St. wrote:
> "John VanSickle" <evi### [at] hotmailcom> wrote in message 

>> Shot B was rendered at 320x240, shot C was rendered at 40x30,
> 
>    40x30 ?? Why that size? Speed?

No, to produce the blockyness.  I used mega anti-aliasing and rendered 
the exact same scene at lower resolution; when scaled back up to 320x240 
in the masking render, it produced a shot in which each 8x8 block of 
pixels was a solid square colored to match the average of the same block 
of pixels from Shot B.

>>The only thing to remember is that the objects which are present in both 
>>the B and C shots MUST be positioned identically, relative to the camera, 
>>and if the masking is based on the objects in the scene, then the mask 
>>object must properly fit the objects in the scene, usually by having the 
>>mask objects positioned, relative to the camera, in the same way.
> 
>    Phew! I was going to mention in my first reply that 'masks' were possibly 
> used, but thought against it because it looked 'too neat'. Good job with 
> what you've done there John.

It's actually quite easy to make sure that everything is positioned 
right, relative to the camera; simply cut and paste the scene code for 
the positioning of the camera and the objects.

It was slightly trickier with the scene above, because for shots A and B 
the models and the camera are positioned exactly, but for shot C the 
scene required that the robots be turned 90 degrees clockwise and 
translated by <-5000,-10,15500>; but the transform only had to be 
applied in four places in the .INC file, so it wasn't that hard.

>>Once you get that going right, the technique is really quite easy.
> 
>     I counted 36 animations on your website. I think I'll leave the 
> animations to you... Let's see double that number.   ;)

I'm working on it.  My next one is already up to 1440 frames (a full 
minute), and it will probably be closer to two minutes when I'm done.

Regards,
John

Post a reply to this message