|
|
Le 24/11/2015 02:10, Mike Horvath a écrit :
> On 11/23/2015 2:17 PM, Le_Forgeron wrote:
>> change cam_dirc to <-1*sqrt(2)/2,-1*sqrt(2)/2,1>
>>
>> The usual factor for x/y is not sind(045), but that's another story.
>
> But sqrt(2)/2 equals sind(045), doesn't it?
the x/y factor for an orthographic camera should reflect the image
ratio, as long as you expect something like square pixel and round circle.
a 640x480 image has a ratio of 3:4, which would be reflected with
#local cam_rgvc = +x * camera_area;
#local cap_upvc = +y * camera_area * 3/4;
In fact, image_height & image_width are usually used instead of hard
coded ratio, when the squareness of pixel is to be kept whatever the
final image ratio. And in the tradition of cinema on tv, the adjustment
is done on the horizontal vector instead of the vertical one: a change
of 3:4 to 16:9 keep the same skyline, but widen or shorten the view
horizontally. (which is somewhat of a problem for the duel scene of The
good, the Bad and the Ugly, in which two actors are at each extremity of
the 2.33 picture... when cutting for 4:3 tv, and still for 16:9 screen,
they get removed)
sqrt(2)/2 is 0.707, which is not a usual ratio of picture, even when
inversed to sqrt(2) (1.414), unless you are doing ISO-216 paper print
(All the A5/A4/A3...B5/B4/B3 series have sqrt(2) ratio, but not the C
serie). But that would be without margin.
computer screens do not have that kind of ratio.
Did I tell that it was another story ?
Post a reply to this message
|
|