Tim Moore wrote:
>-----BEGIN PGP SIGNED MESSAGE----- >Hash: SHA1 > >Harald JOHNSEN wrote: > > >>Vivian Meazza wrote: >> >> >> >>>John Wojnaroski >>> >>> >>> >>> >>> >... > > >>>>I've yet to see a system that IMHO tops what Mark Harris did >>>>a few years >>>>ago part of his doctoral thesis. >>>> >>>>See http://www.lfstech.com/img/sfo_clouds.jpg. >>>> >>>>Someone made a decision a few years ago to replace rather >>>>than improve. >>>>Think we all lost a very promising >>>>implementation, but there might be an opportunity to recover >>>>what was lost. >>>> >>>>Stay tuned... >>>> >>>>John >>>> >>>> >>>> >>>> >>>> >>There are methods today to do real volumetric display with slices from a >>3d volume, so yes there is method to draw clouds a lot better than those. >>The old implementation of the Harris code in fg was using hard coded >>cloud shape, hard coded cloud relative position between clouds, hard >>coded group of cloud around ksfo. The next implementation could do >>clouds everywhere on the planet, parametrable cloud shapes in an xml >>file (Harris was using a non editable binary format), parametrable cloud >>formation type. >>Wasn't that some kind of improvement ? What could have been improved >>then is some new texture for the cloud particles or new shapes or >>whatever, frankly anybody could enhance what I did. >>Hm wait, you did not realize that Harris clouds are *slow* and ppl on >>this list were using depreciated hardware... >>Using his method to render the clouds is very easy to integrate in our code. >> >> >For what it's worth, I recently ported Harris' old SkyWorks code, on which his >contribution to FlightGear was based, to Linux for some playing around. I'll >make >that code available soon, but I have to agree with Harald's critique of the old >code. In particular, the key illumination step for a particular cloud is *very* >expensive -- in the SkyWorks code it repeatedly reads back color values from >the >frame buffer for each particle in a cloud -- which is why particular clouds >were >hard-coded. Harris has since described algorithms that are more dynamic and >which >require heavy-duty GPU firepower, something that always seems to be sensitive >topic >here. > >A free cloud library based on Harris' work, integratable with other OSG >applications, >would attract a large following and would be a very cool thing. In the mean >time, >in order to learn some of the issues, it would be very good to port Harald's >plib-based >work into OSG FlightGear. > > > Are you suggesting a library based on Harris' work or Harald's? or both? John ------------------------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/ _______________________________________________ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
