P3D future development

[kevinfirth]

Here are my thoughts, in no particular order:
1. the nav database. FsAerodata is stopping, keeping the impenetrable nav database is a factor in that. A reworked nav database would be great...why not make a deal with navigraph? One historic Airac bundled with the sim, and regular updates with our navigraph subs?

2. An easier development environment with a better implementation of the sloped runways functionality. You can make it work ok with default assets, but I understand that it's almost impossible to overlay a ground poly over non flat bterrain (hence why no third party devs have taken advantage of the new feature, opting to remain with full flat airports instead.) MSFS has tools that make this much easier I think. Perhaps LM should give some consideration to creating a tool that allows both quick and easy airport terrain management, (preferably through a GUI in SIM), and the quick and easy overlay of ground polys to reduce the investment of time necessary. All the scenery content development tools over the years have been borne from community efforts, and have become standard issue, but it has taken years for them to mature. I think LM has until PMDG sorts it's word not allowed out and gets add-on aircraft in MSFS, otherwise I can see P3D users further splitting in favour of MSFS (warts and all) which could inexorably lead to the add-on development ecosystem for P3D becoming much smaller and unprofitable completely. If LM wishes to retain the value of the ecosystem, it needs to help and protect it by implementing features in such a way to make development easy, not hard!

[kevinfirth]

On 3/6/2021 at 6:20 AM, GSalden said:
I would prefer a complete overhaul of the AG.
In P3D all houses are separate while in real life there are many house blocks with houses attached to eachother...
Orbx has made houseblocks in their True Earth series but they are connected to the Scenery Complexity slider.
Some addon developers have them with custom products. Most of the time all houses are still seperated from eachother.
Beside house blocks I would like to see houses having more realistic colors and a more realistic build.
Compare the AG houses from MSFS with P3D and see the difference.
Even with LC scenery there is room for improvement.

All of that is already possible within the P3D SDK, and has been since FSX.

Earth Simulations did some amazing autogen with chimneys, solar panels etc even in FSX times. I think other devs have never exploited autogen to its true extent because it's a bit harder to do, there used to be a significant performance impact, and a lack of demand from customers.

Rowhouses have been in the SDK since forever. Different textures and autogen models have been possible since at least 2006.

Again, what LM could do is make it easier to do this kind of stuff. Back in 2015 Earth Simulations used Arno Gerretson's tools to place custom highly accurate autogen using over 30,000 lines of code...

[kevinfirth]

I've got a few that are a bit more 'out there'...

Automatic interpolation of segmented taxiways into smooth curves. We've all seen ATC aircraft 'shuffling' step-by-step through a turn on the ground. It's very immersion breaking. P3D should recognise this taxiway pattern and smooth it out automatically. Problem solved.
I'd really like to see them open up the sim more for developers. Not that it isn't open now, but it could be *more* open, and more simply. I'd like .NET bindings for the PDK, which I don't think exist at the moment because I don't think you can have in-process DLLs written in .NET (I could be wrong, but that's how I read the docs). I'd like to be able to write native gauges in C# vs C++.
A way to run standard gauges (C++ and XML) outside the sim process on remote PCs. Of course right now you can undock and drag a gauge to another display, but the gauge code still runs in the main sim process. It ought to be possible to build a network-connected virtual machine in which normal gauge code can run without alteration but which is outside the sim itself so you can move FPS-unfriendly gauges to separate PCs. Obviously you can achieve the same goal with avionics packages like ProSim etc, but there isn't offboard avionics software for most aircraft. This would probably be very complex to do, but hugely worth it for advanced setups.
Basic multichannel capability in the Academic / Pro editions. Maybe limited to 3 channels. Caters for lower-end multichannel setups which currently have to use Wideview because multichannel via Pro Plus is way too expensive and aimed at highly complex commercial sims. Allow us to turn off the frame sync so each client can run as fast as possible.

Edited Friday at 09:47 AM by neilhewitt

[kevinfirth]

My thoughts on why some features arent used as much as they could be...

1. Workflow often involves a lot more work, creating models and texture sets.

2. The autogen system in FSX (carried over into P3D) is more forgiving performance wise with default than bespoke models. Better hardware mitigates this to some extent now but even so, with full slider settings you could bring a cutting edge rig to its knees now if you went all out.

3. The logic required to actually make their custom autogen look as it did took Earth Simulations years to develop, mostly through trial and error. I've since looked at it as a personal project and the theory is simple, but the execution is hard!

4. Availability of datasets to make things happen - especially for commercial use? (Accepted this is not something LM can impact on much, if at all!)

5. ROI is a big one, take Orbx TE for example. They essentially determined what the price point was for their product and worked in a level of detail that fitted it. They could have included location specific vegetation for example, but didn't. I sent them datasets for tree species locations across the entire UK. It's theoretically possible to create vegetation groups to make sure the correct species of autogen tree are used to a 1km square level of detail. I'm certain it wasn't used because the ROI would just not have been commercially acceptable.

ROI is where LM (respectfully) lose the plot. Because it's possible to do something in P3D, doesn't mean it will happen. I get the impression LM take the somewhat myopic view that as soon as they've facilitated the use of a specific feature in some way, that's it, job done. Not enough thought seems to be given to HOW that feature can be leveraged and the time and difficulty inherent in it.

Autogen is an example, so are speedtrees, and sloped runways. Can we imagine what P3D would (or might in the future) look like if Devs had been much more easily (read cheaply) able to use much better looking buildings, trees and terrain? I'd hazard a guess MSFS wouldn't have split the market to the degree it has done.

