@ZkouskaMikrofonu You probably should not be using a timeline for this to start with. You have the Action Value pin on your input - it should be set to an axis and produce a float or a swizzled vector, or even a world-transformed ready-to-go vector (check input modifiers). And then be multiplied by frame delta.
I am making game, where AD gives a rotation to a motorbike.
If you use a TL here, you tap a key once then wait, let’s say 5s, during which the bike turns automatically without further input - is that what you want? Or is this real-time? The longer we hold key, the more we lean & bank?
Timelines work with time - you play it and it runs for x seconds. It is frame-rate independent but you can make it frame-rate dependent by plugging something into the Update that does not belong there (or you’re making something funky). The TL’s Update fires every frame - the more frames, the more updates. The real question is what are you doing with the data during the update.
Timelines work best with Lerps by producing Alpha which is then sampled during the Update, the more frames, the more steps, but the same end result.
- this will run pretty much the same at 10 and 100 fps:
Providing the track spits out a 0-1 range, it will rotate 90 deg but it will be 10x as fluid at 100 fps.
- this will produce dramatically inconsistent results at different frame-rates:
Delta time will be needed here.