I have been experimenting with using zengl.camera with fov=0 for a ortho lighting projection matrix, and found that it's much better performance than using numpy or glm - however, it seems like we can only specify size as an integer, whereas with numpy/glm we can specify left/right/top/bottom.
This seems to mean that the corner of the projection area is always centered to my shadow map texture, and needs to be twice as big as I need it in order to cover the whole scene, thus reducing the shadows quality.
It'd be great if it was possible to specify left/right/top/bottom instead of just size.
I have been experimenting with using
zengl.camerawithfov=0for a ortho lighting projection matrix, and found that it's much better performance than using numpy or glm - however, it seems like we can only specifysizeas an integer, whereas with numpy/glm we can specify left/right/top/bottom.This seems to mean that the corner of the projection area is always centered to my shadow map texture, and needs to be twice as big as I need it in order to cover the whole scene, thus reducing the shadows quality.
It'd be great if it was possible to specify left/right/top/bottom instead of just size.