Hi there, I have encountered a strange result whilst using RInterpTo. I’ve provided a screenshot of the setup I have, although it’s really just a slight alteration of the BP Input Content Example project.
What I’m experiencing is that the amount of rotation is essentially doubled when the project is packaged when compared to running it in the editor.
I’ve debugged some values which seem useful, so the total rotation when run in the editor on the z-axis is either -16 to 16 degrees when the ship moves either -2000 to 2000 units.
However when the project is packaged the z-axis rotation is -30 to 30 degrees and the ship is still only moving between either -2000 to 2000 units. The - to + value is simply accounting for moving left to right and you start centred on the screen at 0.
This seems very strange as I’m using the same setup as the Epic provided example (The same issue occurs in the Example content if the values are exaggerated) and the calculations are using Delta Seconds so should be frame rate independent etc.
This seems more like a bug with RInterpTo.
Really interested to see what could be causing this.
Hi Rudy, thanks for the quick response.
Happy to try and help narrow this down.
Yes I just made a new project and I get exactly the same behaviour.
The only thing I need to do to test the effects are to attach a print to screen node after the setActorRotation. And print the relevant axis of from GetActorRotation. Then if you play in the editor you get one set of rotation values. If you package the project and play, you should expect a different set of printed values, in my case it’s roughly double. I’m packing to Windows x64.
The new project I made for the clean test is relatively small, I was hoping you wouldn’t mind downloading it and taking a look, that way you can see all of my Blueprints, there is not very much there and definitely nothing fancy or hard to debug to see that nothing strange should be happening.
When i Play from the BluePrint Editor, the game starts to be a bit Laggy and the Player Movement is roughly the Half, as if i click “play” from the Standart Viewport. So different Viewports, make different things…
So for example:
Ive setted up a Player Move on Z with a Value of 30.
Clicking on “Play” from the BP Editor and Playing from the Player Blueprint, the Move is about the Half, as if i click “Play” from the Standart Viewport. Also the Game runs smoother and faster in the Standart Viewport.
Also when the Game is Packed for Win-64bit, the same results as if i would play from the BluePrint Editor.
After further testing it appears that this is working as intended. In parts of your setup (get direction for example) you are multiplying by delta seconds. This means that the editor which usually runs at ~120 fps (unless changed in the project settings) will be around twice as fast as a packaged game (packaged game is capped to ~60 fps). Delta time is the amount of time in between frames. This means that if you have half as many frames that your delta time will be twice as much. I hope that this information helps
Thanks a lot for your quick replies. I am a little confused now though and I hope you don’t mind trying to explain something to me.
I thought that using Delta Seconds was meant to help make things frame rate independent.
I’m going to make up some random numbers here to try and help explain my understanding:
I thought the higher the fps, the lower the value returned by Delta seconds was, so at 120fps it will return 2.5f at 60fps it will return 5f and so on… Therefor meaning that things like rotation and movement will end up at roughly the same value regardless of the the frame rate.
I’d be really grateful if you could shed some light on my misunderstanding.
