• Advertisement
Sign in to follow this  

Skeletal Animation Key Frames: Translation and Rotation separately?

This topic is 2893 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

I'm currently implementing a basic skeletal animation system and wonder how I should organize my keyframes. Until now I always had rotation and translation (no scaling) together in one "key", however I recently stumbled over an animation where those are treated separately. Now I am unsure whether I should adopt that approach as well, or rather merge the separate values together. What do you recommend? Keeping translation and rotation separate is a bit more flexible, but seems to be an wasted effort often. Than again, its not that complicated. Any opinions?

Share this post


Link to post
Share on other sites
Advertisement
In our engine we store rotation, scale and translation separately. We do this for a number of reasons. We quantize each type of value differently based on different joints, sometimes completely removing scale or translation if only rotation needs to be there for the animation. Since the data is variable size per node in the hierarchy, it's impossible to store a fixed size key that contains each piece of data; instead we store a stream for each type, be it rotation, etc. Furthermore, our system interpolates between keys, so having them next to each other in memory is more cache friendly.

Best of luck.

Share this post


Link to post
Share on other sites
Thanks for the nice explanation. Sounds like a good advice!

Share this post


Link to post
Share on other sites
Sign in to follow this  

  • Advertisement