• Blogs
• Forums
• Tradeshows
• Partners
• White Papers
• Podcast Show Notes
• Events

 

NaturalMotion Endorphin 2.0

Nov 1, 2005 12:00 PM, By Jordi Bares
The Mill

Realistic Character Animation

While computer graphics is still a very young field, it has moved rapidly into successful arenas in both film and commercials. Our job has sometimes expanded artistic vision and hopefully contributed to making better films. In any case, while CG has made certain projects possible, we have found limitations that create a gap with hyper-realistic animation.

Today, we can build highly realistic environments with modern tools and effects that, together with real footage, make the viewer totally unaware that a story has been interpreted through animation. This is important because the idea is not to change things for the sake of it, but to interpret reality and create a particular vision. Just as painters interpret their subjects, many directors take advantage of technology to tell their story in their particular way. From color correction to altered cityscapes, these effects are now all known and accepted as just more tools for storytelling.

One of the few areas we haven't been able to explore fully is hyper-realistic human motion, mostly because the technology has not been available. Although many advances have been attained in shading and making appearance as realistic as possible, the animation itself has always been confined to extremely long timeframes and exceptionally talented artists, like the ones who gave life to the Gollum character in The Lord of the Rings trilogy.

To that end, we have been toying with many solutions. One of the tools we have found that can create that realism is Endorphin from NaturalMotion. We used Endorphin 2 on Pepsi's “Glue Boy” commercial because it gave us the ability to build highly convincing digital stunts, ones that didn't simply end up with the character as a rag doll.

With Endorphin, you can see tension building up; the character reacts to the environment's forces, from either impact or his own body limits. The actual demands of this project required a natural blend of a real person and a digital character, something that always proves tricky because of the inherent complexity of the human body.

We chose Endorphin because it provides speed and versatility when it comes to creating the exact simulation you want. It gives you more control in a directable way. I needed the simulation to start in a certain way and to end in another certain way; how to arrive and what to do in the middle had to be as realistic as possible. In this particular project, performance was key, which is why Endorphin proved crucial for us.

In this specific case, animating these shots with this degree of sophistication would have been a killer, because it would not only have been slow, but also very painful. Changes would have easily extended the animation work timeline up to a full month. Now, looking back, the decision to use Endorphin was a good one. We did three different over-length (by 500 frames) versions in only two days.

Next, the director gave his input on the animation, changing it to what he felt was the right performance. We did a final version, mixing those different ideas without breaking the continuity from the previous shot and the one to follow. In doing this, we saved a lot of time that could then be invested into other tasks that needed some attention.

In this particular project, we researched actors in the studio, trying to mimic what the digital stunt double would have to do. This aided in lighting measurements, and most importantly, it helped capture the details and actions we wanted to see in the digital stunt. In other words, it was not about falling, but about that particular character falling, i.e., not falling as anyone or anything else. It was about designing a stylized fall.

Endorphin is a tool we must explore to its full potential because the possibilities are huge. So far the most obvious applications are stunts, but we think there is plenty of room to animate with this modern tool.

For more than 100 years we have been animating by keyframing, and although it is a fantastic approach, the truth is there are not many people who are really able to use that method to animate humans with all their unique nuances. Keyframing is something we borrowed from Disney. The technique has allowed us to do a great job — truly great — but we need to move on to a new level. This is where Endorphin comes in by using a biomechanical approach to motion and artificial intelligence for behavior. It gives us that extra bit in terms of achieving realistic human nuance.

Sooner than we think, we could be fully animating certain types of characters. Nowadays there is no way we can do certain things, such as create two characters colliding correctly and trying to protect themselves. It is not that an animator would need much more time to do this, it's that I doubt many could reach the level of sophistication we are aiming for.

Endorphin is directable and simple enough to make animation enjoyable, but I also believe it is a toolset of a higher level, which will add more sophistication to our work. Additionally, NaturalMotion released a free learning edition, making Endorphin more available.

In terms of what gap Endorphin fills for us, it is now pretty clear after looking at the kind of work we are doing with environments and characters, such as crowds that can be fed with Endorphin-generated animation and vice-versa. Also, motion capture can feed Endorphin, as well as other tools like XSI, Maya, and Massive, so we are effectively sharing data at many stages. I believe there is a huge opportunity to raise the level of our work and move toward digital stunts and digital actors.

In terms of what Endorphin technology means for other industries, like gaming, which differs from ours in many areas, I would guess professionals in those fields will be able to cut down on the development of animation assets, use less motion capture, and create more motion synthesis.

Probably in the not-too-distant future, I would dare to say there may be a trend to generate all motion on the fly, based on such AI. It is such an obvious step, considering the next generation of powerful machines like PlayStation 3 and Xbox 360. I expect big developments in Endorphin, and we want to be there.

We are now helping NaturalMotion beta test Endorphin 2.5. (Editor's note: Endorphin 2.5 is now shipping.) From my point of view, the most exciting new feature is the Active Animation event. This allows me to dynamically blend from simulation back into any moving animation or motion capture. What does this mean? Endorphin can be used to create many new types of moves and even multiple simulation/animation transitions. Also, we are able to be more creative through quick turnaround times by using Endorphin.

The software's development is moving very fast. We will continue to make use of Endorphin in upcoming projects. Just wait for more cool content!

---

Jordi Bares is a visual effects supervisor at The Mill, London's Oscar-winning visual effects house. He enjoys working with manufacturers, often helping them test products early in development. Last year, he was first to use Massive on a commercial, Sony Playstation's multiple award-winning ad “Mountain.” To find out more about Bares' work, please go to www.the-mill.co.uk.

Continue the discussion on “Crosstalk” the Millimeter Forum.
© 2012 NewBay Media, LLC.