Jump to content

1.17 Development


Captain Kitten

Recommended Posts

Spherical buffer = Perf upgrade nr 2

This update is basically another performance update. This was one of the 5 performance paths we had in the 1.14 update plans. We did not implement it then since it will take us at least a month to implement and test if it even works in the X-Plane environment. We decided to add it later so we wouldn't delay 1.14 longer than necessary. Fingers crossed that we're not in a dead end. The reward will be great if it works and so far it looks promising!

photo_2021-01-12_08-43-37.jpg.9fa3384690d79f47904f801bcd4947d9.jpg

First ever image of Spherical Buffer in X-Plane. And in any sim as far as I am aware.. Don't quote me on this 😄
Yep - Minecraft comes to mind. It's the tech that counts at this point.


With the Spherical Buffer we hope to be able to get to the point where we can start adding more features like thunderstorms, multi screen support and VR. If you wonder what Spherical buffer is - read up on space observation tech. How we implement it and parts of it's functions are trade secret at this point 😉

photo_2021-01-12_08-43-44.jpg.6ade7315f22c9ea1de08bf856b6ce6f7.jpg

With a bit of tweaking we're back to more familiar clouds. But with some huge tricks up their sleaves I seriously hope we can exploit. Crossing fingers and toes that this will turn out well in the end. We just won't know until we add back all features in the next test runs.

photo_2021-01-12_08-51-57.jpg.fa34394b7744d7c75a55889a3902501e.jpg

photo_2021-01-12_08-51-49.jpg.9ae397f838587f7c570ef94475638802.jpg

Let me explain in short what happened in the past updates:

1.14

  • Major performance update
  • Additional functions and features as a result of a complete rewrite of the core
  • This was the incredibly long update that took a year to finish... 

1.15 

  • Vulkan update (11.50 compatible)
  • Dynamic seasons only available in OpenGL since the needed access is not available in Vulkan.

1.16

  • X-Plane 11.51 hotfix (X-Plane 11.51 and above)
  • Not compatible with earlier versions of X-Plane (in Vulkan)

 

A comment on dynamic seasons
I see that there is confusion about dynamic seasons. And as much as we want dynamic seasons ourselves, it has never been a part of our feature list. It is an experimental feature because we could add it in our pipeline and get it to work. In OpenGL. We will try to convince Laminar that X-Plane needs dynamic seasons and MAYBE we will get the access needed to actually execute that feature without the flaws it suffers under in its current state in OpenGL. So far we have had a closed door but we will continue the talk. (We have a poll running in the facebook group where you can add your vote for our seasons - so maybe Laminar will see the community whish for such a feature.) You can also address this to Laminar directly. Any pressure from the community will help us in getting the access we need.

But I repeat: dynamic seasons has never and will never be promised as a feature in xEnviro. And we are running two pipelines of development currently to keep your whishes for this feature active in OpenGL. How long we can keep this up is a huge question mark though.

photo_2021-01-12_08-52-06.jpg.662e45b2d3267042835466991a60444b.jpg

photo_2021-01-12_08-52-01.jpg.583169c541cfcab374d7b97f0519213a.jpg

 

Link to comment
Share on other sites

hello world

1.17 task #1: Spherical buffer - ongoing

Status: 1.17.t30

  • Spherical buffer initial proof of concept test - done
  • Spherical buffer core code conversion (t01 - t28) - done
  • Spherical buffer pipeline initiated (t29 - txx) - done
  • Spherical buffer adjustments and adaptation - ongoing
  • Spherical buffer evaluation - ongoing
  • 1.17.txx test team push - imminent
  • 1.17.bxx test team push - not started

a319_-_2021-01-04_14_29_23.png.77e4445d525e0ff9d7e28dbf15a17d23.png

(Image by beta test team member xE 1.16)

I have been pretty silent lately, so far this year in fact, and I apologize for that. In a sense I am eating my own words at the moment. "...but we are trying to get away from the huge gaps between releases."  1.15 was released 25th of december. 1.16 was the Vulkan hotfix and... Yeah. Time flies. But there has been a lot that has demanded the teams attention in real life both in regards of our day jobs and privately. So progression on the new update has been slow. Since we have some new users asking about the known issues we listed in 1.15 I'll repeat them here. Several of these will probably be solved by the spherical buffer conversion and if they persist we will have room for improvement in later updates. 

