|
![](/i/fill.gif) |
Le 21/08/2013 15:42, clipka a écrit :
> Am 21.08.2013 15:24, schrieb Bald Eagle:
>>
>>> I have no idea what look_at does when you apply rotations and
>>> translations first. "Something unexpected" would be my first guess. (I'm
>>> actually quite surprised it doesn't give you an error message.)
>>
>> I assumed that since I applied those transformations BEFORE the
>> look_at part,
>> that it would change the camera position, and then look_at would be
>> invoked to
>> ..... look at.
>
> It may give you some consolation that your assumption was perfectly
> reasonable. POV-Ray's camera statement, unfortunately, is not.
>
And IIRC the code, it is even sensible to the actual version of the SDL
when the camera is parsed (as well as the type of camera and other side
effect too).
Just play it safe: transform the location vector beforehand, then keep
the camera simple (with or without lookat)
--
Just because nobody complains does not mean all parachutes are perfect.
Post a reply to this message
|
![](/i/fill.gif) |