Hello,

On 24. Mar 2012, at 12:37, W.Xu [via Software] wrote:

> For example, to render a scene with shadow there are two stages: first the 
> shadow map from the light's view, then the scene from the observer's view. 
> Each of these two steps requires entirely different settings including 
> viewport, frustum and output buffer sizes. And sort first/sort last 
> decomposition is needed by both stages to provide scalability. 

Task decomposition is not something we support yet natively in Equalizer. 
Getting it to scale in a non-trivial setup is still a research topic. ;)

That said, you might be able to hard-configure something which suits your needs:

> compound
> {
>     compound
>     {
>         channel "final_channel"
>         inputframe { name "illumination0" }
>         inputframe { name "illumination1" }
>         compound
>         {
>             channel "illum0_channel"
>             inputframe "shadow_map0"
>             outputframe "illumination0"
>             compound
>             {
>                 channel "shadow0_channel"
>                 projection { ...... }        # Shadow map need a different 
> view
>                 outputframe "shadow_map0"
>             }
>         }
>         ##############################
>         # Another similar illumination channel
>         ##############################
>     }
> }
> 
> 
> In the above configuration, although I struggled to set a new view/frustum 
> for the shadow map, I cannot control its viewport size. As the viewport size 
> of "final_channel" changes, all the outputframes changes their sizes.

The pixel viewport is inherited from the destination channel down to all leafs. 
You can work around that by the following structure:

- compound
-- compound "final", 2x input frames 
-- compound "illum", 1x output frame
-- compound "shadow" 1x output frame

In this case, you've got three destination channels, each controlling its own 
pixel viewport. If you use view/segment destination channels, you can control 
their frustum by setting the wall of the corresponding views.


HTH,

Stefan.



--
View this message in context: 
http://software.1713.n2.nabble.com/Is-it-possible-to-build-a-multi-stage-rendering-pipeline-with-the-Equalizer-library-tp7401123p7408566.html
Sent from the Equalizer - Parallel Rendering mailing list archive at Nabble.com.

_______________________________________________
eq-dev mailing list
[email protected]
http://www.equalizergraphics.com/cgi-bin/mailman/listinfo/eq-dev
http://www.equalizergraphics.com

Reply via email to