I want to bring a potential issue of the cloud detail range system to 
everyone's attention.

What it does it reducing sprite count by dropping sprites in the back half of 
the cloud if the cloud is more than a configurable distance away, so faraway 
clouds have less overall density.

This works great for Cu-type clouds which are generated from 20-30 relatively 
opaque sprites - faraway Cu clouds still look decent with half, and in 50 km 
distance would work well with a single sprite, as this is still very opaque. 

The problem are faint Cirrus-type clouds which use 2-3 sprites, or in extreme 
cases just a single one (they have a transparency of more than 90% and are 
really just a faint haze modulation). Here, dropping half of the sprites 
essentially mutilates the appearance of the cloud. I have just comitted some 
very nice undulatus pattern distributions, and dropping half (actually in 
practice about 2/3 of the sprites seem to go...) just changes the appearance of 
the sky drastically.

I am unsure of what to do about it. The detail range works as advertized, and 
is clearly useful for reasonable performance, but currently the user is unaware 
that he's even in light cloud cover not seeing Cirrus patterns which should be 
there. So it's mainly an awareness issue, setting the cloud detail range to a 
large value makes everything re-appear.

I think ideally a way to drop sprites only for certain types of clouds and 
protect the faint ones would be desirable. Do we for instance  have an unused 
attribute slot where the number of sprites used to generate the cloud can be 
passed?

* Thorsten
------------------------------------------------------------------------------
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
_______________________________________________
Flightgear-devel mailing list
Flightgear-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/flightgear-devel

Reply via email to