I agree with you default weather options now have much smaller clouds, but have you tried moving cloud density slider up to max? BUT, if you use ASN (haven't tested other weather engines yet) you will get much better clouds and performance.

 

 

 

But you might also want to try to changing the multiplier from 0.5 to 0.75 and see how that works (or not for you).

 

Thanks Rob... I tried some different values, but unless they are even the clouds will look stretched in one direction or the other. I'm sure this fix will aid those using weather engines that create complex weather data, however I use my own custom weather themes, so can tailor things to my own taste.

 

There is one issue this will resolve, apart from performance, in that the deck altitude will be correct, whereas the oversized clouds will give a lower deck. Weather engines setting a lowish base altitude will often result in clouds intersecting the terrain, which of course is not correct.

 

Despite this I really prefer the look of the bigger billboards... and performance is much the same.

 

Max density @ 5/8

Pre Fix.

 

Same settings post fix.

That first picture isn't 5/8 though it's 7/8+ where the cloud actually is. Even the second one looks a bit much for 5/8. 

can someone post a few screenshots of what it's suppose to look like, i don't want to mess my sim up and i am not good with text editors

That first picture isn't 5/8 though it's 7/8+ where the cloud actually is. Even the second one looks a bit much for 5/8. 

 

 

No, it is 5/8... I have just checked the save I used to take both images. The extra size of the pre fix clouds give far greater coverage, and does indeed look more like 7/8 which is why performance suffered, as there is more cloud being drawn at the same density setting.

 

 

can someone post a few screenshots of what it's suppose to look like, i don't want to mess my sim up and i am not good with text editors

 

The bottom image IS what it should look like, at 5/8's coverage and max density (much the same as 2.2 except for the new noise patterns)... back-up the cloud.fx and Shaders file so you can revert.

The point I am making is that even after the fix the cloud coverage is still too much, so a 0.4 scale may be even more appropriate. The textures used may influence this to some degree though.

 

 

The point I am making is

 

Ah okay... much smaller and you risk getting silly looking cotton wool balls, even at high densities. Plus you get no merging of the base, with lots of separate dark areas. The larger clouds do merge to give some very realistic looking weather IMO... but the issues they cause make it difficult for them to be used, especially with a weather engine.

 

 

Despite this I really prefer the look of the bigger billboards... and performance is much the same.

 

I like both of those images ... think I'll use both settings pending what I'm doing ... for my custom weather I'll go without the change (remove multiplier or just go 1.0) and for Weather engines I'll go with (0.5).  I have two files that I'll just swap in/out and make sure I delete the shaders afterwards.

 

Since opinion on this matter obviously varies, perhaps the best solution would be to make the values adjustable via the P3D UI - maybe next to the cloud density setting.

 

Cheers, Rob.

I guess I don't understand shader technology (OK there's no guessing about it), but when you place an object in the sim you can scale it to any size you want and I've placed some at horrendous scales but I've never seen any performance impact at all as a result. I'm talking about several thousand vegetation placements and experimenting with scale values between .3 and 2.0 on the entire group, so if it were to have an impact I think I would have seen it at some point. It seems odd to me that scaling these clouds actually has an effect but obviously it does. It keeps coming back around to pixels displayed on screen so I wonder what would happen if you restored clouds.fx to original and blacked out the alpha channel on cumulus01.bmp making the clouds completely transparent? Guess there's only one way to find out... :smile:

I hope LM fixes this but not at the expense of the great looking clouds we are having now.

My system can handle them smoothly everywhere but Seattle , where drops to 19 fps.

 

No sttuters though.

 

GTX 780 Ti

God bless you LM.   My aircraft is a happy plane again.

I think some of you might be missing the point.

 

According to Beau, "Some of the cloud math got tweaked during 2.3 development and the size of clouds got doubled." and "Scale the width and height down by half with this change".

 

The edit is to restore the clouds to what they should more correctly be rendered at. I'm using ASN and with live weather, the cloud coverage at a nearby airport (after this edit/fix and enabling the #define VOLUMIZE) is very close to what it actually is at my home. Also the clouds no longer drop lower than what the ASN weather gives (when there's a 5000 foot ceiling, the clouds no longer obscure the top of a 4500 foot mountain).

 

EDITED: I forgot to mention that there is a setting in ASN, under Cloud options, called 'Enhanced Overcast conditions,' that is disabled by default. If you check/enable this, overcast cloud layers will "be enhanced to provide thicker coverage at the cost of reduced performance." I tried enabling this and found that the reduced performance wasn't nearly as much as what I was getting with the pre-fix (probably because it is less extreme, and more like real coverage). So, if you have ASN and want somewhat thicker cloud coverage, this is another way to have it.

Arwen is correct. This is all about a math error. That doesn't prevent anyone from messing with the shader code, but an official patch will probably correct the original mistake.

No point being missed here :smile: I understand this cloud size was in fact an error, but for me, and seemingly loads of others, it was a rather nice one. I also understand the issues, as I have alrady stated, regarding the altitude of the cloud base, and how this error screwed it up, by placing the deck far too low.... performance less so, for me.

 

As Rob says it would be great to have an option to control the size, but the base altitude would also have to be recalculated to compensate for the bigger size...  I assume its based on the middle of the billboard?

 

To me the scale of the clouds is perfect, and personally I would like to maintain the bigger versions.

 

 

It seems odd to me that scaling these clouds actually has an effect but obviously it does

 

It seems odd to me too. Maybe is more to do with the way AA is applied to the clouds, rather than simply the larger scale?

Works fine with still nice clouds and about 10% better FPS. Not bad.

Spirit

It seems odd to me too. Maybe is more to do with the way AA is applied to the clouds, rather than simply the larger scale?

 

I think that may be correct, and I completely forgot that I'd turned on 4x SGSS in Nvidia Inspector a few days back because I was fed up with the jaggies in the cockpit. As a result my frames were really getting hammered, quite a lot more so than without on a test flight I've been repeating the past few days so I think that proves your theory.

 

 

I wonder what would happen if you restored clouds.fx to original and blacked out the alpha channel on cumulus01.bmp making the clouds completely transparent?

 

The follow up on that is the frame rates still descend into oblivion at the same place in the above mentioned test flight that they did with the normal cloud texture but without a (visible) cloud in the sky. Very impressive lightning show though when not obscured by clouds! :smile:

 

On the subject of lightning, I was blaming ASN for overdoing the lightning just a wee bit, for example the other night I did a flight where all the local metars for miles around were reporting FEW035 or similar yet I was in a lightning storm almost the entire flight. Now after this cloud scaling fix I don't think I'm seeing the excessive lightning show anymore. I always thought the lightning was depicted directly by the weather engine but apparently it's tied to cloud coverage instead. Anyone else noticing 'more reasonable' lightning after the cloud scaling fix?