Reply To: Ryanocerous’ Polybook (Feedback and advice welcomed)

[ryanocerous777]

It’s been a while, but I finally finished one big project I’ve been working on. Please see the video embedded below. I say “finished”, but I’m still very much open to constructive criticism or ideas on how I could have done something better/more efficient.

Most of the models I created myself, with the exception of the shells, cat, and some of the greebles on the spaceships. The planets I assembled from photos of real planets that I then edited to be more alien.

Regardless, I learned a LOT doing this project. Most notably, I learned a lot about the NLA editor. While exceedingly powerful, it has to be the most unintuitive part of Blender I’ve encountered by a long shot. It’s also real hard to find quality tutorial videos on it.

It’s real subtle in this one, but I also learned a good bit about using surface imperfections. Even now I’m seeing I could have used them to greater effect, but I can also see the differences in realism they made even with the small amounts that I did end up using.

@waylow do you have any advice on how to set up NLA strips in the editor? Like is there a convention or set of conventions to follow? I’ve seen a couple people have it set up so that NLA strips are dedicated to each “body part” of an object. I’ve seen other people that have them roughly correlated to actions.

When using the NLA editor, I made it so that most tracks were added on to the one below, and were more like “poses” than movements. However, some strips, like head bobbing while licking paw, were indeed motions. I then just blended in and out of these “pose strips” using the blend setting and ample use of additive extrapolation. However, this strategy meant that it was a little harder to use force curve modifiers effectively. I got around this somewhat by putting two sets of keyframes of the pose position, at the beginning and end of the strip, and then f-curving between them. However, that got messy and relying solely on blending between strips to transition between actions limited how much control I had over the animation. I don’t know if that makes sense. In the future, the additive method wasn’t so bad of an idea, but it would probably be best to break the transition motions into their own actions. Someone please correct me if I’m wrong.

I see a lot of areas I’d like to improve (e.g animation realism, the scale of the spaceships, more surface imperfections, etc), but overall I’m happy with how this turned out given my current level of experience.