Этот сайт лучше всего просматривать в современном браузере с включённым JavaScript.

How to get Image Sequences to use the same transforms?

AxiomVerge

This is hopefully a simple question since I was able to do this on another file, but now I can't remember how I did it. I'm using
3.7.15 beta.

Basically, in the image below, I want to put a sequence of images under "HeadSlot" and have them match up.

So, I select them in hierarchy view and drag them to HeadSlot. Below you can see the results with the first image in the sequence selected. It defaults to rotated 90 degrees (to counteract the bone I presume) and centered on the Headbone.

However - this might be a bug - not all images come in with the same transform, as the last image in the sequence is rotated 270 degrees, even though it appears upgright. If i change it to 90 degrees, it becomes upside down. I could get around this by using world transforms only but I can see how it might cause an issue if for whatever reason you needed them all to have the same local orientation.

Furthermore, I want to have the images shifted upward so that the Headbone is at their base. But when I transform one of the images, the others don't follow. In this case I only have 8 images but there could potentially be hundreds; what is the preferred way to move an entire image sequence so that they stay aligned?

AxiomVerge

Erika

The quickest way I do it is by parenting all these images to a dummy bone created for the occasion, manipulate the bone to get them right where you need them, with the orientation you want etc. then parent the images to the correct bone and delete the dummy bone. Otherwise you'd have to copy-paste the transforms for each image, which is a long process in cases such as yours!

AxiomVerge

Erikari написал
The quickest way I do it is by parenting all these images to a dummy bone created for the occasion, manipulate the bone to get them right where you need them, with the orientation you want etc. then parent the images to the correct bone and delete the dummy bone. Otherwise you'd have to copy-paste the transforms for each image, which is a long process in cases such as yours!

Okay, that works.

Now that I think of it, is there any kind of file referencing in Spine? It would be cool if somehow image sequences could be grouped into a separate file and then in the parent file you key what Animation you want them to play as opposed to picking the individual frames. You could also use this for bone animations if a character is really complicated.

Pharan

Yep. The idea for a separate skeleton animation attached to another one is called SkeletonAttachment. It's still in the backburner.

If you meant another independently playing animation on a character (eg, your character is holding a pinwheel and you want it to keep spinning in a loop regardless of the action), you can always do that.
Spine's runtime AnimationState is a multi-track animation playback system. You play your main actions on track 0, and play the pinwheelspinning animation track 1.

Playing an "animation" within another animation is another thing we've talked about. It sort of figures into a Pose system where you're sort of keying "blend shapes". But instead of being restricted to mesh blendshapes, you have a "pose" comprised of keyed values like bone positions, rotations, slot attachments, and mesh deformation keys. But also rather than just blending between discrete poses, you scrub parts of an animation. So "poses" are just points in a separate animation's dopesheet.

Image sequences are also another thing. Nate originally had the idea for a RegionSequence attachment. But we also noticed that people like to map image sequences onto meshes so they can be weighted and deformed. We hope to incorporate that somehow. The image sequences on meshes thing is still currently possible using linked meshes though, but you have to key the frames manually.