In the online Tips on radiosity (2.3.4.3), we have some statements that 
are greatly misleading for both version 3.6 and 3.7, and may be 
leftovers from the original implementation.
See section 3.4.4.3.5 in the included documentation for version 3.7 for 
Windows.

There is a mention to " drop the low_error_factor to something small 
like 0.6".
It's strange when the default for low_error_factor is 0.5.
Something small should be less than 0.5.
It may explain why, in rad_def.inc we have low_error_factor set to 
something like 0.7 to 0.9 for most presets...

Another inconsticency:
"To decrease these artefacts, use a pretrace_end of 0.04 (or even 0.02 
if you are really patient and picky)."

pretrace_end 0.04 is the default. This tip should mention something like 
"To decrease these artefacts, use a pretrace_end of 0.02 (or even 0.01 
or 0.005 if you are really patient and picky)."

Those should have been ammended around version 3.5 or 3.6 to match the 
implementation.

There is no longer a need to recommend using always_sample off as it's 
now the default.

It should be noted that using a smaller pretrace_end and 
low_error_factor values can greatly help reduce the artefacts when using 
a multicore system.





Alain

Post a reply to this message