18 hours ago, scotchegg said:

disappointment that it’s clearly just using cumuloforms to try and approximate something else

This isn't the case, though. Seb has tried to explain this at least a couple of times, to my recollection. All clouds are made up of voxels, not a type of cloud, which are the volumetric, 3D equivalent of a pixel. The data fed into the sim is supposed to tell those voxels what type of clouds to form into, in the live weather engine. However, Seb did discuss limitations. I think he was talking about, at the last mention of it, the necessary size of the voxel units, and needing to make them smaller in future for more, and finer, rendering complexity. He also referred to increasing the complexity of the data in future. I think he said it's possible to do these things, but I think he was talking about waiting to see what performance headroom is spare after implementing other things that people want or expect the sim to have.

I'm sure they hinted at the data being, perhaps, simply wrong sometimes, as they mentioned seeking new data partners.

But there are certainly going to be limitations in the volumetric rendering path. Home hardware still isn't up to the standard that I think users have in their minds. Perhaps if the sim wasn't doing many of the other things it's doing simultaneously, we'd see even more realistic cloud depiction already.

I'm personally still surprised at just how my 1070 card can render the current clouds on ultra setting along with everything else on ultra.

Edited April 18, 20224 yr by March Hare

1 hour ago, March Hare said:

I'm sure they hinted at the data being, perhaps, simply wrong sometimes, as they mentioned seeking new data partners.

Blame meteoblue, goodbye meteoblue, hello metar and other free alternatives. Sorry about the clouds looking bad now but you must understand home hardware can't handle it.

 

All planned from the start.

1 hour ago, March Hare said:

All clouds are made up of voxels, not a type of cloud, which are the volumetric, 3D equivalent of a pixel.

Why does it only ever produce cumuloforms? Even the stratus layers I’ve seen (more frequently recently, thankfully) are all still flat-bottomed cumulus (which I know exist in real life but are certainly not the only type of stratus).

On 4/17/2022 at 4:22 PM, KERNEL32 said:

Asobo PLEASE consider hiring a team of professional weather experts/artists much like you did for many other parts of this sim and FIX this substandard atmosphere. And it will SHINE! But with the totally repetitive, ugly CLOUDS we've had since SU3 - our atmospheres will remain DULL and uninspiring to fly.

Maybe the weather experts were at work and you are exposing your ignorance? Are you aware, that a large number of cloud types exist in real world? E.g. clouds can look like these ones:

Föhnlinsen, Wolkenstimmung in den … – Bild kaufen – 70077951 ❘ Image Professionals

Or these ones:

 

12 minutes ago, mrueedi said:

Maybe the weather experts were at work and you are exposing your ignorance? Are you aware, that a large number of cloud types exist in real world? E.g. clouds can look like these ones:

Well, it appears he was flying over the ocean so Lenticular clouds wouldn't be seen there....and since we cannot get proper cirrus yet, I really don't think they have introduced a rarely seen cloud type such as Lenticular and doubt they ever will.  Asobo has mentioned they cannot yet produce the typical cirrus clouds we are used to seeing but might be able to stretch out the clouds using shaders to mimic them.  It doesn't sound like we will ever see the wispy sharp edged cirrus clouds we have been used to in the old sims that used texture-based clouds.

44 minutes ago, scotchegg said:

Why does it only ever produce cumuloforms? Even the stratus layers I’ve seen (more frequently recently, thankfully) are all still flat-bottomed cumulus (which I know exist in real life but are certainly not the only type of stratus).

It doesn't. The upper clouds ringed in this image from the first post aren't cumuliform. Predominantly it does produce that type, yes. But that will be owing, at least in part if not entirely, to the data.

 

2 minutes ago, Flic1 said:

Well, it appears he was flying over the ocean so Lenticular clouds wouldn't be seen there....

Lenticular clouds are often seen at the crest of wave created by jetstream interactions. Terrain not required, I've seen them over the mid-pacific.

I wouldn't call the clouds in the OP lenticulars, but again, there's nothing wrong with them. They very much look like a common type of cirrus layer, as opposed to the stereotypical mare's-tail variety. 

13 minutes ago, Stearmandriver said:

Lenticular clouds are often seen at the crest of wave created by jetstream interactions. Terrain not required, I've seen them over the mid-pacific.

I wouldn't call the clouds in the OP lenticulars, but again, there's nothing wrong with them. They very much look like a common type of cirrus layer, as opposed to the stereotypical mare's-tail variet

I stand corrected😄  In any case I don't see much wrong with these either....I had more of an issue with the volcanic ash clouds which I have not seen in the SU9 beta as of yet...

