How To Setup Split Content for Softedge Projection

[Jan Huewel]

When creating softedge projections you might want to use individual files per projector to maximize the image quality by using all available pixels.

This is especially usefull when the overall pixel count is much higher than HD (1920x1080).

Since you might not want to waste performance as well as image quality by stretching a 1920x1080 clip to 2x,3x or N-times its original size, it can be very effective to create split media files out of a master composition of the entire screen.

The most important parameters to consider for the setup of the media are size and overlap.

What is the screen physical size in width and height or what is the total pixel count including the overlap area of the projectors?

How much overlap can you get per projector?

For example you have two projectors that output 1600x1200

and your horizontal overlap is going to be 20%

20% of 1600px is equals 320px

This means that your entire composition is 1600px -320px+1600px = 2880px Width x 1200px Height



When creating compositions in AfterFX or similar compositing tools, its a best practice to first create a master composition that matches the screens aspect ratio or projection screen surface.

In the case describe here I assume that we use the entire output of both projectors.

The master composition pixel size is 2880x1200px
and the left and right split composition are 1600x1200px

I hope this gives you a good starting point on how to create your media for softedge projections.

Jan Huewel
Geschäftsführer / CEO

coolux GmbH
Venloer Str. 703
50827 Köln
Germany

Tel: +49 221 130 654 0
Fax: +49 221 130 654 22
huewel@coolux.de
http://www.coolux.de

Registergericht: Amtsgericht Köln HRB 51649
Geschäftsführer: Jan Hüwel

[selor]

Hey Jan,

This is a good begining for beginners . But the interesting thing for me i am needing is to understand the Setup IN PB in order to project this two splitted medias for example .
I am just learning PB and my first problem in Head is about the virtual universe and Axis. Axis are not represented in the software so sometimes you're lost in the position of cameras or layers.
For you which is the best method to setup Layers and cameras of one server for your example ?

Because i am lost with values . Tell me if am right with thaht :

The universe is define from 0 to 65535 Values (That are not Pixels just because of 16Bits but where is the conversion or how to identify pixels and values shown in PB ? )
This for the 3 Axis X,Y and Z ok ?
So Imagine my first camera by default is in the middle of the universe so 32768 X, 32768 Y, 32768 Z .... and my layer 1 is the same ... so i can see my image centered in the camera ...
But how with precision know or calculate my position of the second layer or second camera to match your case ??

In my actual test i am with 2 * 1920*1080 HD projectors, with a softedge of 10% (192 Pixels) so i have created an image of 3648 by 1080 Pixels and split it in two 1920*1080 HD files ok ?
I am trying to setup with precision this case and don't find the way to make it with precision . I am trying to put my Cam 1 showing Layer 1 with my left splitFile and the Cam 2 showing Layer 2 with my Right splitFile. Its really hard to have it good with precision. O, you can put a Test Pattern and try to do it with eyes but its really Approximativ .. not really Perfect. ..
Did i take a bad method ? or can you explain a little the meaning of the universe .. Axis .. or ...detail more the Point of view of cameras ... and values asociated to setup ...


My english not perfect i hope it understandable.

Thanks

[Jan Huewel]

Dear Selor

as Pandoras Box renders in 3D space, avery pixel of a video or image must be represented and rendered onto textures.

The textures can then be manipulated in any way in terms of colour or form.

So as these textures are positioned always in a 3d environement getting pixels accurate output depends on the setup of the cameras and layers.

The best way to do this is in fact by dialing in the x or y offset on the cameras. The xyz axis parameters will , depending on the mode, position the camera in space which will lead to a perspective view from a side and will not allow you to get the result you are looking for.

If more resolution in terms of precision is needed layers can also be placed further away from the camera and scaled up this way you will increase the accuracy of the positioning of the layers to be subpixel accurate.

So for a setup I would suggest to use the geomertry testpattern and adjust the offset parameter until the center cross matches perfectly in the center of the softedge blend zone.

Best regards

Jan

[Daniel Kaminski]

Hello,

I have created a Excel Spreadsheet and a Adobe After Effects CS4 Project as a Demonstration of how it works.

The Demo Files are for a 4 Projector Softedge with 15% Overlap.
The projectors are running at a resolution of 1280x1024 Pixel.

I hope this helps to get started.

Enjoy and feel free to modify.

Daniel