So here's the list of known issues

  • Precipitation follows camera
  • Snow coverage may appear in the regions where no snowfall encountered recently
  • Default X-Plane night triggers lights artifacts
  • Cloud edge artifacts
  • In cloud artifacts
  • Cirrus clouds may affect colorization of clouds beneath
  • Low level clouds may be combined

You can find an updated list at xenviro.net

To add to the silence, there has not been anything interesting to show visually at this stage of 1.17 development. xEnviro has been in peaces again. 1.17 has just left pre' Alpha state which is just pure code. We are currently done with the code conversion and are now adapting all of xE code to the new tech. We are constantly evaluating if this is a sensible direction to take.

If you are wondering what spherical buffer is, you're not alone. 🙂 I won't disclose what spherical buffer is other than that it's nothing new. But as far as we know this has not been done before in this application, let a lone in X-Plane. So we're treading unknown grounds again but we're taking this risk because the benefits are huge if we pull it off. Treading new ground is nothing new. We were the first to go public with volumetric clouds in flightsim scale. We think that was the only sensible choice at the time and one of the best decisions we have ever made albeit controversial and demanding. 

As per usual I do not give any ETA's for the release but we have made progress and this is just an update to tell you all that we're still alive and working on 1.17.

 

 

Link to comment
Share on other sites

  • 5 months later...

Alright,

let me first say that I am sorry for not updating all of you for a while. I haven't had the time for it first of all and second: I don't believe in vacuum hyping. Which means hyping something for sales alone when you can't back it up. But I see that several people question if we're even alive - or if xEnviro is in a dead end. So here is an update to lay that ball dead. Some won't accept any form of "excuse" what so ever and will always demand an update to xEnviro regardless. Let me just start with answering those with the fact that you bought the version you got at purchase. Not a certain update cycle.

With that said:

We're still developing. It has taken a looong time and I am aware that this has been the case for several of the major updates so far. Development has been slow because of several factors. First of all we have day jobs that have demanded a huge chunk of our lives. Ever since Covid started several of us in the dev-team have what is called critical professions that are under high pressure and have been for a long time. Apart from that we have private lives that also demand it's time and in periods have made it impossible to concentrate on other things. xEnviro is developed by a small team and we're vulnerable to halts.

unknown.png.c5c0918d8a88567fe09ac598bb53a15c.png

Image from 1.17t43

So where is development per today?

Spherical buffer is a complex transition and a new tech developed by us within noise rendered volumetrics. It will give huge benefits both in performance potential and visually. How much will remain to be seen. At launch of 1.17 clouds will probably not look very different than 1.16 (apart from artifacts fixes and performance). But xE will rest on a structure that we hope will give it a huge advantage going forward. This is the 5th performance path that has been mentioned in the past. Also in XP12 when that launches. But about that...

We're still waiting for some clear answers from Laminar in regards to if we get the needed tools to continue developing weather at this level in xp12. We know that xp12 most probably will get volumetric clouds but the tools have been very limited so far and this might be even worse in xp12. xE has always been independent of Xplane to a large extent and that will be our savior also this time around. We'll see if laminar wants to have 3rd party weather addons in the future but if not; we'll manage to brute ourselves in. As we have done for the entire xE life so far anyway.

Anyway, sorry for not having updated you and that this update is not of a "here's what we got" or with images of 1.17. We don't have that much to show anyway since the work is on the back end. 1.17 is over 90% done. But the remaining 10% doesn't translate in to time. I won't promise you an update soon either. I am currently stealing time from my day job to write this and I have to get back to that now. But if we have any news worth mentioning I will update asap.

 

Link to comment
Share on other sites

  • 2 weeks later...

1.17t57 

Before I start

let me just say that we will not stop development with the release of XP12. We firmly believe that we will still be relevant in X-Planes future. The teaser shots released from Laminar proved that to us. I had a whole rant laid up after a discussion with Andrey about the screenshots Laminar presented. But I quickly scrapped that rant since it basically would be a long educational text on how light behaves in the real world and why the lighting model Laminar has chosen is completely unrealistic. XPlane12 still doesn't have atmospheric calculations for light scattering. Everything is coming from palettes. Photometric means "something metered". Metered in advance and represented as a statistical table. And Laminar's choice of the light source having a set color is just 🤦‍♂️..
This is why XPlane12 still won't have realistic lighting and will be dependent on addons to get sky colors and general light behavior right.

