What is the best method?

Renderer.type(RT_SINGLE_PASS);
D.multiSample( 4 );

or

Renderer.type(RT_MULTI_PASS);
D.superSample( 2 );
D.edgeSoften(EDGE_SOFTEN_BOTH, 0.7f);

Suggestions?

Multi_Pass has always better quality, but need better PC.
What exactly do you mean 'best method'?

menajev Wrote:Multi_Pass has always better quality, but need better PC.
What exactly do you mean 'best method'?

Alright, Well I want to use hardware anti Aliasing only, so I go with multiSample( 4 );?

Quote:Multi_Pass has always better quality

I must be missing something here. Could you please explain.

My understanding was that "RT_MULTI_PASS" was actually "Deferred Rendering".
The advantages for Deferred rendering is that for significantly larger number of lights , this way of doing things produces faster FPS.
For smaller number of lights the "Forward Rendering" modes like "RT_SIMPLE" and "RT_SINGLE_PASS" are faster.
The actual quality of the image should be identical.

If one uses AA in forward rendering then it is a matter of how efficient the video card happens to be with regard to handling AA in the hardware.
With deferred rendering the video card hardware can not handle AA mode properly, so if one wants to use a simulated AA mode, then one actually renders the image at 2X width and 2x height and remaps it down to the original smaller resolution.......this is a very expensive procedure for its effect on FPS.

The big disadvantage about using "RT_SINGLE_PASS" is that this engine has limited special effects built in to the library for people to use. EG: the infinity water is not available under "RT_SINGLE_PASS".

Fu** I again mistook RT_Simple with RT_Single_Pass (which author ask about and i never used it)... ;/