2.14beta Clouds

[Amerio.Stephane]

Hello,

The new clouds are just superbs! I manually added all clouds materials to the ig_config.xml so I can use them with the IG. I still have a few remarks when using them within an IG though:

1. MAJOR: The sun looks broken. Time 13:00 on timezone 0 is night at 43N5E.
2. MAJOR: Layer 0 (fog) is still not supported in beta. Will it be there in 2.14 final?
3. MINOR: As fog is not (yet) supported, we use Layer1 with Type=cirrus. But the layer actually starts at elevation+some unspecified offset. The real thickness is also half of the specified one. This makes it difficult to simulate a precise ground fog layer.
4. MINOR. The transition bands (below/above cloud layer) is not used. Quite useful with a overcast sky.
5. CRITICAL: When adding a layer with coverage>0, the visibility between the clouds and above the clouds drops. This is problematic, especially because when above the clouds, the visibility should be the globally set one (1E6 typically). You can best see this with the Cumulonimbus and coverage=100, there is a single large cloud (ok) with a very darkened sky all around (arguably, coverage=100 is meaningless for this type, but it shows the problem very well. Try also with cirrocumulus and go above the clouds). This is critical because most Instrument flight condition specifically requires that visibility range is "perfect" above and below the cloud layer. At the very least, the sky aspect/visibility should return to normal when above the cloud layer.
6. Question: how do you set the world position of single clouds like thecumulonimbus? Are Region/entities supported?

Thanks

 

[silent]

Hi Stephane,

Thanks for the feedback! I will specify issues that we are plan to fix in the upcoming 2.14 release:

• 1 (broken sun), 5 (visibility between layers)

These will require heavy Meteo refactoring (that wasn't possible until the new clouds system):

• 2 (fog), 4 (overcast) and probably 6 (right now you can only move cumulonimbus if you will manually create an entity for that type of cloud and control it manually)

Regarding the #3 - that's currently expected behavior, the cloud layer will always have these margins (they are used in height falloff and noise falloff). Maybe we can come up with some kind of a solution for this from the IG point of view, but that's hard to do a exact estimation on this right now.

[Amerio.Stephane]

Just to be clear on point 5:

This is the current 2.14beta state with clouds cirrus (type 5), with coverage 80, above the clouds:

As you can see it's very dark, when actually it should be as bright as with coverage 0 (same view point, same time of day):

Even below the clouds, the luminosity drop feels too important, almost night-like.

 Thanks!

 

 

[qanva]

Hi Stephane,

Could you please give us more information about sun issue (#1). Right now we can't reproduce it and we need to know:
- your PC timezone
- accurate date wich you set in IG
- log and ig_config.xml
- steps to reproduce the bug ( video if it possible )

Here's our current behavior, no obvious bugs: 
 

Thanks!

[Amerio.Stephane]

Hi, 

(back on the initial topic, sorry demostenes maybe this post should be split by the moderators)

In 2.14 final, the *critical* point 5 is still present (sky is way too dark when going above clouds). Is this the intended behavior? If so, how can I get rid of it?

Thanks!

[silent]

That should be fixed.

Please, make sure that you have the latest ig_config.xml or try to reproduce in a new empty scene. You should have the similar lines inside config:

<cloud_type type="1" name="altocumulus" material="ig/clouds/materials/altocumulus.mat" light_density="0.2"/>

By tuning light_density you can control the level of brightness (1 - full dim, 0 - no dimming) for each cloud type.

Btw, if you are using debug camera (Ig_debug 1) - darkness is the expected behavior.

Thanks!

[Amerio.Stephane]

22 minutes ago, silent said:

if you are using debug camera (Ig_debug 1) - darkness is the expected behavior.

That must have been it! I'll check again then. FMI, will it be possible to have the correct behavior even with the free camera? I use it a lot for fine adjusting with the "pilot" by my side.

Thanks!

[Amerio.Stephane]

On 3/24/2021 at 6:20 AM, qanva said:

Could you please give us more information about sun issue (#1). Right now we can't reproduce it

Can't reproduce it in 2.14 final, so let's call it done :) Thanks (FYI: France here, timezone GMT+1/+2)

[silent]

Debug camera is created for a different task (to be able to actually see the how the scene is actually looks like).

I guess the easiest way to fix this would be to just create your own free camera from code and use it for additional fine tuning :)

Thanks!