EA Studies the Use of Deep Learning for Combat Animations

The latest version of the algorithm was presented at SIGGRAPH 2021.

EA has recently presented their research on how deep learning can be used to generate more natural-looking combat animation for video games. The paper was prepared for this year's SIGGRAPH.

The project focuses on interactively synthesizing novel combinations and variations of character movements interactive which is still quite a complex task that requires tons of manual work. The team described a deep learning framework that allows generating a large variety of martial arts movements in a controllable manner using raw motion capture data.

"Our method imitates animation layering using neural networks with the aim to overcome typical challenges when mixing, blending and editing movements from unaligned motion sources," wrote the team. "The system can be used for offline and online motion generation alike, provides an intuitive interface to integrate with animator workflows, and is relevant for real-time applications such as computer games."

You can learn more about the project on GitHub. Don't forget to join our new Reddit page, our new Telegram channel, follow us on Instagram and Twitter, where we are sharing breakdowns, the latest news, awesome artworks, and more.

Join discussion

Comments 1

Anonymous user
In my opinion this is one of the most over looked aspects of modern gaming. I feel that we the gaming community as a whole are too content with mediocre mechanics for even simple basic human movement. We've come so far in the cosmetics of character and environment design that we've neglected matching the animation. Our characters appear more and more life-like with each new AAA game. The same can be said with simulations in liquids and physics. Yet still the realism is lost every single time the character starts moving.

The best in AAA games always appear to be more in lined with this simple yet complex concept. Games like RDR 2, Last of Us and Ghost of Tshushima seem to know when to push and when to pull on human realism in motion. Let me not forget Death Stranding. Need I say anymore?

The only games that get away with poor animation are usually 1st person or indies. The games that lack mastery of locomotion never do as well as they should have. Still even in indie game much is changing. Take the up coming Little Devil Inside. That appears very simple in regards to triangles but the human movement and interaction novel and poetry in motion. Inside and the ICO series are other really good examples of how powerful small but perfectly time movement really is. It can literally make or break a game.

Interestingly enough I've always felt EA was the biggest culprit. Fallen Order was amazing but could have been an even better game if giving more innovation in the animation department. I strongly feel NBA 2k always had the edge on Live not simply because of more playmaking gimmicks but because of the feel of the game in totality of movement. Realistic iso-motions compared to the somewhat ice skated moments. A las the final chapter of the Live series was so much better in every way than 2k in my opinion. Unfortunately, by that time NBA Live's credibility was pretty much lost to consumers on that particular franchise.

Hopefully this study can lead to a fresh return of the OG Company and its giant in sports. I believe the average players aren't always as vocal as awareness of this isn't as obvious but it makes a vast difference. The average person with zero effort can read much through body language. In my humble opinion if we avoid taking that for granted: I believe the industry will be better for it and the first to master that will dominate. And it won't matter if it's AAA or indie.
0
Anonymous user
·3 years ago·

EA Studies the Use of Deep Learning for Combat Animations

Join discussion

Comments 1

You might also like

We need your consent