RH4oxqpmDLf65XvNRd0ThsGMHW2cgD181qeYaNw8RsWZFm9kdxn4ANiWDy4pGNR0Nvyq1c5pghq1o4w1k8Syaom5P-FMTmnkC_oZr-1aOiwh5F9YymD5BdvHVAv0BhymP1j17EMD=s0
XP12 released teaser image
 

Where are we?

Here: 

  • 1106284349_2021-10-0312_10_22-xEnviroDevCenter-Vivaldi.jpg.b277c4e788e8ac1930776c2012d110a4.jpg

Spherical forward buffer is done. It's finally done! And it works!

1927198570_2021-10-0113_11_49-unknown.png-IrfanView.jpg.46dfec8326e5869d3261270c782b245b.jpg
First images in the new true full volume rendered environment based on spherical buffer. (1.17t57)
716346840_2021-10-0113_12_49-unknown.png-IrfanView.jpg.4c8780bed4d3e1c080b17776b89abc93.jpg
Not in any way representative of final result! A lot of the 1.16 code is obsolete and needs to be cleaned out. 

Spherical buffer has never been used before in this application and is a tech we are proud of pioneering in the industry. We were the very first with volumetric clouds in a major sim. And we are first out with this tech as well which we hope will open up some possibilities in X-Plane with all kinds of cloud formations and wanted features. We have good results. As it looks now we will have some possibilities I am reluctant to publish.

This is a whish list based on what we have seen - some items might be blocked later depending on the path we choose (And Laminar's access in XP12) but we're hopeful that all will be doable with time.

  • General performance increase for all resolutions
  • True full volume rendering. (this is already done and gives us a huge advantage)
  • Unlimited cloud layers
    • Mixed cloud layers
    • Unlimited formations
    • Thunderstorms
  • Cloud on cloud - and self shadowing
  • High performance on high resolution clouds
  • Wake vortex turbulence effects
  • VR compatibility

Not all of this will be in 1.17 but the road to get to the items above is so much shorter now. And we won't add to the current future feature list now but we're dreaming of some possibilities now that we knew could be done before but with huge compromises we might not have any more. We're already at a 10-20% performance increase from 1.16 and we haven't even started optimizing the code. Where we end up is not a set figure. We'll see. 

231117964_2021-10-0312_02_56-xEnviroDevCenter-Vivaldi.jpg.8b734127faa79e20cb35778edfd44b70.jpg

We currently have 14 items left on our agenda. This can change if any of them gives us problems. But as it looks now we should have a set item list to finish. Light is observed at the end of the tunnel
😃 Back to work!

image.png.e29361b26e9ae5e2694576c28a87d55a.png

 

Link to comment
Share on other sites

Progress. And a lot of it lately. 1.17 is shaping up well. I'll try to make a more detailed update this time. I usually don't have time for this but upon request I'll try to squeeze a couple in.

image.png.2f9f216761e4e498c8bdc2b43d3b6da0.png

Following screenshots are work in progress. Blending, artifacts and so on will be mended. Performance is still at 15-20% even with the higher detailed clouds we have now. And it is foreseen to stay like this when we start mixing cloud layers. As mentioned before Spherical buffer unlocked seamless layer transitions. So we hope to be able to mix cloud types soon to execute one of the final proof of concept items. If it were to fail we're confident that Spherical buffer is the way to go regardless. But if it works we'll have cloud rendering on a level we didn't think was possible when we started out with volumetric clouds. Exciting times. 

In the following images ignore blending and artifacts. We're shaping cloud types. Initially we have planned eight formulas for clouds to form from.

719555014_2021-10-3000_24_03-unknown.png(25601080)-Vivaldi.jpg.a10b2224f735cdd7386cd9356b3351c4.jpg

1.17.t71

Details have been boosted pretty heavily from 1.16. We hope to not have to reduce that too much. If we get any artifacts like we had in 1.16 we will reduce detail. But so far we're well within our aim and we haven't even started optimizing yet. 

