There's at least two possible sources for changes in shape in the gimbal video:One question. In two of Micks videos where he demonstrated the glare effect, the glare-shape does not change at all. Although Mick ist changing distance, is changing orientation of the light source and even he lets it rotate.
Shouldnˋt it be in the case of the gimbal-thing as well? In the gimbal video, the shape is changing.
1. The image processing filters applied to the image seem different in black hot vs white hot mode.
2. Whatever's producing the glare, presumably a jet engine, is being seen at a different angle. So the shape of the source changes, which means that the shape of the glare (which is given by a convolution of the source with the point-spread function of the optical system) will also change.
This is very rough and dirty and is not intended as representative of what's happening in the gimbal video (in particular, I have made no attempt to match apparent sizes to plausible jet engines or whatever), but it is illustrative of the concept.
import numpy as np import imageio import scipy import scipy.signal import matplotlib.pyplot as plt round = imageio.imread("round.png", as_gray=True) gibbous = imageio.imread("gibb2.png", as_gray=True) kern = imageio.imread("kern2.png", as_gray=True) convr = scipy.signal.convolve2d(round, kern, mode='full') convg = scipy.signal.convolve2d(gibbous, kern, mode='full') convr /= np.max(convr) convg /= np.max(convg) fig, ax = plt.subplots(2, 3) ax[0,0].imshow(round) ax[0,1].imshow(kern) ax[0,2].imshow(convr) ax[0,2].contour(convr, levels=[0.15],colors='r') ax[1,0].imshow(gibbous) ax[1,1].imshow(kern) ax[1,2].imshow(convg) ax[1,2].contour(convg, levels=[0.15], colors='r') plt.show()