vitroid
Fake water
Simulation of water ripple. It would be useful to hide the tripod and stitch failure.
In the calculation, I put a virtual nadir plain and a virtual zenith plain. Zenith plain contacts with the projection sphere at the zenith, while the distance from center of sphere to nadir plain is variable. In this picture nadir plain is placed just below the center of the sphere. Then upper hemisphere is projected onto the zenith plain and the reflection in the nadir plain is calculated. In the calculation, i put a periodic perturbation to the reflection angle, which produces the ripple. Reflection image is cut in appropriate shape in Photoshop and translucently merged with the original equirectangular image.
See in Immersive viewer how it works.
The pixel at latitude θ' North is projected to θ South in reflection image by the following formula:
L/tan(θ) + (1+L)/tan(θ+δ) = 1/tan θ',
where L is the distance from center of sphere to nadir plain. δ is the perturbation angle and I defined as δ=A sin 2πNx, where x=L/tanθ.
Such a post-processing is possible because orientational information is conserved in equirectangular image. Even in regular retilinear shot, it will be possible if lens FOV is known.
The program is written in C, but it will be easy to be written in mathmap.
Fake water
Simulation of water ripple. It would be useful to hide the tripod and stitch failure.
In the calculation, I put a virtual nadir plain and a virtual zenith plain. Zenith plain contacts with the projection sphere at the zenith, while the distance from center of sphere to nadir plain is variable. In this picture nadir plain is placed just below the center of the sphere. Then upper hemisphere is projected onto the zenith plain and the reflection in the nadir plain is calculated. In the calculation, i put a periodic perturbation to the reflection angle, which produces the ripple. Reflection image is cut in appropriate shape in Photoshop and translucently merged with the original equirectangular image.
See in Immersive viewer how it works.
The pixel at latitude θ' North is projected to θ South in reflection image by the following formula:
L/tan(θ) + (1+L)/tan(θ+δ) = 1/tan θ',
where L is the distance from center of sphere to nadir plain. δ is the perturbation angle and I defined as δ=A sin 2πNx, where x=L/tanθ.
Such a post-processing is possible because orientational information is conserved in equirectangular image. Even in regular retilinear shot, it will be possible if lens FOV is known.
The program is written in C, but it will be easy to be written in mathmap.