I know it doesn't seem to be an API question (and it's actually not completly one) but I thought developers may have figured this one out. I basically have no problem finding out what Maya does to compute the screen coordinates of a viewing frustum (basically what getRenderFrustum does).
I am more puzzled by what's happening when one changes the Fit Resolution from lets Fill to Overscan. Maya scales the screen coordinates (at least for me) by the ratio 1.77 / (film aperture ratio). I have 0 idea where it gets the 1.77 thing from ? Maya doc says :
"kHorizontalFilmFit In this case the digital image is made to fit the film back exactly in the horizontal direction. This then gives each pixel a horizontal size = (film back width) / (horizontal resolution). The pixel height is then = (pixel width) / (pixel aspect ratio). Now that the pixel has a size, resolution gives us a complete image. That image will match the film back exactly in width. It will almost never match in height, either being too tall or too short. By playing with the numbers you can get it pretty close though."
The problem is that I have no idea what they are talking about 
Where is the horizontal resolution coming from ? The result has actually nothing to do with the image resolution in the render globals. If you change those settings (image resolution) it wont do anything to the Film Gate !!! The only thing that changes the film gate is when you modify the film aperture...
I am completly lost so any clever input from a clever person would greatly appreciated.
Thanks a lot,