Wouldn't you just love to see and understand the tools at the disposal of the Asobo team working on this aspect of the sim, and learn what they discuss about everything that needs to be considered including performance impacts and so much more for clouds and weather.  The last recent interview one of the key players said they are going to be working on cloud morphology down the road.  It seems plausible that before they can really go there they have to decide exactly on the specific architectural features they are using to make clouds of any kind.  Perhaps as that gets better sorted out we will start seeing stabs at other cloud morphologies.  I know the specific way clouds are formed or at least rendered is very different from the initial release cloud structure.  Initially you could get very close to clouds proximity-wise and they looked vastly better than they do now.  They have taken on a certain plate-like structure that from a distance can be very effective but up close looks really terrible in spots.  And this methodology kind of seems doomed--using parallel plates to approximate 3 dimensions.  What is intuitive is making a cloud truly 3D from all directions, not plates which are 2D structures.  Presumbly this is secondary to efficiency needs and if that's the case I hope we see scalability for hardware that can handle it.  

50 minutes ago, March Hare said:

It doesn't. The upper clouds ringed in this image from the first post aren't cumuliform. Predominantly it does produce that type, yes. But that will be owing, at least in part if not entirely, to the data.

 

Those look like stretched / flattened cumuloforms. If not, what cloud type do you think they are?

38 minutes ago, scotchegg said:

Those look like stretched / flattened cumuloforms. If not, what cloud type do you think they are?

They look like an imperfect attempt to depict some form of stratus. Alto-, cirro-? They really don't resemble cumuliform clouds to me.

3 minutes ago, March Hare said:

They look like an imperfect attempt to depict some form of stratus. Alto-, cirro-? They really don't resemble cumuliform clouds to me.

They still have the clustered, bumpy edges of cumuloforms to me. I haven’t seen any clouds for a long time that don’t have these clustered, bumpy edges either in quasi or actual 2D or 3D.

Edit: although I agree they look like they’re trying to make stratus, maybe even cirrus, as I said in my original post.

Edited April 18, 20224 yr by scotchegg

1 hour ago, scotchegg said:

Those look like stretched / flattened cumuloforms. If not, what cloud type do you think they are?

They're wavy stratus; cirrostratus I would guess.  They lack any vertical development or defined updraft that would brand them as cumuloform.

Examples: 

https://www.google.com/search?q=+cirrostratus+from+air&tbm=isch&ved=2ahUKEwjoq7Tq3573AhXEGDQIHU-xCesQ2-cCegQIABAB&oq=+cirrostratus+from+air&gs_lcp=ChJtb2JpbGUtZ3dzLXdpei1pbWcQAzoHCCMQ7wMQJ1CSCViNDWCpEWgAcAB4AIABaYgBqwOSAQMzLjKYAQCgAQHAAQE&sclient=mobile-gws-wiz-img&ei=9vFdYujYFMSx0PEPz-Km2A4&bih=630&biw=360&client=ms-android-verizon&prmd=isnv#imgrc=OkRdXU1FzxbkSM

Edited April 18, 20224 yr by Stearmandriver

4 hours ago, March Hare said:

This isn't the case, though. Seb has tried to explain this at least a couple of times, to my recollection. All clouds are made up of voxels, not a type of cloud, which are the volumetric, 3D equivalent of a pixel. The data fed into the sim is supposed to tell those voxels what type of clouds to form into, in the live weather engine. However, Seb did discuss limitations. I think he was talking about, at the last mention of it, the necessary size of the voxel units, and needing to make them smaller in future for more, and finer, rendering complexity. He also referred to increasing the complexity of the data in future. I think he said it's possible to do these things, but I think he was talking about waiting to see what performance headroom is spare after implementing other things that people want or expect the sim to have.

I'm sure they hinted at the data being, perhaps, simply wrong sometimes, as they mentioned seeking new data partners.

But there are certainly going to be limitations in the volumetric rendering path. Home hardware still isn't up to the standard that I think users have in their minds. Perhaps if the sim wasn't doing many of the other things it's doing simultaneously, we'd see even more realistic cloud depiction already.

I'm personally still surprised at just how my 1070 card can render the current clouds on ultra setting along with everything else on ultra.

I appreciate your explanation of voxels, which is not a term I had heard before.

The sim is certainly capable of producing more realistic skies than it currently does though - Rex Weatherforce is using the same data to produce better skies. I have posted these two before and shows the sky from the same metar report.  There were two cloud layers, highest of which was overcast at 25,000ft. REX nailed both layers where Live Weather gave an awful rendition.  

Clearly MS/Asobo could represent the weather better than they do - perhaps REX could be part of the weather team?

Rex Weatherforce.

Live Weather

The next two images shows the issue again.  Two high level cloud layers were present. 

REX does a remarkable job with both layers.  Look closely through the bottom layer (top left especially) and you can see a whiter cirrus layer represented. 

Live weather shows the high layer but to my eyes is not depicted as well. To be fair to MS/Asobo, this sky doesn't look bad at all, but not as good as someone else did with the same data.

Edited April 18, 20224 yr by MrBitstFlyer

2 minutes ago, MrBitstFlyer said:

Rex Weatherforce is using the same data to produce better skies

This is very interesting. I’ve been swapping between REX and MSFS live weather for the last few weeks, and overall it seems REX is more convincing.