You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, the vertex data (which contains global positions) gets updated only when the scene's stats become "outdated" (i.e. when an object is added or removed). This means for most scenes, objects that transform (translate, rotate, scale) don't appear as transformed in the frame buffer, because their vertex data is not updated.
This could be temporarily solved by uploading vertex data for every frame, but due to the way the frame buffer view works currently (each pixel accumulates colors over time and get averaged out by the frame count), the transformations appear as motion blur and eventually get so averaged out that they are not visible.
Some ways we could solve this moving forward:
Uploading the vertex data every frame has performance implications. We can instead switch to a buffer of model matrices as proposed in Feature: Model matrices #4.
Reset the frame count back to 0 when the matrix buffer is updated.
The text was updated successfully, but these errors were encountered:
Currently, the vertex data (which contains global positions) gets updated only when the scene's stats become "outdated" (i.e. when an object is added or removed). This means for most scenes, objects that transform (translate, rotate, scale) don't appear as transformed in the frame buffer, because their vertex data is not updated.
This could be temporarily solved by uploading vertex data for every frame, but due to the way the frame buffer view works currently (each pixel accumulates colors over time and get averaged out by the frame count), the transformations appear as motion blur and eventually get so averaged out that they are not visible.
Some ways we could solve this moving forward:
The text was updated successfully, but these errors were encountered: