[edwardk]

The question is this. If I set the graphics settings for runactivity.exe under the control panel, should the results be the same if running OR in either Windowed mode or non-windowed mode? The first screenshot shows OR with Windowed mode not set. You will notice the moire pattern on the tracks. The second screenshot is with OR running in Windowed mode. The moire pattern is not present and this is with the same settings under the control panel. Could this be a possible OR issue that was missed or something else. I just recently upgraded from a GTS250 to the GTX760 SC. I can't recall if this was like this with the GTS250.

Edward K.

Attached thumbnail(s)

• 
• 

[markus_GE]

I too am experiencing that moire pattern in full screen, with an ATI / AMD Radeon Mobility HD5650 graphics card, that has been installed for five years.

IDK what about windowed mode (will try and report back), but I know that some revisions ago (a few weeks ago, to be precise B) ) the pattern was not there in full screen. I think the issue was reported back then, but I can´t recall for sure.

Anyway, would be cool if it could be removed. :)

Cheers, Markus

[James Ross]

edwardk, on 15 April 2014 - 12:37 PM, said:

The question is this. If I set the graphics settings for runactivity.exe under the control panel, should the results be the same if running OR in either Windowed mode or non-windowed mode?

As far as I can see, you're getting the same visual rendering in both screenshots. The shots are taken at different angles, so they can't be compared properly, but I can see moire on both. But even if you had two screenshots at the same angle with different renderings, there's nothing in OR that selects or uses different rendering for fullscreen and not, so your problem would entirely lay with NVIDIA. (I'll save you what I think about NVIDIA's drivers.)

markus_GE, on 15 April 2014 - 12:55 PM, said:

Anyway, would be cool if it could be removed. :)

For NVIDIA at least, you need to turn up the texture quality from the default of "quality" to "high quality" because NVIDIA's drivers are taking liberties with the rendering. We flipped one rendering option - enable alpha blend - and apparently that also degrades the texture rendering quality. There's no technical reason this should be the case and as evidence of this you can fix the problem by changing the driver's own settings, so I'm not willing to compromise OR for this at this time.

It's worth remembering that the majority of settings in NVIDIA's control panel (and AMD's) are for things that applications have absolutely no direct ability to control in the first place. There is nothing we can do to control the "texture quality" in NVIDIA drivers, for example; it's up to the driver to not screw up the result, something that hasn't happened here.

[markus_GE]

Thanks for your explanation, James :)

It really seems like that change you mentioned caused the change in the visual quality. I did not touch my graphics card´s settings after that cahnge, yet image quality suffered.

Anyway, out for some fiddling again :)

Cheers, Markus

[R H Steele]

I discovered the same thing. Angle of view has a great deal to do with minimizing or maximizing the moire patterns as the simulation is running. Careful observation shows that a pattern begins, the distortion peaking at a certain distance and angle, then diminishes and finally stops. I finally read an explanation that in MSTS it's caused by overlaying or combining two textures that causes these patterns. A byproduct of how buildings, tracks and other world objects are constructed I suppose. Some adjustment is possible with very high end graphics monitors - which usually results in softening the overall image, but little can be done with card settings. I don't think anyone has yet produced a graphic card with drivers that please most of the people some of the time. (is that right?)

[markus_GE]

Something CAN be done with graphics settings: Increase anti-aliasing, at the cost of (quite a few) FPS.

It´s about finding the right balance, one again. I was able to completely remove the moire pattern, maxing out AA and dropping my frames to the low single digits. Well, that said, try it yourself, but be careful about performance!

Cheers, Markus

[R H Steele]

markus_GE, on 15 April 2014 - 01:51 PM, said:

Something CAN be done with graphics settings: Increase anti-aliasing, at the cost of (quite a few) FPS.

It´s about finding the right balance, one again. I was able to completely remove the moire pattern, maxing out AA and dropping my frames to the low single digits. Well, that said, try it yourself, but be careful about performance!

Cheers, Markus

I will. Although I don't know what else I can do. Here's the AA setting being used for X2173 (and most X version prior to that) Maybe changing the "override app setting"?

antialiasing fxaa....................................................ON
antialiasing gamma correction.............................ON
antialiasing mode..................................................OVERRIDE APP SETTING
antialiasing setting................................................32X CSAA
antialiasing transparency......................................8X (supersample)

http://www.elvastowe...225#entry146225 Link to thread Hash House/PC Systems Discussion about moire patterns. Some very good information here. I'm still reading and digesting the material. I'm really interested in trying out Genma's shape file edits on a couple of corrugated metal buildings that go really moire crazy.

[edwardk]

Evidently the screenshots can't show everything, but all I know is that with one standard setting for runactivity.exe, the windowed mode is much cleaner compared to the non-windowed mode and it has nothing to do with the angle of the camera view. For example, the moire pattern is much worse in non-windowed mode.

Edit: In the post above this, my settings is similar to his, but the overall appearance is different between the two modes.

Edward K.

[D&RGW]

Hello,

I'm not sure if this'll help or not, but one thing to consider is how anisotropic filtering is being handled by the game engine and/or graphics card (at least, that's what I thought of when you mentioned viewing angle). Antialiasing, as far as I know, has to do with smoothing "jaggy" edges, while anisotropic filtering affects the quality of textures when being viewed at shallow angles. MSTS definitely doesn't support anisotropic filtering but I would imagine that ORTS does. Here's the Wikipedia page for it and a sample image:

http://en.wikipedia....ropic_filtering


http://upload.wikimedia.org/wikipedia/commons/8/80/Anisotropic_filtering_en.png


I don't know if this will be of any help but I thought I'd throw it out there for consideration.


Thanks,

Eliot

[James Ross]

D&RGW, on 16 April 2014 - 12:15 AM, said:

I'm not sure if this'll help or not, but one thing to consider is how anisotropic filtering is being handled by the game engine and/or graphics card (at least, that's what I thought of when you mentioned viewing angle). Antialiasing, as far as I know, has to do with smoothing "jaggy" edges, while anisotropic filtering affects the quality of textures when being viewed at shallow angles. MSTS definitely doesn't support anisotropic filtering but I would imagine that ORTS does. Here's the Wikipedia page for it and a sample image:

You are correct that this problem is about texture filtering and not antialiasing. Open Rails requests anisotropic texture filtering and has done for years. The problem here is that, by changing something else, NVIDIA have stopped actually using anisotropic texture filtering and are doing something else that looks bad. This is something I mentioned in my earlier post but might have been lost in the other details:

James Ross, on 15 April 2014 - 01:16 PM, said:

For NVIDIA at least, you need to turn up the texture quality from the default of "quality" to "high quality" ...

That one simple change of texture quality in the NVIDIA control panel completely eliminates the moire effects for me by causing the anisotropic texture filtering in OR to be respected.