Harald Which boolean flag do you mean? Clip End Mix Out Duration is a float parameter, and Don't End with Clip directly above disables the mix-out at clip end to the empty animation.

Sorry I mistyped on the field type. Basically what I've observed is that when the spine clip is the very last thing on the timeline asset, Clip End Mix Out Duration doesn't seem to work regardless of how much duration I provide it. I want the spine animation that ran on spine track 1 to blend back to empty so that the idle animation on spine track 0 can play. Instead, the spine animation on track 1 seems to just hold and the spine rig never goes back to its idle animation and holds still.

Harald If you could send us a minimal Unity project which demonstrates this issue of incorrect playback when you are jumping to fixed positions in the timeline, then we can have a look at what's going wrong and hopefully come up with a bugfix. You can send us your reproduction project as a zip package to [contact@esotericsoftware.com](mailto:contact@esotericsoftware.com), briefly mentioning this forum thread URL so that we know the context.

I'll put one together by end of day today to send over. Thanks for being so responsive!

Harald

I grabbed the newest UPM timeline package and it seems to work now. The spine animation follows the playhead as expected. This is wonderful! Thanks for being so quick on the uptake!

I just want to mention the Clip End Mix Out Duration issue again where the last timeline clip on the spine animation track index doesn't blend back to empty. However, I have a workaround for this at the moment but it might be worth investigating.

Again many thanks for being so quick on addressing the main issue!

Harald

Maybe I'm not understanding the feature as it is designed. So what I thought Clip End Mix Out Duration does is to blend the animation on the spine track index defined in the timeline track back to empty when the clip is done playing.

Example scenario I have a simple timeline asset of a single SpineAnimationStateTrack set to track index 1 and a single SpineAnimationStateClip with a jump animation reference asset assigned. Using the SkeletonAnimation Spineboy gameobject assigned to the playable director's track and idle set as its initial animation, I press play and expect to see the spineboy jump and then fully return to idle. Instead I see the Spineboy do his jump animation then get stuck in the last frame of his jump animation. He still moves his hip bone due to the jump animation not modifying it. So he's in the state where hes stuck in the last frame of the jump animation but his hip continues to move from the original idle animation, which I guess was playing throughout the duration of the jump but isn't noticeable.

However if I then toss an empty signal emitter on the marker track past the last frame of the SpineAnimationStateClip, which extends the duration of the entire timeline asset past the last frame of the clip, I can see the Spineboy empty out his jump animation frames and return fully to idle. So this leads me to believe that in the case where the SpineAnimationStateClip is the last object on the timeline, it doesn't do Clip End Mix Out Duration properly.

I did do a debug breakpoint step through of the code before to try to understand what was going on. What I think was happening is that in SpineAnimationStateMixerBehaviour script when the code hits the OnGraphStop method, isPaused is set to true in the scenario where the clip is the last thing on the timeline asset and thus never fires HandleClipEnd.

I hope this is enough information to try to reproduce the issue I am seeing. Unless I misunderstood the feature entirely, it's something I am seeing on the most simple timeline setup. Thanks again for your time.

Harald

It's cool. I have a "good enough" workaround in the meantime to alleviate the behavior for our project.

I also just noticed a different issue with the latest spine timeline package v4.1.10 that is hard for me to properly explain. It might be a result of the fix I requested for nonlinear traversal but when I have sequential clips in the timeline now (without any clip mixing overlap), the first clip's animation reference is ignored and the one located in the following 2nd clip is used instead. This only happens for the first SpineAnimationStateClip in the sequence as any following 2nd or 3rd or 4th clip seems to behave as normal.

I'll be sending a sample reproduction project built upon my previous one with multiple playable directors and timeline assets in them to show off the behaviors in separate scenes. I've also bundled up the Clip End Mix Out Duration issue in there as well with the workarounds I've discovered in a separate scene.

Thanks again for being so quick and responsive.

Harald BTW: If you enable the clip option Don't pause with Director at your last clip, it will also mix out to the empty animation. Just in case that helps.

AH! Yes, I'll keep that in mind. Thanks for the heads up.

Harald Sorry to hear you've encountered another problem. Thanks in advance for sending the reproduction project!

It's all good. I'm more concerned about other people grabbing the latest version and finding out they can't chain clips together.

Harald

Sorry for the delayed reply. So far in my tests with the new package and our non-linear timeline implementation, it all seems to be working great. I'm able to loop variable amount of times without issue and all the regular timelines are working fine as well. Thanks again for the fast turnaround!

I'll keep an eye out for the Clip End Mix Out Duration whenever you guys get that resolved. In the meantime our workaround is good enough. Many thanks!

Harald

So in my testing reproduction project, I pulled down the latest package to test out the Clip End Mix Out Duration setting and it seems to work as expected now for the final clip, which is great.

I took a look at the logic that was changed and to me it looks a little confusing. Namely line 71 in SpineAnimationStateMixerBehaviour where it says bool isStoppedNotPaused = playable.GetGraph().IsPlaying(); // end of track was reached or graph stopped.

The observed behavior does seem to work as advertised but I just wanted to make sure this is the intended logic. From just reading the code one might assume the logic is reversed but with the way the timelines fire off the event timings, it's probably correct logic but just looks unintuitive.

Again thanks a bunch for the help! I believe all my issues have been resolved.