#Stereogram Generator
11 messages in this thread
Are there people out there using 3DS to model scenes that are to be converted
to a Stereogram ('magic eye image' / '3D image') ?
If so, are you interested in an IPAS process creating such an image from your
3D Editor scene ?
Just asking, because I am close to finishing that and I am curious how much
interest there is.
Boran, uniVirtual pictures
I've noticed that many of the images in the "Magic Eye" series were created in
3DS. Do you have a routine that produces an RDS of similar quality?
– G
>>Stereogram
Are you talking about random dot stereograms or something else? I use 3DS to
create primarily vertical interval holographic stereograms and have that pretty
well worked out . . . although the ability to offset the rendering viewport
would simiplify this. . . ( Gary? )
Paul -Rising Star Graphics
Paul,
I am creating SIS with my routine.
How do your >>vertical interval holographic stereograms<< work ? Can I find
them somewhere ?
Boran (uniVirtual pictures)
>>vertical interval holographic stereograms
Boran,
Sorry, I've been out of touch -I just got back from the NAB ( National
Association of Broadcasters ) convention.
You are not likely to have seen much of my 3DStudio generated holographic work.
I've done a couple of in-house image pieces for myself and my holographer, a
few tradeshow booth pieces for DuPont and a boiler manufacturer ( incorporating
a ring of 36 IPAS Flame jets ) , etc.
You are more likely to have seen a series of dance images distributed by the
holographer that I work with. These are marketed under the name "Holographic
Moments". These are mostly 'live' images captured with a patented multiple
camera system. I developed various techniques for processing the images as
well as compositing with 2D and 3D elements. Lately we've been incorporting
more and better 3D stages for the live models. Next week Inside Sports magazine
will sport a Michael Jordan hologram we produced literally overnite! This is
mostly 2-1/2D animation on 3 planes. The number 23 on his jersey crossfades to
45, he dribbles the ball down and forward of the frame and he floats in front
of the Chicago skyline proclaiming the 'Second Coming" <G>. This was all
pretty routine except for having to reconstruct the hand that was hidden behind
the ball.
In talking with my holographer and scanning other messages in the thread,
neither one of us is sure what you mean by SIS. What's that?
Paul -Rising Star Graphics
What is a "vertical interval holographic stereogram?"
JKJ
John,
Between NAB and Easter my correspondance has lapsed terribly, sorry!
>>What is a "vertical interval holographic stereogram?"
That's the mantra I've learned to say over the last 12 years in the presence of
holographers when I'm trying not to be dismissed as an ignorant peon.<G>
Sometimes it works.
Basicly it discribes the technique of encoding a series of images in vertical
stripes on a holographic plate that are then projected onto a 2nd plate to
produce a white light visible image. The series of images can represent
various stereo views of a scene or animation over time. If you are carefull
you can present animation without distroying the stereo effect. These are the
reflective mylar embossed images that you often see on packaging, baseball
cards, and magazine covers.
-Paul
>>SIRDS IPAS…
I think there would be a moderate interest if the routine were fairly complete,
meaning support for more than just random dots. Will your routine handle bitmap
background strips?
I hope this is not discouraging, but I'm afraid such a routine might not sell
too well unless it was fairly inexpensive, since free programs like RDSGEN make
it easy to play with SIRDS.
If your program supported 24-bit output it would be useful, since RDSGEN is
8-bit only. Also, generation directly from the 3DS z-buffer, on-screen
previews, animation capability, and enhanced depth resolution might all help
make an IPAS worth considering.
JKJ
John,
thank you for your rich input.
it was my fault not be be more specific in my first message. All the points
you mentioned are handled by my routine:
>>SIRDS / texture<<
What I am exactly generating is Single-Image-Stereograms (SISs) producing the
textured type. I dont support Random-Dots.
>>24-bit support<<
Stereo.IXP will support any filetype 3DS supports, both for input and output.
>>z-buffer and enhanced depth resolution<<
a first version was using a 'fogged' frame input, thus using grey-scale 8-bit
resolution. I am upgrading to use 3DS4's z-buffer, that passes me
floating-point values. The improvement will be limited, though, by the output
images' resolution, which (typically) can not resolve better than 6 bits in
depth.
>>animation capability<< … is automatic – you simply render a segment of
frames.
>>on-screen previews<< actually, I would not know how to drastically accelerate
the stereogram generation process still resulting in 'visible' images, but I
could think of a preview for the tiling options (texture scaling).
>>fairly inexpensive<< I am targetting low shareware regions.
Considering these features and the fact that you can actually _render_ your
stereogram from 3D Editor, I found the IPAS solution very straight-forward and
handy.
Boran
Hey, it sounds like you've done all your homework! Your routine
should be great fun and a useful tool. Dispite the "fad" appearance,
there is still a lot of interest in stereograms. I make some
customized for special occasions – with people's names, etc. My
12-year-old son has made some for gifts at school and even sold a
few. A seamless output thru 3DS would make this easier. Here's a
deal for you: if it is ready sometime before Siggraph, I'll be glad
beta test it for you and still pay for it – how about that?
>>z-buffer, limited improvement…
My thoughts on z-buffer were based primarily on the potential ease of
use, not necessarily image improvement. With the atmosphere/fog
method, you have to specifically set the lighting and materials in
the before rendering a source image. By using the z-buffer, you
could create a SIS without changing anything in the scene. Please
correct me if I'm "foggy" on this. <g>
>>on-screen previews…
I actually had two things in mind here.
First, since you are in fact texture based, I would like to
see a way to view, adjust, and edit the texture bitmaps. I
spent a lot of time in an external program adjusting
bitmaps. If you provided a visual means to scale, and select a
strip from a bitmap it would make things easier. In SIS
utopia, a simple paint option [have you seen MeshPaint?!]
would allow quick adjustment of a bitmap. This is not a real
problem for me, personally, since I have a second networked
machine for my bitmap editing, but would be a useful feature
in general. (and a pain to implement!)
Second, I was thinking of a way to view the completed result
without having to use Render/View/Image. Now that I know your
routine is an IXP, this should be automatic.
Your routine sounds great! And your sharware target fits my
philosophy that many things in life are worth a considerable effort
even without a promise of getting rich. <gasp! – heresy!> Best of
wishes to you.
John,
we have a deal.
>>z-buffer, easy to use<< – you are seeing totally clear. Rendering with no fog
option can also deliver the 'answer-to-the-riddle-image' with nice, realistic
colors (like I prefer to create it), at the same time as generating the
z-buffer information.
>>on-screen previews<<
As I don't feel my destiny is programming all day, I will implement a tiling
preview with scaling interaction only – sorry for all non-networked workers.
I have to admit I did not look into Meshpaint so far, but you finally made me
donwloading it.
And, you are right, there might be nice solution for changing some
SIS parameters without rendering over again. The routine could keep
the z-buffer information, maybe save it to a file, and re-use.
Thanks for your good karma, too.
Boran