Hi Juan, Just to expand on Cedric's answer:
On 16. Feb 2011, at 18:44, Juan Hernando wrote: > Is this a problem in the config file or is it a bug in the parser? It is a problem in the config file. When you are using view/segment description, the frustum is derived using the following priorities: 1) Use the view frustum if specified 2) Use the segment frustum if specified 3) Use the canvas frustum At least one of them has to be defined, otherwise the assertion is triggered. You should also start declaring the version of your config files, otherwise our loader converters may do stupid things. The difference between 1.0 and 1.1 is related to the new runtime stereo switch[1]. 1) A view frustum is typically only set by the application programmatically, representing a non-physically-correct projection, as opposed to the physical frustum description in the segment. Sub-channels of the view receive a planar sub-frustum of the view. 2) This is typically used for non-planar canvases. See next point. 3) This is used for planar canvases. The segment's frustum is automatically calculated using the canvas frustum x segment viewport, unless a segment frustum (2) overrides this calculation. [1] From <http://www.equalizergraphics.com/documents/design/stereoSwitch.html>: The semantics of the compound eye attribute changes from describing the eye passes to be rendered to the eye passes this compound can render. The actual selection of the eye passes will be done by the view of the destination channel(s). That is, a compound using eye [ LEFT RIGHT CYCPLOP ] will render all three eye passes in Equalizer 0.9.1, but only the cyclop (view is in mono mode) or left and right (view is in stereo) passes in the new implementation. Consequently, the default setting shall change from eye [ CYCLOP ] to eye [ LEFT RIGHT CYCPLOP ]. HTH, Stefan. _______________________________________________ eq-dev mailing list [email protected] http://www.equalizergraphics.com/cgi-bin/mailman/listinfo/eq-dev http://www.equalizergraphics.com
