- #1

KLoux

- 176

- 1

Hello all,

I have a system that consists of:

I am using the following process:

Steps 2 and 3:

Step 4:

Step 5:

Step 6:

I am assuming that the "small light ray" is small enough that the edges of the points of intersection with the projection surface can be connected by straight lines without great loss of accuracy. I intended to also assume perfect diffusion of light by the projection surface. I believe this is done by including the two cosine terms, but I haven't found examples that go from light source to surface to eye point, so I'm not 100% sure if including two cosine terms is correct. Several sources have stated that for a perfect diffuser, luminance is the same in all directions (implying that maybe I shouldn't need the second cosine term), but the accompanying images seem to show that the direction is important (i.e. different magnitude vectors in different directions), like this image (from https://www.bksv.com/media/doc/18-231.pdf):

I know from measurements on the physical system that I should expect the luminance to be ~10 cd/m^2, but my calculations yield much larger numbers (off by ~3 OOM). I believe the arithmetic is correct, so I assume that the error is in my understanding of luminance. Maybe I'm using the wrong solid angle or projected area?

Thanks,

Kerry

I have a system that consists of:

- A projector with a known position, orientation, and brightness (luminous flux, I guess - units are lumens),
- A lens with known distortion characteristics (i.e. a known light ray vector for every point on the projector's image-generating panel)
- A projection surface with known geometry (not necessarily a planar surface) and assumed reflectivity, R
- An eye point

I am using the following process:

- Calculate luminous flux for the small light ray (##\phi_{V0}##) by dividing the area of the square on the panel by the total panel area and multiplying by projector luminous flux (##\phi_{V0} = \frac{A_0}{A} \phi_V##) in lumens.
- Average the angles of incidence where each of the four light direction vectors intersect the projection surface and evaluate the cosine of this angle (##\cos(\theta_0)##).
- Calculate the luminous flux of the light reflecting off of the projection surface (perpendicular to the surface) by multiplying the quantities above (##\phi_{V1} = \phi_{V0} \cos(\theta_0)## in lumens).
- Average the viewing angles of incidence (vectors from the eye point to ##p_1## through ##p_4##) and evaluate the cosine of this angle (##\cos(\theta_1)##).
- Calculate the solid angle subtended from the eye point toward the area bounded by ##p_1## through ##p_4## (##\Omega##) in steradians.
- Average the vectors from the eye point to ##p_1## through ##p_4##; this average direction vector defines the normal to a plane. Project the area of the projection surface bounded by ##p_1## through ##p_4## onto this plane and compute the area (##A_P##) in m
^{2}. - Compute the luminance as ##L_V = \frac{\phi_{V1} R \cos(\theta_1)}{\Omega A_P} ## in cd/m
^{2}.

Steps 2 and 3:

Step 4:

Step 5:

Step 6:

I am assuming that the "small light ray" is small enough that the edges of the points of intersection with the projection surface can be connected by straight lines without great loss of accuracy. I intended to also assume perfect diffusion of light by the projection surface. I believe this is done by including the two cosine terms, but I haven't found examples that go from light source to surface to eye point, so I'm not 100% sure if including two cosine terms is correct. Several sources have stated that for a perfect diffuser, luminance is the same in all directions (implying that maybe I shouldn't need the second cosine term), but the accompanying images seem to show that the direction is important (i.e. different magnitude vectors in different directions), like this image (from https://www.bksv.com/media/doc/18-231.pdf):

I know from measurements on the physical system that I should expect the luminance to be ~10 cd/m^2, but my calculations yield much larger numbers (off by ~3 OOM). I believe the arithmetic is correct, so I assume that the error is in my understanding of luminance. Maybe I'm using the wrong solid angle or projected area?

Thanks,

Kerry

Last edited: