Mailinglist Archives:
Infrared
Panorama
Photo-3D
Tech-3D
Sell-3D
MF3D
|
|
| Notice |
|
This mailinglist archive is frozen since May 2001, i.e. it will stay online but will not be updated.
|
|
[photo-3d] Re: shutter glasses / 3d video workstation
- From: mramstad@xxxxxxxxxx
- Subject: [photo-3d] Re: shutter glasses / 3d video workstation
- Date: Sun, 10 Dec 2000 08:40:27 -0600
Boris,
I've been fantasizing about the same thing for about 3 years now.
I have done experiments with a single Canon Optura digital video camera
with a
homemade beamsplitter (mirror box) and by using motion parallax to
create stereo
pairs. These experiments, although not exactly what you have in mind,
have some
of the same requirements. For example, when using motion parallax to
create
stereo video, you must be able to combine lots of images into pairs
efficiently.
And in both cases, you want to be able to apply stereo editing functions
to lots
of images efficiently as well as general editing functions such as color
editing
and resizing.
These software functions might already be available somewhere for some
price, but
in any case, I decided to write a low cost software package which could
do these
functions rapidly. As you probably guessed by now, this software
package is the
PokeScope Image Manager software which comes with the PokeScope 3d
viewer. This
software probably needs a few additions to reach its full potential for
this
application, but here I will describe how the present features are
useful for
working with stereo digital video. Anyone not interested in a
description of my
software should skip the rest of this message.
Currently, the software doesn't work with any video file formats so you
would need
to use other software to generate sequences of bitmap (or ppm or jpeg)
images on
your disk for both left and right video clips. Your video will probably
be
recorded at 640X480 pixels and 30 fps. Using 24 bit color you'll need
about 1 Meg
per image or 60 Megs per second or 3.5 Gigs per minute of stereo video
pairs. If
you plan to do cutting and splicing, you probably won't have to work
with segments
longer than this. So now the problem is how to work with two sets of
files each
with 1800 images in them. (There is no theoretical limit to how many
images my
software can handle at a time - it only loads one stereo pair at a time
even when
managing large clips (lists) of files.
If the files in the two clips have overlapping names, you might want to
rename
them first. To do this you would select the files of each clip in the
Open File
dialog using wild card characters '*','?',and '#' where the '#'
character matches
any sequence of digits. The Open File dialog also has a default option
to override
the annoying sorting algorithm provided by Microsoft which causes images
to be
sorted like: image1.bmp, image10.bmp image11.bmp ... . In my software,
the images
can be sorted correctly alphabetically and numerically.
The clip is loaded when you close the Open File dialog. Now you can
define a
label for the clip and the first file index number. By selecting the
Rename
function, you can rename the files of all or part of the current clip.
Now, before joining the video clips, you may want to do some color
editing of one
clip in order to balance any systematic differences you observe between
the
clips. To do this you, set the color editing parameters for one image.
You can
advance to any image in the clip to view the result of the editing
parameters.
And you can switch clips and view the source images at any time. When
you are
satisfied, you can save the entire clip with the color editing applied.
Next, you would want to join the two clips into stereo pairs. This can
be done by
loading all the images into a single clip using two pattern matching
strings if
necessary. Next, you would use the auto-pair function which
automatically pairs
images in a clip by a selected offset. You would select the offset to
be the
number of images in the right or left video segment with a possible
correction to
synchronize the left and right sides. Now you can move to any image in
the clip
to view the resulting stereo pairs.
Next you will probably want to crop the images and correct for vertical
offset.
You can also stereo rotate, keystone correct, resize, color edit, and
select any
of the supported stereo formats such as anaglyph, interlaced or
side-by-side.
After setting these parameteres on a single image, you can advance to
any image to
check the result. When you are satisfied, you can save the entire clip
using a
new clip name if you choose.
Currently, the images can be saved in bmp. jpeg, and ppm formats. To
convert them
to a video format, you would need some other software.
So far, I have only worked with clips of several hundred images since I
don't have
a video capture card and I don't have the other software I mentioned
above. I be interested in hearing from anyone who attempts the
procedure above with PokeScope IM software.
Finally, I should mention that the PokeScope Image Manager software only
runs on
the Windows 32 bit platform. My apologies!
Monte
http://www.pokescope.com
Dr Monte Ramstad
Graphic Media Research
211 Ridgecrest Drive
Cannon Falls, MN 55009
USA
Phone 507-263-4611.
Boris wrote:
> But what about digital video (and HD video)?
>
> If you want to build yourself a system to take and edit 3d video, and to
> demonstrate same on the editing workstation itself, what system components
> are needed? Please make recommendations off list if they become too
> technical!
>
> I picture twin DV cams (frame mode capable) for taking the video, twin
> processor Pentium with video capture card + editing software, humongous
> hard drive array, high refresh rate video display card, shutter glass
> system (ideally a shutter screen, where audience of several persons can use
> simple polarized glasses), etc. DVD-RAM drive for cutting demo disks...
>
|