Precision of units' time scale

[Rhazarian]

Hey all.

I've been researching some things regarding animations and time scaling, and I've found that time scale is very unprecise. Try starting the same animation on two units, with first unit's time scale equal to 1 second unit's time scale equal to 0.5. Stop the first unit's animation after a second (by setting the timescale to zero) and the second unit's animation after two seconds. You expect to see the units in the very same position, but they are not. (Easy to check by placing them in the same point with different tinting colors)

What's more, I tried to do similar tests with, e.g. 0.5 time scale and 0.53 time scale and... they are the same time scale. The units stop in the same position. However, if you try to test 0.53 and 0.54 time scales, you will see the units finish in different positions. I've found that similar time scales are going in blocks. Like all time scales in [0.5, 0.53] are the same, and all time scales in [0.54, 0.57] are the same, however 0.53 and 0.54 do differ a lot.

Does anyone have more details regarding time scales implementation and what are the REAL values for each interval? (Since as I mentioned above, 0.5 time scale does not slower exactly in 2 times, it's something other...)

 

[Planetary]

Seems like odd behavior. I glanced at some code here Reverse Animation - Wc3C.net which runs the animation manually via an offset timer. Could be a solution?

 

[FeelsGoodMan]

I think it's affected by the units blend time, and what animation it was currently in before you play the animation you're testing the time scale of. Try resetting both units' animations prior to playing the preferred animation.