1 hour ago, akita said:

Question is what happens to those extra cpu cycles with the new flight model and other simulation features unrelated to graphics.

I believe the immediate benefit is that plugins will have more room for computations on the CPU 🙂

1 hour ago, Gulfstream said:

I was just reading through some threads on "another" forum from only a few years ago regarding X-Plane's clouds.

Everyone was mentioning how we need 3D clouds but that LR wasn't listening, and others coming to the defense that this "simply isn't possible" and "would kill every machine's frames".

When it was pointed out at the time, and this was years ago, that other sims do it without an issue, there was all sorts of excuses from the naysayers that "those don't deal with real time weather", "the abrupt transitions are necessary due to the complexities of X-Plane", etc.

Heh. Feels like it was yesterday! And the next step past "not possible/not necessary" is of course "other priorities", as can be observed even in this thread 😄

11 minutes ago, RXP said:

Prior to this was at least FSW

There were a lot more failed attempts other than those in my list, for example xEnviro had 1.09 in May 2018

I didn't include them because it was more a question of when volumetric clouds actually became technically feasible, and who achieved it/when.

Personally I dont think any of them have come close to AC7 (yet) - they still set the bar.

 

Edited November 15, 20214 yr by mSparks

9 years ago, this was possible (yeah it's technically not a flightsim):

9 minutes ago, rka said:

9 years ago, this was possible (yeah it's technically not a flightsim):

Sure, but its still garbage compared to what it is actually trying to represent:

 

26 minutes ago, mSparks said:

Sure, but its still garbage compared to what it is actually trying to represent:

This XP12 preview with the Varjo Aero is stunning! 🙂

 

59 minutes ago, mSparks said:

Sure, but its still garbage compared to what it is actually trying to represent:

That's not the point. The point is that this stuff is a decade old.

 

29 minutes ago, rka said:

That's not the point. The point is that this stuff is a decade old.

 

It is the point, it doesn't look as good as cloud art and performs worse, that isn't an improvement.

These clouds are all based on

https://www.semanticscholar.org/paper/Real-time-rendering-of-volumetric-clouds-Häggström/89e9153a091889c584df034a953a0eff4de45ee9

published in 2018.

There is none before that because it (realtime volumetric clouds) wasn't invented before that.

Its also only half the problem - its one thing to get real time rendering of volumetric clouds, something else entirely to generate the volumes in a manner that looks remotely realistic - or even believable.

Edited November 15, 20214 yr by mSparks

1 hour ago, rka said:

9 years ago, this was possible (yeah it's technically not a flightsim):

...

Wake me up when there is hardware that can run ArmA 3 with a rendering distance that is comparable to a true(TM) flight simulator.

6 minutes ago, Bjoern said:

Wake me up when there is hardware that can run ArmA 3 with a rendering distance that is comparable to a true(TM) flight simulator.

Good morning from a few years ago.

3 hours ago, RXP said:

Instead, I'd bet this was more a combination of Dev & Time resources already committed on refactoring entirely the underlying rendering engine architecture to cope with modern constructs imposed by VK as opposed to "managed" constructs imposed by OGL. It would have therefore made no sense to invest time in doing volumetric clouds prior finishing the refactoring first.

The same reasoning goes for the GPU based water waves already implemented a long time ago in XP11 code base but never released to the public, and many other things "rendering" related in the X-Plane product line.

In any case and whatever the reasons, the fact they are doing this with XP12 is just good news anyways!

@RXP You are spot on. 

They milked their cash cow as long as it was possible (which means in this case: until a strong competitor showed up).

Yet people go to great lengths to explain why they had better to do over all those years than improve their product.

I really laughed (and cried a bit) at the part "they improved the waves long time ago, but didn't release the improvement to us customers and that's a good thing!".

Great clouds! XP11 already surpasses the competition in many things, now it will also surpass them in the environment and weather department! 😍

Edited November 16, 20214 yr by Murmur

5 hours ago, rka said:

I really laughed (and cried a bit) at the part "they improved the waves long time ago, but didn't release the improvement to us customers and that's a good thing!".

Yes it was good thing. Had they release it before finishing Vulkan, it would had the performance of a painting in the Louvre Museum in Paris, giving you more reason to complain. 

They didn't that way because they knew better, they maybe slow at getting things out but they're not stupid.

Edited November 16, 20214 yr by BobFS88

36 minutes ago, BobFS88 said:

it would had the performance of a painting in the Louvre Museum in Paris, giving you more reason to complain. 

The main issue with the water update back then was it didnt react to land

Not really seen any sign that issue has been fixed...

For those who havent spent much time with non vulkan xplane 11, or never really gave much thought to why non vulkan xp11 was such a nightmare (e.g. because of flying most of the time on autopilot)

It always got talked about as "stutters" - "fixing the stutters" etc. But this description really didn't do justice to just how huge a performance issue they were, especially when visually they could be barely visible or easy to ignore.

The visual aspect of the stutters was only a tiny part of the problem.

By far the bigger problem was that during a stutter any manual control input would be ignored, which would then be compounded by xplanes flight model calculations exploding to nonsense for that short time because there was so much more time than 50ms between the two frames (the maximum time between frames that flight model calcs are valid), the net effect of which could easily be your aircraft randomly falling out of the sky for no apparent reason (when not on autopilot)

It is "a good thing" they threw all hands on deck to fixing that at the expense of everything else they and everyone else would have liked and had been working on, because opengl in xp11 was/is basically unusable for hand flying. (they also exist in xp10 but are rare enough that no one noticed)