Finishing loop + transition = weird transition timeline behavior

At the moment, I’m quite convinced that the best workflow (at least for me) with audio segments that need to seamlessly loop or transition somewhere else is by including a reverb tail at the end of the audio assets, and let it play in the transition timeline. Though it is possible to achieve great transitions from or towards the middle of an instrument using FMOD’s transition timelines, depending of the musical content it’s not always perfect. Blending the reverb tail of the source instrument to the beginning of the destination instrument always makes a theoretically perfect transition, in every case. And creating an asset with a reverb tail is easier to achieve (especially in Reaper) and more polyvalent than baking a seamless loop in the DAW.

BUT. I found something, which I first thought was a bug, which I now think is a “feature”, that goes against this workflow. An illustration of what I’m talking about would be something like:

with something like that inside the transition timeline:

In that case, when playing the transition “to A”, instead of playing my reverb tail, it plays… the beginning of the loop! I sure understand that this behavior can have its interest, when using a baked seamless loop, so there will always be something to “feed” into the transition timeline, without having to manually repeat the instrument. However, things becomes weirder when using transition timelines both in the transition and in the loop (which is what I want) ; in that case, the loop still plays instead of the reverb tail BUT without its own transition timeline, which maybe is a technical limitation (impossible to play a transition timeline inside another transition timeline?) but makes no sense conceptually (if a loop has been set with a transition timeline, there’s no point in playing it without).

Could you confirm I’m correct about my assumptions on this behavior?

In fact, to use my reverb tail workflow, I had to remove every loop and replace them with transitions. That’s not a big deal at all, but it’s a bit of a pity FMOD doesn’t handles by default what I assume to be a great workflow.

Here’s the demo of the track where I used this technique, to loop or transition somewhere (without any loop region, as said):

Hi Alcibiade,

You are correct in both cases:

Loops are followed in transition timelines
Transition timelines are not followed in other transition timelines

It was done this way as we were thinking through how we wanted transitions handled while in a transition timeline. We were thinking in particular about jumping between looping music sections.

We knew we wanted user control over when, but wanted to avoid another checkbox option. It seemed to us that there was a reasonable case for following loops in a transition timeline, but not for following other transitions.

So, as you’ve noticed, you can work around it with a normal transition. If the transition is occurring based on a parameter, you could also add the inverse of that condition to the loop region.

That seems reasonable, at least with FMOD 1.10, where transition timelines on loop regions didn’t exist yet. But you didn’t answer to my remark about the loop being followed without it’s own transition timeline, which I find a bit absurd. If it’s not technically possible to play a transition timeline (the loop’s one) in another transition timeline (the transition’s one), then you should bypass this behavior at least in the case where a loop transition does exist. It makes no sense to follow a loop that has been set with a transition timeline if it can’t be played with it.

All this can simply lend to cases where transitions aren’t as smooth as they should, and either we don’t really understand why, or we don’t notice it (by not being careful enough). If we don’t take the time to understand why (as I did), we’ll miss something. I think the doc should make this clearer.

That does seem like a reasonable change for us to make (to not follow loops with transition timelines, only loops without them). I’ll add a task to change that over and to outline the behaviour in our documentation.

1 Like

I just retested that behavior in 2.02.05 and it now seems to have been fixed as I suggested (not looping inside a transition region if the loop has a transition region itself). @tristanjl, can you confirm I’m correct, or did I missed something? I usually check the fix list at each update, but I don’t remember seing this one ; maybe this kind of minor fix isn’t detailed?

Hey Alcibiade, it was fixed in 2.01.12/2.02.03. It’s in the studio tool release notes under

Loop regions with transition timelines are no longer followed in other transition timelines.

1 Like