Austin Interview on Threshold

[Kopteeni 819]

26 minutes ago, Airfighter said:

Compare XP9 with XP10 and see how big of a leap it was. Now compare XP10 with XP11 and see. again, the big leap here. Now, standing on XP11, envision how big the leap will be, and if will land XP12 into "next gen game" territory. 

Most do not acknowledge this: Up to X-Plane 11 release (4 years ago), everything has to be written for OpenGL, and fit within 6 DVDs. And still X-Plane 11, in many aspects was a big leap. You also know that you have, within the next years, to re-write the whole graphics engine, and once you start doing so, you can't go back and add stuff to the sim. It is like building the foundations of a new house and someone wants to put in the furnitures! Now that Vulkan/Metal is closer to be "stable", will come the time to move to the next gen stuff.

If you give X-Plane, as is now, volumetric clouds and improved atmospheric lighting effects, you are almost there.

What we know so far. Ben said that XP11 was the last version that has that type of cloud rendering. We know that a much improved autogen is on the works. 11.50 beta users, might have noticed that there is a small improvement in atmospheric lighting. Not some that blows you away, but a start. Orthophotos... next X-Plane version, probably will be just compatible, so if Laminar provided or not, is almost irrelevant. And there more things that Laminar is already doing, which will be announced when they think is good to do so.

So, you don't need faith. Next X-Plane version (probably just X-Plane?) will be there with the next gen games!

And I leave you with this. Does MSFS has fantastic clouds? Yes sure. Can anyone else do it? Yes sure, should the norm from now on...

 

thanks for the clip and woah. Never heard of UHawk before but holy hell.

[CFIJose 5,485]

1 hour ago, Airfighter said:

Compare XP9 with XP10 and see how big of a leap it was. Now compare XP10 with XP11 and see. again, the big leap here. Now, standing on XP11, envision how big the leap will be, and if will land XP12 into "next gen game" territory. 

Most do not acknowledge this: Up to X-Plane 11 release (4 years ago), everything has to be written for OpenGL, and fit within 6 DVDs. And still X-Plane 11, in many aspects was a big leap. You also know that you have, within the next years, to re-write the whole graphics engine, and once you start doing so, you can't go back and add stuff to the sim. It is like building the foundations of a new house and someone wants to put in the furnitures! Now that Vulkan/Metal is closer to be "stable", will come the time to move to the next gen stuff.

If you give X-Plane, as is now, volumetric clouds and improved atmospheric lighting effects, you are almost there.

What we know so far. Ben said that XP11 was the last version that has that type of cloud rendering. We know that a much improved autogen is on the works. 11.50 beta users, might have noticed that there is a small improvement in atmospheric lighting. Not some that blows you away, but a start. Orthophotos... next X-Plane version, probably will be just compatible, so if Laminar provided or not, is almost irrelevant. And there more things that Laminar is already doing, which will be announced when they think is good to do so.

So, you don't need faith. Next X-Plane version (probably just X-Plane?) will be there with the next gen games!

And I leave you with this. Does MSFS has fantastic clouds? Yes sure. Can anyone else do it? Yes sure, should the norm from now on...

 

Very nice!

[Colonel X 1,430]

UHawk seems to be far ahead even compared to FS2020.

[Ianrivaldosmith 5,317]

That looks good. I’m amazed no one has made a decent volumetric cloud system for xplane. A good one. Or is working on one for vulkan. I’m sure everyone would buy it

[Pastaiolo 1,688]

xEnviro is the closest thing to that, with all its downsides including slow development (few guys on it), the performance impact and how they have to build it on workarounds because of default XP weather limitations.

I really do hope XP12 will offer something pleasant from the start, no more billboards and 2d clouds.

[Airfighter 150]

2 hours ago, Colonel X said:

UHawk seems to be far ahead even compared to FS2020.

Now, FS2020 does not seems out of reach, right? 

2 hours ago, Ianrivaldosmith said:

That looks good. I’m amazed no one has made a decent volumetric cloud system for xplane. A good one. Or is working on one for vulkan. I’m sure everyone would buy it

 

1 hour ago, Pastaiolo said:

xEnviro is the closest thing to that, with all its downsides including slow development (few guys on it), the performance impact and how they have to build it on workarounds because of default XP weather limitations.