1936819231_2021-10-2919_55_26-unknown.png(25601080)-Vivaldi.jpg.4b807145e28e130a624067587924cc88.jpg

904455201_2021-10-2919_43_09-unknown.png(25601080)-Vivaldi.jpg.de137ffb201490d28947186a353acb57.jpg

1.17.t71 above

1.17.t72 below

1777875482_2021-10-3000_13_30-unknown.png(25601080)-Vivaldi.jpg.ca6d595f15d575eae5b145931cf20f55.jpg

1444072834_2021-10-3000_12_55-unknown.png(25601080)-Vivaldi.jpg.8ae8a01b37cf4547f2161f61d65902c8.jpg

I can't stress enough how much work in progress this is. 

 

 

 

Link to comment
Share on other sites

  • 3 weeks later...

xE 1.17t75 The first video of our shaping test runs. This is taken on a laptop with a 1060. We're currently at between 35-80 fps
(not optimized so fluctuating a lot) So we're still well over our goal for performance.
This does not say anything about the end result. Still a lot to be done. But a lot has been done too so we're happy.


https://www.youtube.com/watch?v=7z5a8ZxCxig

Link to comment
Share on other sites

1.18t80 / t81

While we are shaping cloud types we disable several items and filters so just like any other post I have made take it as work in progress. I know many of you want to know how much is left and when we release but I still won't give ETA's. So I'll update with the Nürburgring analogy 😉

We're right here:

image.png.17259256b6e1878a21956b9f5dbee168.png

We're at the (1) mark. Or maybe even a bit further. Things are moving fast these days.

So yeah. It's been a while but most of the time that has passed since 1.16 update has been occupied by day-job and life in general that has been more demanding than ever. Now that there has been time for xEnviro we're racing towards the finish line. We still have a lot to finish but the recent progress has been very fast and very visual. 

Have some more teasers:

1622086909_2021-11-2016_55_14-AsPowerBar.jpg.ed70cc9ed2410b872f9d91c90ad72645.jpg

258792364_2021-11-2017_04_00-AsPowerBar.jpg.50499a452a113a2b2b68852e04995c5b.jpg

 

897915610_2021-11-2022_40_16-unknown.png(25601080)-Vivaldi.jpg.306f3ed5c0ca1467f712ac122ee4e02a.jpg

874369699_2021-11-2022_39_48-unknown.png(25601080)-Vivaldi.jpg.03136543d12cdc6acdf2afb10ee00e3c.jpg

372630963_2021-11-2022_39_39-unknown.png(25601080)-Vivaldi.jpg.73a77df6a63ac2c8a480af63c4c18106.jpg

769758413_2021-11-2022_39_30-unknown.png(25601080)-Vivaldi.jpg.d7aab6852b508edc8c201f60cdaa2201.jpg

(I've opened up for commenting - but I do want to keep this thread to still be about the update. So any other conversations will be deleted. I will be strict on that one 😉 )

2021-11-20 22_39_22-unknown.png (2560×1080) - Vivaldi.jpg

Link to comment
Share on other sites

this thread will soon heat up, having come that far, like the good days we had in recent release cycles.

am always keeping an eye on this precious hiding place.

having watched the last published video, i see it very much promising to end up better than msfs cloud depictions at near distance.

Link to comment
Share on other sites

Bought in at the start of version 1.10 and always been amazed at the different sky scapes and sky colours X Enviro produces with Its own weather engine, all this on a middle of the road desk top, looks so real and the new rendition looks to even surpass this, really looking forward to this version, are the frame rates still looking respectable in your opinion and are you slipping in any surprises for us that you would of liked to do in previous versions.

Link to comment
Share on other sites

  • Recently Browsing   0 members

    • No registered users viewing this page.



  • Don't have an account?

    An account is required to interact with and download content from the Threshold Forum.
    Forum is currently also locked down for guests because of heavy bot invasions. Sorry about that.

  • Threshold News

  • image.png.5d3ebd5638179a05dc5cfa67eaf4bf3a.png

×
×
  • Create New...

Important Information

Please read the Terms of Use