|
This RayBrid
KeyMaker seems to be a commercial implementation of depth
matte for keying.
https://youtu.be/Z5BkEikUrW8
You can see the two depth cameras used on top and under of the
main camera @ 3.26.
Do these solutions support transparency?
On 16-Dec-19 20:14, Phil Spitler wrote:
Totally agree, the Kinect v2 is a great choice with Processing.
A few years ago I wrote an article about using the
Kinect for visual effects.
It might have some good references for you.
Phil
| Phil Spitler  Creative Technologist |
RealSenses thus far have been
very noisy (although the latest one announced looks like
a huge spec bump)...and software support is still fairly
DIY.Â
The Kinect series (including
Kinect v1 clones like the Xtion and Carmine) has the
most mature software support.
If you don't need native
performance, imo Processing and the SimpleOpenNI
library is probably the most flexible solution
around right now.
Wonderful information, thank
you all. Saves days of research!
"Depth Camera" is a new and intriguing
term for me. Diving in deep in that promising
direction. Looks like IntelRealSense is a packaged
product that allows us to get this done fairly
easily with a proven SDK available in a variety of
languages.Â
Cary
check a bit around z mattes , its
often better quality
kinect or similar (better) cameras
, can provide you with the threshold depth you
want.
if you prefer difference matte, it
depends often too much on lighting,
transparency, color.
I've not done it but looked into it
before. Memory is a little foggy, but if
it serves, there was an old After Effects
plugin that could do live difference
keying, I think. Conduit. Not sure if it
still exists or is developed.
Other than Unity, you might also look into
max/msp for this.  (I think that's what
I'd decided I was going to do at the time)
OBS might be one to check also - I don't
think there's any sort of diff matte in
there now, but it might be possible to add
it in as a shader (lots of that code out
there in the googleverse)
Or openFrameworks using C++
(compile in Xcode or Visual Studio)
Not sure what kind of
camera you will be using but this is
where depth cameras can really shine.
Cheers,
Rachel
Unity will do that
or you could use OpenCV in
Processing.
Personally I would
use Processing.
https://github.com/atduskgreg/opencv-processing
Cheers
Phil
| Phil
Spitler  Creative
Technologist |
This may
be a long shot for
this group
I
need to build a
box that pulls a
difference matte
continuously.Â
We'll
setup a camera
in a rock solid
position. Have
complete control
of the
background. Need
to replace it
behind talent in
near real time
for an
installation.
C,
C++, Python,
almost any
language will
work.
Thanks!
CaryÂ
|
