Use normalized direction vector or not?

[qhdqhd]

AS you said: in

mipnerf/internal/datasets.py

Line 193 in 84c969e

# Distance from each unit-norm direction vector to its x-axis neighbor.

BUT you use directions (not-norm) rather than viewdirs （unit-norm）:

mipnerf/internal/datasets.py

Lines 187 to 195 in 84c969e

	directions = ((camera_dirs[None, ..., None, :] *
	self.camtoworlds[:, None, None, :3, :3]).sum(axis=-1))
	origins = np.broadcast_to(self.camtoworlds[:, None, None, :3, -1],
	directions.shape)
	viewdirs = directions / np.linalg.norm(directions, axis=-1, keepdims=True)
	# Distance from each unit-norm direction vector to its x-axis neighbor.
	dx = np.sqrt(
	np.sum((directions[:, :-1, :, :] - directions[:, 1:, :, :])**2, -1))

This is different,right?

mipnerf/internal/datasets.py

Lines 194 to 200 in 84c969e

	dx = np.sqrt(
	np.sum((directions[:, :-1, :, :] - directions[:, 1:, :, :])**2, -1))
	dx = np.concatenate([dx, dx[:, -2:-1, :]], 1)
	# Cut the distance in half, and then round it out so that it's
	# halfway between inscribed by / circumscribed about the pixel.
	radii = dx[..., None] * 2 / np.sqrt(12)

If you use directions (not-norm) ,dx (or radii) will be the same for different pixel rays in the image .
If you use viewdirs （unit-norm）, dx (or radii) will be smaller for pixel rays away from image center, and bigger for pixel rays around image center.

[qhdqhd]

@jonbarron Help me please,thank you!