|
|
Greetings,
I'd like to be able to combine rendered (raytraced?) objects and digital
photographs in a fairly realistic manner. I've tried projecting my images on
a "movie screen" then placing rendered objects in front, but the foreground
looks wierd. Also, if there are reflections from the object, they don't show
what I expect.
Does anyone have experience in doing this? If you reply off-line, (to
pea### [at] nwnet) I'll compile a collection of responses and post them back
here.
Tools that would help this process include:
Cylindrical and hemi-spherical projections (like Apple's quick-time
VR) This way I can project a panorama into a cylinder and
wrap it arount a collection of objects.
One-sided objects (so I can put a bitmap on the front and see thru
them from behind.) This would allow tricks like seeing your
own face in a mirror.
Thanks for your suggestions,
Harold Miller
pea### [at] nwnet
Post a reply to this message
|
|
|
|
Harold Miller wrote:
>
> Greetings,
> I'd like to be able to combine rendered (raytraced?) objects and digital
> photographs in a fairly realistic manner. I've tried projecting my images on
> a "movie screen" then placing rendered objects in front, but the foreground
> looks wierd. Also, if there are reflections from the object, they don't show
> what I expect.
> Does anyone have experience in doing this? If you reply off-line, (to
> pea### [at] nwnet) I'll compile a collection of responses and post them back
> here.
> Tools that would help this process include:
> Cylindrical and hemi-spherical projections (like Apple's quick-time
> VR) This way I can project a panorama into a cylinder and
> wrap it arount a collection of objects.
> One-sided objects (so I can put a bitmap on the front and see thru
> them from behind.) This would allow tricks like seeing your
> own face in a mirror.
> Thanks for your suggestions,
>
> Harold Miller
> pea### [at] nwnet
I've done some of those. Not especially realistic, but I did try to "merge" the
images. Using the realworld-image in a POV-scen will not work I think so I just
cut'n'paste the POV-object into the photograph. To make this look good you have
to make sure the lighting in the POV-scene is more or less the same. Also, use a
background with a color typical to the photograph and use anti-aliasing. I0ve
done some manipulations where I could fit a POV-figure in a real-world
environment, even with some overlaps. A bit of retouching here and there...
But perhaps this is absolutely not what you mean.
Good luck,
Remco
Post a reply to this message
|
|
|
|
I have some experience with this. I strongly suggest rendering your scene
with an alpha channel so that you can subsequently merge it more easily. For
most cases merging the rendered image and the photo is not very difficult. I
suggest PhotoShop 4 or 5, but it is not a cheap tool. As for reflections, in
most cases (if they are simple enough) you can fake them with image mapping.
In PShp 5.0 there'e even a play-around-with-3d-image-wrapping kind of
plug-in, so if you want an orb in a forest (?!?) it comes down to mostly 2-D
processing.
If you are using PShp I can help you more (rather than throwing some
sensless tips at you). Contact me via e-mail if this is the case.
Peter
Post a reply to this message
|
|