Some pretext on how the scene is set up. I have a preanimated vehicle rig animated to drive into position. This plays out in the intro, then the animation comes to a stop when the car stops. Its set up as a first person scene with the ability to control the camera. Before 4.9, this scene was pretty much perfect, aside from being empty and unpolished.
Now that i’m nearing the end of the project, I wanted to address this issue ASAP. So what is happening here is that it looks as if the post-processing is going haywire. Ever since I upgraded to 4.9 all the Post Processing felt OFF. Like completely off. Circle Depth of Field was so different my dynamic DoF function needed a lot of adjustments. Now the Circle DoF looks like a noisy mess with sparkling white dots.
Same thing with motion blur. Now in the intro drive, everything is unusually motion blurred, unlike before. Disabling the Motion Blur fixes the problem but now everything looks like it is ALIASED. Seen in the shots below. I also notice some ghosting of the environment during the drive. THIS NEVER HAPPENED BEFORE.
Now when the car comes to a full stop, everything looks perfect. The aliasing is gone. The shading looks correct. No abnormal motion blur.
There’s no going back to a previous engine version anymore as I put in a lot of work in this engine version, on top of utilizing a lot of the features 4.9 offers.
Anyone know whats going on? I’m surprised there aren’t any other complaints with this.
Bump. Anyone? This is such an ugly bug that I just can’t fix one way or another. AA has nothing to do with it. Tried all forms, still getting this ugly Motion Blur. Disabled all Post Process, fixes the problem but now everything is unlit and ugly as all hell.
Well the motion blur artifact occurs even on the smallest amounts. Its fixed by disabling it all together. Which is fine, I can live without the Motion Blur, but sadly disabling it all together makes the object moving aliased even on FXAA. Even the Circle DoF looks much worse in 4.9 even though it was supposedly improved in 4.9