Hemispherical Dome Projector

Document the build progress for your screen or computer setup.

Postby mweaver » Tue Oct 11, 2016 7:33 am

I have started the creation of our hemispherical dome projection system. It is based on the iDome concept by Paul Bourke http://paulbourke.net/dome/iDome/

I have sourced a projector, dome and mirror, and I am beginning the testing this week. Stay tuned for a small build log. I will be using Nthusim to warp the image to the dome. Comments welcome!

dome.jpg
small dome
dome.jpg (48.38 KiB) Viewed 1765 times
mweaver
 
Posts: 6
Joined: Sat Jun 04, 2016 2:59 am

Postby BHawthorne » Tue Oct 11, 2016 9:18 am

Looks cool! I'm definitely a fan of Paul Bourke's iDome research. I'm also really interested in seeing how the security mirror works.

If you have any feature request to make warp calibration easier let me know.

-Brad
Image
User avatar
BHawthorne
Product Manager
Product Manager
 
Posts: 1291
Joined: Thu Sep 20, 2012 11:02 pm

Postby mweaver » Wed Oct 12, 2016 12:22 am

I have a 1st surface mirror in a crate for when I am comfortable with the setup. I am using this 30 dollar mirror for prototyping. I will photograph the image generated by the security mirror soon. As expected, it generates an okay image, albeit with slight ghosting on high contrast images. The company that provided the first surface mirror suggested that I never even so much as breathe on it, as it can not be cleaned.
mweaver
 
Posts: 6
Joined: Sat Jun 04, 2016 2:59 am

Postby BHawthorne » Thu Oct 13, 2016 8:59 am

I've pondered picking up one of the regular security mirrors for experimentation. Is the ghosting tolerable for experimental purposes?
Image
User avatar
BHawthorne
Product Manager
Product Manager
 
Posts: 1291
Joined: Thu Sep 20, 2012 11:02 pm

Postby mweaver » Tue Oct 18, 2016 3:16 am

Ghosting is very tolerable. It shows itself most in high contrast situations, like black text on white background. NTHUSIM is not function on the 2d desktop, even when configured to do so. It has worked in a full screen 3d application. Here is the error log. Any suggestions?



Mon Oct 17 12:05:31 2016 : [3668] [HLAPI] Failed to load blend map C:\Users\USERNAME\AppData\Local\Nthusim\default\blendmap_1.png

Mon Oct 17 12:05:31 2016 : [3668] [BE] Could not enable 2D correction : desktop composition is off

Mon Oct 17 12:05:31 2016 : [3668] [BE] Could not enable 2D correction : desktop composition is off

Mon Oct 17 12:05:34 2016 : [3668] [BE] Could not enable 2D correction : desktop composition is off

Mon Oct 17 12:05:36 2016 : [2828] DirectX 11 32bit requested by C:\Program Files (x86)\Mozilla Firefox\firefox.exe

Mon Oct 17 12:05:55 2016 : [3860] [BE] Could not enable 2D correction : desktop composition is off
mweaver
 
Posts: 6
Joined: Sat Jun 04, 2016 2:59 am

Postby BHawthorne » Tue Oct 18, 2016 6:59 am

When it says desktop composition is off that means you're not using an Aero themed desktop for hardware acceleration. Which Windows version are you using? Don't use a basic theme for the desktop, use a hardware accelerated one.

In "Control Panel > All Control Panel Items > System" select the "Advanced" tab.
Under "Performance" select the "Settings..." button.
Select "Adjust for best appearance" to ensure all hardware accelerated portions of desktop are enabled.
Then make sure you're using a hardware accelerated Windows theme.

That should fix the desktop composition issue.

A side note, the blendmap error message isn't related and won't affect Nthusim use.
Image
User avatar
BHawthorne
Product Manager
Product Manager
 
Posts: 1291
Joined: Thu Sep 20, 2012 11:02 pm

Postby mweaver » Wed Oct 19, 2016 5:18 am

Great thanks, that worked well. You had asked earlier if there was anything you could do to help with image warping on a dome. I have used the default configuration test pattern, and I believe it works fairly well, however Paul Burke uses a 'polar test grid' for his calibration. Is there a way to replace the default test grid with one similar to this?
http://paulbourke.net/dome/iDome/warp1.jpg

Otherwise, here is an image after warping with the current test grid, it works okay for the center, but the edges are heavily deformed. This is with a small (18 inch) dome, I also have a 26 inch dome that I am testing, one or the other may be better suited to the geometry of the acrylic dome projection surface.

small-dome.jpg
small-dome.jpg (66.69 KiB) Viewed 1685 times


One issue I have with menus and other high contrast areas is the zig-zag lines that are formed by the warping, I have attempted to force anti-aliasing through the graphics driver but it had little to no effect.
mweaver
 
Posts: 6
Joined: Sat Jun 04, 2016 2:59 am

Postby BHawthorne » Wed Oct 19, 2016 5:51 am

I'll ask the brisbane office to add in the polar grid option for 4.2.
Image
User avatar
BHawthorne
Product Manager
Product Manager
 
Posts: 1291
Joined: Thu Sep 20, 2012 11:02 pm

Postby BHawthorne » Wed Oct 19, 2016 9:17 am

For clarification you're wanting what Paul Bourke refers to as a "truncated bottom" with a 16:9 aspect ratio? Making 100% sure the warp grid I'm thinking of is correct.

I drew this in Adobe Illustrator based on a few examples from Paul Bourke reference pictures. I'll see if I can have open a picture for calibration background as an option added to Nthusim again. That way you can just load this picture up and calibrate with it. We used to have the option around 2009 when we were known as Sol7 Home Edition.

http://imageshack.com/a/img922/3058/et5IyQ.gif
Image
User avatar
BHawthorne
Product Manager
Product Manager
 
Posts: 1291
Joined: Thu Sep 20, 2012 11:02 pm

Postby mweaver » Thu Oct 20, 2016 12:25 am

I believe the image you linked, and the image I linked in a previous post are the results of warping a polar grid, and that an unwarped polar grid should instead be the reference image for domes.

Here is the radial grid file from Paul Burke's Meshmapper software http://paulbourke.net/dome/meshmapper/

radialgrid.jpg
radialgrid.jpg (225.75 KiB) Viewed 1676 times
mweaver
 
Posts: 6
Joined: Sat Jun 04, 2016 2:59 am

Next

Return to Project Build Logs

cron