In reference to the post:
"The image is distorted vertically by 1/sqrt(2) and horizontally by sqrt(2). This effectively makes a square (1:1) come out (1:2) - i.e. is stretched in the horizontal direction by a factor of two relative to the vertical. What you need to do is generate an image that is "pre-distorted" by a factor of 2 vertical direction. In other words make the "squares" in the image you feed to the DLP4500 twice as high as they are wide. When you feed this in, the result will be an image whose squares are (1:1) - i.e. really are squares."
If I want to create a square with a specific dimension when projected onto a screen, I then need to pre-scale the image in both horizontal and vertical direction right? Scaling by a factor of 2 in vertical direction will only make the squares (1:1) but the overall dimension is still not what is expected? Just to confirm.
Thanks
Saad