Heiko: > Thanks, that's now an answer I was looking for. > > Though I'm sure we will get bug reports as well by users complaining > where the shader from 2.6.0 went, or why shaders won't work. > Some of this confusion was already visible in the forum.
I don't want to bitch, but the thread started out with a description of the issue like this: "The problem: A shader for object X isn't an independent entity which you can switch on and off at your leisure independent of what else is going on - it needs to share the fogging and lighting scheme of the rest of the scene, otherwise it creates visual artefacts. At the moment, we have three such schemes: a) default in which the light is always the sunlight and the fog is always exp(-(const * length/vis)**2) b) Rembrandt in which the light can come from multiple sources and sunlight may be blocked by a shadow, fog remains as is c) Lightfields in which the sunlight is the only light source but at dusk or dawn becomes a function of position in the scene and in which the fog has a more complicated 2 component structure Each of the schemes has a justification to be there in terms of features vs. performance. Making a shader of one scheme work in a different scheme is a non-trivial exercise, in the case of water reflection shader into lightfields it cost me about half a week of really good coding time (kids with grandparents). So we simply don't have all shaders in all schemes, and I don't even know if all would run with enough fps in all schemes. How to present it to the user? (...) Please give some feedback - how can we communicate better what is happening and why there are limitations?" My personal opinion is that we should not allow to choose inconsistent shader options from the GUI in a release (and should not have done it in 2.6 either) but should make those only available via a commandline switch for the people who know what they're doing. http://www.hoerbird.net/fgfs-screen-764.jpg illustrates that the random vegetation shader isn't converted - trees are fogged too weakly as compared with the rest of the scene. http://www.hoerbird.net/fgfs-screen-763.jpg illustrates the problems of shader mismatch very nicely - there's a sharp line dividing sky and terrain where fog color and amount do not match. Try the same scene in sunset for a real bad mismatch... Curt: > I think the problem now is that the nice scatter effect sky dome no > longer works correctly with any render mode in the git version. It never did work correctly with any shader before. Now it works with some. Try using a 2.6, set visibility to 3000 m, switch any shaders you want on and see what happens - the whole world is fogged, except the sky is a bright, shiny blue to the horizon. > The scatter effect sky dome will give you very pretty skies -- especially in > combination with advanced rendering, just don't look at the terrain or > your own aircraft model because those are now rendered incorrectly with many > missing effects. I'm lost. Can you explain and give screenshots? I am seeing correctly rendered terrain and reasonably rendered aircraft when using the atmospheric scattering. I'm not seeing detailed aircraft effects because the shader hasn't been converted. In principle, I think here one could think of just allowing to run the default scheme shaders because the atmospheric scattering scheme wouldn't do much extra to your own aircraft - the problem is MP and wrong fogging and lighting for other people's planes. I don't know. I'm not seeing transition effects on the terrain, urban shader or ground relief because these aren't done either. I am seeing wave patterns and snowline because these are done. None of this is supposed to be recent behaviour. The old skydome of 2.6 never ever worked correctly with any other shader. Atmospheric light scattering (lightfields, terrain haze, the name changes to more general as I added features, but it's all the same family and I think I'll stick with atmospheric light scattering) never supported transition shaders for the terrain or urban effect before. They're not suddenly gone, they never were working before in the scheme. If there's anything recently broken, please explain with details - I'm not seeing it. * Thorsten ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
