I’m looking at using motion capture-based animation for the supehero for the simple fact that I can stick (presumably) any animation cycle I find on the web onto it as necessary. Ideally all animation cycles will be tweaked to match the “animation style” of the superhero. The reality is that using mocap data is painful and I’ll explain why.
Here is my workflow:
Build Hero model with default mocap friendly rig (bone hierarchy - no constraints - no IK)
Import BVH into scene - creates bone hierarchy and applies animation
Import hero model in scene - character in T-pose with bone hierarchy
Transfer animation from mocap bone hierarchy to superhero bone hierarchy!
The last step has not worked at all even with two rigs (hero + mocap) having similar but different hierarchies and due to the suckiness of Maya Trax transferring animation has been a nightmare.
So I’m toying with the idea of using MotionBuilder. There is a FREE Learning Edition that works but does not export FBX files for you to use in your 3D app and is for non-commercial purposes (i.e learning!)
MotionBuilder has it’s own idea of what a rig should be. It wants a bones named in a certain way or you will have to manually map your bones to their bones.
This requires a few extra steps so the workflow to get mocap animation into a Maya-based character is as follows:
Export Maya character as FBX (bones + geometry)
Import FBX into MotionBuilder
Create character definition for Maya rig - maps maya bones to MotionBuilder bones
Import BVH/AMC (mocap data) into MotionBuilder
Create character definition for mocap rig - maps mocap bones to MotionBuilder bones
Transfer animation from mocap to Maya rig (pretty easy)
Export animation into FBX
Import FBX into Maya and onto the character
The hard part is still transferring the animation because of differences in the rig. Sometimes re-targeting is necessary, but if you have to do it… MotionBuilder seems to be one of the best tools for that task.
At the recent Toronto GameCamp, j2Play presented their SDK/service for self publishing your game on social networking sites using their “universal gamertag” (that’s what I call it)
j2play is the not the publisher - they only provide the framework for you to connect your games to Facebook and other social networking sites. What do they get? A piece(15%) of the advertising revenue in exchange for the infrastructure they provide like Leaderboards, Badges, and Challenges - a great deal if you ask me because they ask for no money upfront. Details here.
As an indie game developer, you can take two viable (let’s assume) directions:
Self-fund and beg artists to work for free. Stay small and the game stays indie.
Sell out and get external funding. Grow to a small crew and the game goes tries to become commercial.
Well I’ve taken the second route, but I must admit the past few months have been as Alexander would say “Terrible, Horrible, No Good, Very Bad” due to the stress and uncertainty When a game requires more than one person to build, it stops progressing when others are not available - which is usually due to lack of funding to pay them or if they are working for free, spare time, because they need normal jobs to pay rent + food. Thankfully, I’ve recently acquired some significant funding.
If I had gone route No.1 like Petri then life would carefree and simple, right? At some point I’d run out of resources(time + money), of course and I’d have to return to the workforce.
I’m very late to the party, but TIGSource had a game competition where developers created a game using a title generated by the “Video Game Name Generator”
If you’ve read Microserfs, you will understand what this post is about. I recently dug up some old e-mails lost on a CD and found my old Jeopardy categories when I was 18/19 (and obsessed with economics and how banking systems work).
Circa 1996 (university-era):
Programming Languages
IPOs and CEOs
Computer animations
The Fuck-ups of the Personal Computer industry
Stock and Currency Markets
Pop psychology
Failures of Economic and Government Policy
Now:
Game development.
3D computer graphics.
Children’s literature post-1980s.
American indie music.
Moder cartoons and artsy-fartsy animated shorts.
NHL hockey post-2006.
Mix tape of useless info: Motorcycles, Traveling, Animation pipelines.
It’s rare to see commercial games deviate from the pixel-art / photo-real mode, but in the past few years you see more and more non-photoreal games with a distinctive art style like The Blob, Team Fortress 2, Paper Mario, Okami, No More Heroes, etc.
Madworld by Capcom (Sin City / line art style):
Two weekends ago, I played Team Fortress 2 for the first time as part of Steam’s free TF2 weekend. I think the global illumination-like lighting and illustrative shaders are amazing.
Menus in AAA games have become as fancy as TV shows.
Colin McRae’s DIRT’s UI is out of this world:
GameTap has a nice 3D UI:
Psychonauts + Katamari + The Simpson’s movie game all have stylish menus.
Another approach is to get rid of menus altogether like Burnout Paradise and GTA IV and throw the player into the game using a montage:
Like films, I believe the first 5-10 minutes sets the tone for the entire game. If you plan the game to be exciting and fast, then throw the player into the game ASAP.
After some back and forth with my modelers about PolyCube maps and how that influences their workflow, I’ve resigned myself to that fact that modelers will be using UVs (even though they hate them) for another 10 years for a couple of reasons:
UV mapping tools are now fantastic + cheap + widely available. See UVlayout, Unwrap3D, Unfold3D, built-in tools in Maya, 3DS Max, XSI, etc.
Smart UV mapping saves tons of texture space by intelligently mirroring / folding UVs - although this causes major problems with normal + light mapping where mirroring UV will cause normals to be flipped.
All modelers are taught how to UV map and have developed extensive UV mapping experience by now.
UV mappers / texture artists are so efficient at using texture space with UVs and understand all the normal + lighting mapping problems now. I doubt they’d be willing to move to a 3D projection paint + PolyCube methodology until we have more texture memory and Zbrush (or apps like it) becomes the standard modeling tool.