I really do hope XP12 will offer something pleasant from the start, no more billboards and 2d clouds.

With OpenGL is almost impossible to have volumetric clouds with big performance penalties.

[Ianrivaldosmith 5,317]

36 minutes ago, Airfighter said:

With OpenGL is almost impossible to have volumetric clouds with big performance penalties.

Yes, hopefully someone picks this up in Vulkan, because in P3DV5 truesky, using a modern graphic card, there appears no performance penalty 🙂 

[Airfighter 150]

57 minutes ago, Ianrivaldosmith said:

Yes, hopefully someone picks this up in Vulkan, because in P3DV5 truesky, using a modern graphic card, there appears no performance penalty 🙂 

This will be Laminar. I don't know if ever Laminar allow 3rd parties to draw in 3D in Vulkan. It is not trivial.

[Paraffin 2,197]

9 hours ago, mSparks said:

I am more than 95% certain "next gen scenery" means significantly more than just more landmarks. At the very least current gen scenery ala witcher 3 is implicit in such a statement.

That can't be done without massive increase in user storage space. That's the Achilles Heel of XP's local scenery method vs. MSFS streaming approach.

Since we're using Witcher 3 as a reference, that game takes up 57 GB on my drive (including a high-res texture mod). The scenery is gorgeous and detailed, but it's not actually a large physical area if you're flying over it at aircraft speed instead of traveling on horseback.

Unless Austin moves to a streaming model in XP12 with the infrastructure to support it, there will be a limit to how far scenery quality can be improved going forward. I don't see XP12 being very popular if it requires terabytes of storage space to load Witcher 3 quality scenery for the entire world. A few of us FS geeks will set up systems like that to use ortho scenery, but the general user market won't support it. Especially with game developers and platforms moving towards streamed "games as service" models. 

Not trying to be too negative here, and I do want to see what Austin & Co. come up with for the next version. Maybe improved procedural terrain textures + procedural seasonal changes + landmarks and far more regional autogen building varieties could be "plausible" enough to compete with the ortho approach.

 

[Paraffin 2,197]

Scenery aside, I think the main risk for XP in the future (which Tony mentioned earlier) is developer migration, especially aircraft developers because that's really the heart of the sim. The planes in the default hanger aren't enough.

As it is, we're often snubbed by developers. I've been waiting for an XP version of Carenado's Beech 18 ever since the FSX/P3D model was released. I love that plane, it's always been one of my favorites. All we have is an ancient Heinz model, good for its time but nowhere near state of the art. At this point it looks like we'll never get that new model for XP, either because sales would be too low, or Carenado is shifting focus to MSFS. 

Future versions of XP will no doubt break the engine and systems modeling of current XP11 versions, like they've done before. Even within the current product cycle! That's just Austin's need to continually perfect the sim. I don't see that trend stopping, and the aircraft devs don't commit to support across major versions anyway.

If aircraft developers don't release new XP12+ versions because they're making more money with MSFS, we'll be back to XP being Austin's personal hobby sim with just a very few models to fly.

We're not there yet, but it's a danger. It doesn't matter how good Austin's flight modeling is, if he gets no 3rd party dev support for new add-on aircraft.

Edited June 25, 2020 by Paraffin

[Airfighter 150]

4 minutes ago, Paraffin said:

Scenery aside, I think the main risk for XP in the future (which Tony mentioned earlier) is developer migration, especially aircraft developers because that's really the heart of the sim. The planes in the default hanger aren't enough.

As it is, we're often snubbed by developers. I've been waiting for an XP version of Carenado's Beech 18 ever since the FSX/P3D model was released. I love that plane, it's always been one of my favorites. All we have is an ancient Heinz model, good for its time but nowhere near state of the art. At this point it looks like we'll never get that new model for XP, either because sales would be too low, or Carenado is shifting focus to MSFS. 

