the cathedral feel more like a giant piece of architecture than a miniature scale model. I think it genuinely helped enhance that effect whilst also smoothing out the harsh transition from inside to outside the cathedral, but I would think that wouldn't I...

I've started to work on modelling and UV unwrapping alongside Nigel on The Little Helper, (Pete's lighting work on one of the sets). My first prop being a cuckoo clock, I have some useful information to go on, requirements for its animation and a photo for reference etc, using that I created a very quick concept sketch-see above, and began modelling. While I like the proportions of the thing, at the moment it feels too perfect and CG, I'll have to find out about the style it needs to be in an hopefully look into creating a more wonky, handmade feel to the prop. Other than that just a bird and the pendulum missing before this is ready for me to start unwrapping and then I can hand it to texturing.

Nelson the supercomputer I have just invested £579 in something else I've been devoting a bit of time to recently. Last week has seen me travelling the length and breadth of Falmouth trying to track down the multitude of packages I've managed to miss. Now all the necessary parts have been assembled and I'm highly anticipating the moment I can begin enjoying the new processing power as I render Pixar quality CG in real time...perhaps I should consider lowering my expectations? Nah, its a beast :)

In my continued attempts to open the minds of my readership to new and exciting ways of thinking about videogames I have discovered these two articles. The first talks about the content of the painfully dumb looking Bulletstorm, and what's appropriate/offensive. I personally feel that the medium needs to go through stages of evolution. According to Shatz's genre theory there was the time when a genre was just discovering itself, before it moved into the classic phase (which can last for an extended period of time), after this it develops self awareness and moves into parody and pastiche phase were humour is derived from the audiences knowledge of the genre's conventions. Finally comes the revisionist stage where films in the genre play on the audience's expectations of the conventions to create new ways of presenting the narrative whilst still operating within the format of the genre. Games are still in the stage where designers don't yet know how best to use interactivity to tell stories, this means we are a long way off reaching the satirical stage that Bulletstorm claims to be inhabiting. While there are certainly gaming conventions that are ripe for comedic material, its mostly because they're ridiculous in and of themselves, we don't need another shooter filled with stereotypes, sexism, questionable motives behind gameplay themes. Just because Bulletstorm has self awareness doesn't mean its being clever and witty just by pointing out what's already laughable in a very obvious way. Lets have some progression and then we can laugh about it all later.

Not sure where that all came from.

This piece references some of the finest experiences the contemporary media consumer can have, and on top of that it makes an excellent point about player agency and the personal nature of interactivity.