function [mov] = make_movie(eyes, foc, asize, bsize)
	%This function makes a movie of the cereal
	%box as it's view from the multiple centers
	%of projection provided by eyes. The user 
	%can also specify resolution and focal 
	%length.
cereal_box = [-5 -5 -5 -5 -5 -5 -5 -5 -5 3 3 3 3 3 3 11 11 11 11 11 11 11 11 11;
-15 -15 -15 0 0 0 15 15 15 -15 -15 -15 15 15 15 -15 -15 -15 0 0 0 15 15 15;
-20 0 20 -20 0 20 -20 0 20 -20 0 20 -20 0 20 -20 0 20 -20 0 20 -20 0 20];
set(gca,'nextplot','replacechildren');
aviobj = avifile('cool15.avi');
aviobj.Quality = 100;
isize = size(eyes);
I = zeros(bsize,asize);
for j= 1:isize(2);
    ceye = eyes(1:3,j);
    	%ceye(3) = -ceye(3);
    	%I had to do this, until I started saving to
    	%an avi file.  I think it has to do with how
    	%the avi object views the frame verses how 
    	%the figure plotter views it.
    [P,foc_v] = Pifw(ceye',foc,asize,bsize);
    for i = 1:24
        curpent = cereal_box(1:3,i);
        curpent = curpent - ceye;
        if dot(curpent,foc_v) > 0
            dotty(1:3,i) = curpent;
        end
    end
    dots = P*dotty;
    	%I do kinda clipping.  An easy fix is to
    	%dot the potential point with the foc_v 
    	%from Pifw and only draw if the result is 
    	%positive. As per Roberts hidden-surface
    	%algorithm, that mean the potential point
    	%is on the same side of the plane through
    	%cop as the vector normal to cop.
    I = Splat(I, dots);
    	%max(max(I))
    imagesc(I, [0 240]); colormap(gray);
	%As discussed in Splat: in order to prevent
	%the dimming of most dots when others 
	%intersect on the screen, I have to clamp
	%the grayscale domain on the image, so that
	%matlab doesn't do its own scaling to fit
	%the grayscale values. (It's the [0 240]
	%above.)
    mov(j) = getframe;
    aviobj = addframe(aviobj,mov(j));
end
aviobj = close(aviobj);
movie(mov)