Future versions of XP will no doubt break the engine and systems modeling of current XP11 versions, like they've done before. Even within the current product cycle! That's just Austin's need to continually perfect the sim. I don't see that trend stopping, and the aircraft devs don't commit to support across major versions anyway.

If aircraft developers don't release new XP12+ versions because they're making more money with MSFS, we'll be back to XP being Austin's personal hobby sim with just a very few models to fly.

We're not there yet, but it's a danger. It doesn't matter how good Austin's flight modeling is, if he gets no 3rd party dev support for new add-on aircraft.

I don't know where is the source of your assumptions, but they definitely wrong. All 3rd party developers are talking with Laminar every day, all day, and I can tell you that if an aircraft will work on XP11 Vulkan, will work almost unchanged in the new version. 

[mSparks 2,086]

1 hour ago, Paraffin said:

That can't be done without massive increase in user storage space. That's the Achilles Heel of XP's local scenery method vs. MSFS streaming approach.

Since we're using Witcher 3 as a reference, that game takes up 57 GB on my drive (including a high-res texture mod). The scenery is gorgeous and detailed, but it's not actually a large physical area if you're flying over it at aircraft speed instead of traveling on horseback.

Unless Austin moves to a streaming model in XP12 with the infrastructure to support it, there will be a limit to how far scenery quality can be improved going forward. I don't see XP12 being very popular if it requires terabytes of storage space to load Witcher 3 quality scenery for the entire world. A few of us FS geeks will set up systems like that to use ortho scenery, but the general user market won't support it. Especially with game developers and platforms moving towards streamed "games as service" models. 

Not trying to be too negative here, and I do want to see what Austin & Co. come up with for the next version. Maybe improved procedural terrain textures + procedural seasonal changes + landmarks and far more regional autogen building varieties could be "plausible" enough to compete with the ortho approach.

 

Have micrsoft actually solved the god awful clouds on the sat photos yet?

That hard problem alone makes it unsuitable for a quality world.

Edited June 25, 2020 by mSparks

[Paraffin 2,197]

1 hour ago, Airfighter said:

I don't know where is the source of your assumptions, but they definitely wrong. All 3rd party developers are talking with Laminar every day, all day, and I can tell you that if an aircraft will work on XP11 Vulkan, will work almost unchanged in the new version. 

That's good to hear about the 3rd party devs, but it still doesn't get me my Carenado Beech 18. 😉  I know, just one example but I think it's relevant. 

Regarding XP11 Vulkan -- I wasn't referring to the current cycle, but what developers will need to do going forward as XP12, XP13 etc. will inevitably involve changes to the underlying systems that may break backwards compatibility. Austin doesn't stop moving. 

That aside, developers need to give customers a reason to buy new updated versions  of their models anyway. They don't make any money on XP11 models they've already sold.
 

15 minutes ago, mSparks said:

Have micrsoft actually solved the god awful clouds on the sat photos yet?

That hard problem alone makes it unsuitable for a quality world.

That will probably never be solved in remote areas, along with other typical ortho artifacts like squashed cars. However, in areas with photogrammetry that get more attention, it shouldn't be a problem. It's those areas that will sell the approach for many people. 

[mSparks 2,086]

1 hour ago, Rob_Ainscough said:

I'll disagree that Vulkan is faster and has better implementation of hardware accelerated ray tracing.  DX12 + DXR extensions were implemented and deployed before Vulkan started working on it's own extensions for ray tracing.

Vulkan is certainly better than OGL, but lacks considerably in the 1% low FPS arena at 2160p.  

I think 4K gaming/simming is pretty common these days and that seems to be where Vulkan suffers on the min FPS front.

Cheers, Rob.

Linux has had hardware raytrace support (and been in widespread use) since something like 2006, the VK extensions come from that.... DXR wasn't announced until 2018 - 12 years later and is still mostly untested, besides a couple of glitchy attempts from a couple of titles..... 

Edited June 25, 2020 by mSparks

[Ianrivaldosmith 5,317]

Has anyone seen the new MSFS videos today? They actually look like Ortho4XP.

Imagine XP11 loaded with Ortho. Then imagine we could have really good AA and weather. Thats what to me, you're looking at in the videos. I noticed the waves were baked in.