------=_Part_11504_2595886.1150432416776
Content-Type: multipart/alternative;
boundary="----=_Part_11505_8304664.1150432416776"
------=_Part_11505_8304664.1150432416776
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
I just uploaded a new MC5D install. There is now a menu with a couple of
stereo options (cross-eyed/anaglyph), so time to bust out your red/cyan
comic book glasses if you've got some (you can get them online at
http://www.rainbowsymphony.com/3d-anaglyph-glasses.html btw). There is also
a new slider that will control the amount of angular difference between the
two images. You can play with this until you find a good effect.
I made it so that stereo would work correctly with the red filter over your
left eye when displaying the puzzle on a white background. If you leave the
background the default black, you'll need to reverse how you wear your
glasses so the blue filter is over the left eye. I had never known this
until playing with it, but different background colors reverse what gets
filtered out. For white backgrounds, the red filter shows the cyan color,
but for black backgrounds, the red filter shows the red color. I don't
understand this, but it is true! I bet there is some color theory
explanation or something. If I wear the glasses backward from intended (e.g.
blue over left eye with a white background), often it just looks flat but
sometimes results in trippy necker effects that make it look like the puzzle
is warping in weird ways. I think white backgrounds work better than black,
and I also get a better 3D sensation with less on the screen (2^5 cube with
some faces turned off). I'm attaching some settings that worked well for me
with a 2^5.
It's sad but anaglyph isn't usable for solving because it really doesn't
work well at all with colors (though I'm sure there's probably some better
way to do it than I am). To get a 3D effect with colors, cross-eyed is
probably better. I've never been able to do the cross-eyed stereo though,
and so I'm not even 100% sure I have it presenting/rotating the images the
right way (except for once, I never could do those magic eye things
either). Melinda, if your offer to write a "how to" for cross-eyed viewing
is still on the table, that would be awesome. Also, let me know if I am not
doing it right or if something could make it better. After trying
cross-eyed stuff, I always worry my vision is never going to go back to
normal :)
Have fun,
Roice
------=_Part_11505_8304664.1150432416776
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Have fun,
Roice
Alright!! This is awesome, Roice. I'm sorry that you can't see the
cross-eyed stereo yet. It's quite beautiful in all it's colorful glory.
I'm sort of surprised that you attempted color anaglyph views because
that loses the colors that you need to solve it, though it's interesting
what you learned about the background colors. I didn't know that
although I did know about a related effect that lets you do stereo in a
single projection, but again not preserving color. (See chromatek.com.
They may even still send out free glasses if you ask nicely.)
You asked about instructions for learning cross-eyed stereo. I've taught
a number of people. It tends to be easier to learn wall-eyed stereo but
that limits you to very small image pairs whereas when you can fuse
cross-eyed pairs you can use the entire screen. I stumbled onto one page
(http://www.physcip.uni-stuttgart.de/phy11733/stereo_e.html) that has
some very good instructions along with some excellent stereo images to
practice on. I'll be happy to give more tips if anyone has questions
after trying that page.
It definitely will not damage your eyes to try this but it can cause
headaches if you are not careful. One of the most important things to
watch for is to keep your head perfectly aligned with the vertical axis
of the screen. If your alignment is slightly off, your eyes will try to
compensate which I remember finding very painful in the beginning. Once
you get the image locked in it should be quite comfortable. You may find
yourself falling out of it a lot at first. I recommend cleaning your
screen clean first and keeping the cursor away from the images while
practicing since small differences between the two eye views can attract
the attention of both of your eyes which kills the effect.
At first I thought that your initial stereo separation choice was too
much but after using it for a while and playing with other values, I've
decided that it is actually a fairly good default. Here are some
problems I found:
* Orientations in which several rows of stickers line up on top of
another row are difficult to view especially if they have similar
colors. A little rotating fixes this.
* Scaling up the image also causes the images to spread. I think the
centers should stay fixed while scaling, and another slider added to
allow users to set the image separation independently from the virtual
eye separation. (I.E. a purely 2D viewport operation.) This will require
you to clip the parts of one view that extend into the other one. You
might get that for free if you place 2 3D views side-by-side and let the
graphics library or window system do the clipping for you but I would
love to be able to zoom deeper into the puzzle in stereo. Currently when
I try that the views just move offscreen.
* You might want to add a wall-eyed option since it would be trivial at
this point and might be a good option if you have trouble clipping
zoomed in views since wall-eyed works best with smaller images.
* Click-twisting doesn't seem to work in stereo although I can finish a
twist in stereo that I began in mono.
That's all for now. Again, great stuff!! This is crazy-cool.
-Melinda
On Thursday, June 15, "Roice Nelson"
>I just uploaded a new MC5D install. There is now a menu with a couple of
>stereo options (cross-eyed/anaglyph), so time to bust out your red/cyan
>comic book glasses if you've got some (you can get them online at
>http://www.rainbowsymphony.com/3d-anaglyph-glasses.html
>btw).
I don't think Rainbow Symphony is a good source. As
far as I can tell, they only sell in bulk. I would
appreciate a pointer to a source for onesies. (The
problem with this sort of product (the cardboard
version) is that it is essentially worthless. You
find such things more often as freebies; but,
unfortunately, that means there is no store where you
can go to buy just one.)
There is another useful technique for presenting
stereo. It is similar to the cross-eyed approach, but
you make one image be a mirror image and you put a
mirror between them for viewing. The surface of the
mirror is positioned perpendicular to viewing surface
and vertically bisecting the line between your eyes.
One eye looks directly at one image, while the other
'thinks' it is looking over on the same side, but it
is actually seeing the reflection of the other image.
This is not nearly as difficult to learn as the
cross-eyed technique, because you do not have to cross
your eyes to use it. It feels like both eyes are
locked onto the same image at the expected distance.
It should be straightforward to introduce this option
for a 3D effect as well. (A single well-placed minus
sign should do it!) Besides, I can find a suitable
mirror more easily than anaglyph glasses.
Regards,
David V.
On Thu, Jun 15, 2006 at 11:33:36PM -0500, Roice Nelson wrote:
> I made it so that stereo would work correctly with the red filter over
> your left eye when displaying the puzzle on a white background. If
> you leave the background the default black, you'll need to reverse how
> you wear your glasses so the blue filter is over the left eye. I had
> never known this until playing with it, but different background
> colors reverse what gets filtered out. For white backgrounds, the red
> filter shows the cyan color, but for black backgrounds, the red filter
> shows the red color. I don't understand this, but it is true! I bet
> there is some color theory explanation or something.
I never realized that before, but it makes sense...
The red filter only lets through red,
and the cyan filter only lets through green and blue.
So if you look through the red filter:
white -> red
red -> red (so you can see it on a black background but not a white one)
cyan -> black (so you can see it on a white background but not a black one)
black -> black
And if you look through the cyan filter:
white -> cyan
cyan -> cyan
red -> black
black -> black
Don
--
Don Hatch
hatch@plunk.org
http://www.plunk.org/~hatch/