#Active People?
8 messages in this thread
Hello. We (Blanchard Hower Associates) are creating an interior room for a
client who wants to show the space as being very busy and active. To
accomplish this we are creating the room in 3D Studio, and then pasting
2-Dimensional images of people onto this image in Adobe Photoshop. Problem:
the only people on file we have are very "motionless." Meaning, they are not
showing any action. We were wondering if you can get us some names of
companies that sells scanned photographs of architectual components in
perspective (i.e. cars, people, plants) form. These images need to be ready to
be easily cut (on black background) and pasted to show a photorealistic look
(fairly high resolution).
Thanks,
Nick Hower
P.S. We are not really interested in using People for People for this
particular project.
Nicholas,
<< …pasting 2-Dimensional images of people onto this image….>>
I can refer you to MMRG5-.TXT in the image and info section of the lib.
jonas[adesk]
Jonas,
I'm not sure how to retrieve or even get to this .txt file. Could you pretty
please tell me how to do this. Thank you, and sorry for my ignorance. Nick
Hower
Nicolas
<< where's the txt file… >>
When reading this, you are in the message area of the ASOFT forum. There is an
associated library broken down into sections that accompanies this message
area. The section that you need is called Index & Info. When using your
software, you should be able to "Browse" or "Search" the library. If you
seached on a keyword "GUIDE", two files would be displayed. You have the chance
to mark files for downloading. You want to "download" the MMRG5-.txt file from
the Index and Info section.
All navigational software should provide tools to accomplish the above.
jonas[adesk]
Just out of curiosity, how do you plan to match perspective with these
clip-people when you find them?
— James E. Coulter — Mindscape Animation Productions —
James,
In response to your inquiry:
We know from our design/architectual backgrounds that all people (with certain
exceptions) in a perspective scene will have their heads or eyes at roughly the
5'6" imaginary horizontal line. If some people are scaled up or down, their
heads must still remain at that invisible 5'6" line. Grade changes must be
accounted for and applied accordingly. That puts people in perspective, but
still does not correct angle situations. If we have a large selection of
different images to choose from, we can pick the most closely matched person or
image that will fit into the scene (i.e. position, lighting, etc.). Another,
more time consuming way, is to actually have these images mapped on invisible
rectangles and then positioned as actual 3-D objects in the scene. This
requires having the same image as an all-white image mapped with the actual
photo. Its more time costly I think, but it is more acurate. if you use
ray-traced shadows, you will have accurate shadows too.
I hope that helps, but write again with any other Q's.
Nick
Thanks for the insight on the 5'6" perspective. I was thinking that
the range of camera lenses, combined with different lighting and
shooting angles, would make matching the images difficult.
I matched a photo of a building to a 3D model I built for an
architectural job. To get the angles, FOV, and lighting to match for
a seamless transition took amost eight solid hours of trial and
error. I did have the coordinates of the camera, but they only
served as a starting point.
I suppose in your case, if you have a sufficiently large set of images
to sample, it should not be too hard to find a match — just time
consuming.
Nick,
I use clip people in my renderings too and they work just fine, even if you
can't match exactly the perspective. I have only a few people that are also
very expressionless. If you ever find clip people to sell or acquire, please
let me know. I'd be interested in them too.
Thanks,
Eni