#camera allignment
11 messages in this thread
I am working on an animation for a multimedia presentation, it is a spinning
doo dad that my client would like to superimpose over a the lower right hand
corner of a bunch of text slides.
…………………..
…..
| |
| Text text text
|
| |
| Text …………..|
| | Ani-
|
| Text | mation
|
|………….
|…………. |
This would be easy if I could use the background from their text slides as a
background in 3DS however they would like the doo dad to cast a shadow on the
background of their text slides. The dillema that I am having is this: in
order to get the doo dad to cast a shadow on their slides I need to map their
background on a rectangle behind the doo dad to recieve the shadow. How do I
allign the camera with the mapped rectangle so that when my animation is
superimposed over the text slides it matches up perfectly with the background
of their text slides?
Thanks ME
Max,
>> they would like the doo dad to cast a shadow on the background of their text
slides. <<
if it's just a still of the "doodad" why don't you just use a paint package to
make a drop shadow (white background, black shadow). You would render the
doodad on black with alpha, then in a paint package select the black area,
colour fill with white (or any colour not present in the doodad itself), invert
the selection, colour fill with black – save as shadow.tga. Then use video post
to composite the 3 layers: slides, shadow (keyed on white), doodad (alpha
keyed).
if the doodad is animated then read "3d Studio: Special Effects" for a method
of how to do it. Good luck.
render the doodad against
Martin,
Thanks for the reply. The doo dad is animated and will be superimposed by the
client over the bottom right corner of hundreds of text slides. Therefore it is
critical that the background for the animation be in exact alignment with the
background of the text slides so there is no seam where the edges of the
animation meet the text slide background.
This is similar to the problem I had of flying into an animation playing on a
TV screen several months ago.
It would be nice if one had the option of having an object be able to cast a
shadow on an image used by 3DS as a background.
ME
Max,
>> The doo dad is animated and will be superimposed by the client over the
bottom right corner of hundreds of text slides. <<
this is certainly feasible within 3ds and video post. The settings are
critical. For a full account of the technique which should help you get a hold
of a copy of the book I mentioned (beg, borrow, steal … whatever).
Can you give the client 2 layers? 1 of the doodad without shadow, 2 the shadow
alone, for them to composite themselves. If so, I would render once with the
doodad alone on black with alpha, then with the doodad as a matte object, but
casting shadows (raytraced) on a "shadow catching" object (Stand-in for the
background) with a pure white, 25% self illuminated material. Make the
background pure white. Turn dither 24-bit off. Crank the spotlight to
255,255,255 multiplier 3. That should give you a clean edged shadow (if you
positioned your spot appropriately) on a white background. The client or you
should be able composite the slides, the shadow keying off the white (possibly
making the shadow slightly transparent), and the doodad with alpha key.
Martin,
Thanks mate,
I hadn't thought of rendering the shadow seperately. If it is somewwhat
transparent that would be perfect.
ME
Max,
>> If it is somewwhat transparent that would be perfect. <<
It isn't transparent when you render it, but it can be composited as slightly
transparent if you want – using a mask in video post of less than 255, 255,
255.
<<It isn't transparent when you render it, but it can be composited as slightly
transparent if you want – using a mask in video post of less than 255, 255,
255. >>
Will try what you suggested.
Thanks again
ME
Max, The easiest way to do what you want is such: Create a box in the editor
make it as close to your final intended resolution as possible. Ex: If you're
rendering to PAR create your box about 752×480 the depth doesn't matter. Align
it so it's straight in the Front viewport. Turn on Safe Frame – set up your
render resoultion so it's correct. Now do a zoom window on the box and get it
to fit just right along the yellow line of the safe frame. Now planar map the
box and apply the different slides as maps. You can now throw your doo-dad just
in front of the object. Set up an IFL to change the slides behind the doo-dad.
I;ve used this technique lots of times and it has worked well for me. Good
luck.
Dave Mennenoh
David
Thank you for the advice. If I understand you correctly your method involves
aligning the camera to the rectangle by eye. I am looking for a more precise
way to do this because I would like to render only the lower right hand corner
of the background where the doo dad is being animated over and giving this
animation to the client to suprimpose over the text slides thus avoiding having
to create an IFL for hundreds of text slides and also resulting in an animation
of much shorter length.
ME
Max,
Just another idea (I always try to find a solution that gives me the least
work!):
If the text background that your client is using is a still, would it be
possible for you to ask them if you could modify it (slightly)? I'm thinking
that you could texture map their background onto a flat rectangle and size it
up in the frame so that it matches their original as closely as you can. Then
render a single frame at the same resolution (or higher, I guess) as their
original. Give this back to them to use as their text background and then go
ahead with your animation- this way they match your animation instead of the
other way around.
Don't know if this solution would fit your problem (personally I'd just
composite it in post), but something similar worked for me.
Scott
Scott
Yes, this is a possibility I will consider if there is no way to precisely
align the camera.
Thanks
ME