z-fighting can also be related to hardware variance and drivers ... it's not just a DX11 and/or 3D model issue.  I've seen nVidia release new driver profiles for main stream games (not P3D of course) that addressed z-fighting issues ... so the problem can be addressed at multiple levels.

 

You can also combat z-fighting by adjusting the 3D model ... however because of DX API differences what may look good in DX9 may not look good in DX11.  Implications being two different 3D models for FSX and P3D which is not an attractive option for 3rd party developers.

 

Cheers, Rob.

I find AA is pretty good in P3D. Making changes in the Samples setting shows differences clearly. In flight, lines on runways smooth up well enough with 4x, maybe not quite enough with 2x looking slightly like dotted lines. But 2xMSAA + 1.5xDSR, or 4xMSAA and zero DSR, usually solves the AA problems well enough. However, some objects, aircraft, or instruments can look poorly antialiased, and only smooth up with the Sparse Grid setting in NI. You'll know it when you see it, apply SGSSAA for those circumstances, otherwise it generally makes no difference. The Wobbly nature of the buildings and objects in the video, are related to geometry and filtering. If wobbly objects are causing a problem it might be worth trying adjustments to the texture filtering setting in P3D and Texture Filtering section in NI.

This is the ClipMode test I duplicated based on Slayer's video (Orbx KMRY).  In my testing not setting ClipMode for the VC does indeed match up to what the SDK says ... Normal.

 

I can confirm that you do NOT want to use ClipMode = Minimum for VC as that will produce the issues Slayer pointed out. 

 

Quick video I made at KMRY:

 

 

As you can see, not having a ClipMode entry for VC does default to "Normal".  I also included how SGSSAA and MSAA + FXAA look ... in this particular case MSAA + FXAA seems to produce the best image quality.

 

NOTE: this is a 4K video source (3840 x 2160) so if you view it at lower resolutions it will skew the AA issue and that would be essentially down sampling the video which will make jagged edges less apparent.  However for ClipMode testing it's valid.

 

Cheers, Rob.

 

 

As you can see, not having a ClipMode entry for VC does default to "Normal".

 

But Rob since many are seeing an improvement like Slayer you would have to assume that in their cases, for some reason, the absence of a clip mode entry is not resulting in a default to normal.

 

gb.

the absence of a clip mode entry is not resulting in a default to normal.

 

I believe Slayer indicated it was EZCA that had changed the values in the Cameras.cfg ... there are other products that will change the values also ... and keep in mind that these values can by changed dynamically by any 3rd party product.

 

I wasn't able to replicate the defaulting issue.  If this issue were to be presented to LM by end users, LM would need to be able to replicate and they would most likely do that from a base (no-add) environment.

 

Cheers, Rob.