[kevinfirth]

18 hours ago, Mace said:
Super important.

I don't know if things have changed, but when I designed scenery, it was extremely time-consuming to make a change in an object, then have to re-load the sim just to verify the appearance of the change. I spent more time watching the loading/progress bar than I did actually designing.

Agreed, LM incorporated the clear shaders function while the sim was running, would it be too hard to have a similar function that cleared and reloaded model data while the sim was still running?

[kevinfirth]

Unfortunate as it is to say, I think MSFS points the way in some aspects. I dont want or expect LM to make P3D MSFSmk2, but it's just fact to say that with their strategic approach to how development can be accomplished for their game, Asobo have conmpletely outflanked LM. I would strongly recommend that LM incorporate some professional gaming expertise in their beta group...I cannot think of any conceivable situation where that wouldn't perhaps be helpful, even if other factors mean they choose not to follow any such input.

[kevinfirth]

So in order of preference, this would be my wishlist please...
1. Ability to make amends to scenery in sim in real time. Mesh and model changes.
2. Ability for ground polys to not have to be flat and to automatically adjust with terrain changes.
3. Truesky improvements
- red colours at dawn/dusk are overblown
- clouds appear in a grid pattern
- clouds still look too low res and pixelated at times
- there is a strong banding effect like an inversion layer that appears almost everywhere in OC conditions
- clouds do not impact on ground lighting conditions well enough
- Weather injections mean you sometimes get super fast streams of clouds
- Still get artifacting around objects and terrain
- Precipitation perhaps needs making more visible
- Ability for devs to accurate extract weather conditions - currently I see that Trueglass effects in aircraft often don't match the visible weather conditions when EA is on
4. Refactored nav database, with Navigraph updateability
5. Expanded autogen capability - more complex autogen models, improved support for speedtrees without devs being crippled by exorbitant licencing charges which mean accomplishable projects are uneconomical to deliver

LM have done a superlative job with performance increases, especially with v5 and the transition to DX12. None of the above should take away from that at all. I'd hope that it would be possible to work with NVidia and to exploit some of their new and emerging technologies as well.

[MPO910]

So in order of preference, this would be my wishlist please...
1. Ability to make amends to scenery in sim in real time. Mesh and model changes.
2. Ability for ground polys to not have to be flat and to automatically adjust with terrain changes.
3. Truesky improvements
- red colours at dawn/dusk are overblown
- clouds appear in a grid pattern
- clouds still look too low res and pixelated at times
- there is a strong banding effect like an inversion layer that appears almost everywhere in OC conditions
- clouds do not impact on ground lighting conditions well enough
- Weather injections mean you sometimes get super fast streams of clouds
- Still get artifacting around objects and terrain
- Precipitation perhaps needs making more visible
- Ability for devs to accurate extract weather conditions - currently I see that Trueglass effects in aircraft often don't match the visible weather conditions when EA is on
4. Refactored nav database, with Navigraph updateability
5. Expanded autogen capability - more complex autogen models, improved support for speedtrees without devs being crippled by exorbitant licencing charges which mean accomplishable projects are uneconomical to deliver

LM have done a superlative job with performance increases, especially with v5 and the transition to DX12. None of the above should take away from that at all. I'd hope that it would be possible to work with NVidia and to exploit some of their new and emerging technologies as well.

I do 100% Agree on this! Especially on the TrueSky parts and 4th and 5th also!

Marcus

[Saldo]

So in order of preference, this would be my wishlist please...
1. Ability to make amends to scenery in sim in real time. Mesh and model changes.
2. Ability for ground polys to not have to be flat and to automatically adjust with terrain changes.
3. Truesky improvements
- red colours at dawn/dusk are overblown
- clouds appear in a grid pattern
- clouds still look too low res and pixelated at times
- there is a strong banding effect like an inversion layer that appears almost everywhere in OC conditions
- clouds do not impact on ground lighting conditions well enough
- Weather injections mean you sometimes get super fast streams of clouds
- Still get artifacting around objects and terrain
- Precipitation perhaps needs making more visible
- Ability for devs to accurate extract weather conditions - currently I see that Trueglass effects in aircraft often don't match the visible weather conditions when EA is on
4. Refactored nav database, with Navigraph updateability
5. Expanded autogen capability - more complex autogen models, improved support for speedtrees without devs being crippled by exorbitant licencing charges which mean accomplishable projects are uneconomical to deliver

LM have done a superlative job with performance increases, especially with v5 and the transition to DX12. None of the above should take away from that at all. I'd hope that it would be possible to work with NVidia and to exploit some of their new and emerging technologies as well.

This is the list that I totally agree on.
P3Dv5 has a solid and stable base and Addons work with it very well using Simconnect and FSUIPC.

AG density/range changes inflight depending on framerate/CPU load without the need for a reload.

Now it is time to start working on this list to take P3D onto the next level as competition is also working on improvements...

[diggusa]

Right Kevin, LM needs to let Navigraph update their data base so we have a seamless update process thru Navigraph FMS manager as an addon since FSaero is supposedly quitting!!! Come on LM do something great! Would stop of CTD crap.

[kevinfirth]

Here's another possible improvement, unless it's been changed already?
Way back in the mists of time (2016 actually), Arno Gerretson discovered a wierd issue with extrusion bridges. See his blog:
https://www.scenerydesign.org/2016/07/u ... n-bridges/
Would be nice to be able to draw extrusion bridges with correct elevations, instead of having to hack them!

[PreparPK]

So in order of preference, this would be my wishlist please...
1. Ability to make amends to scenery in sim in real time. Mesh and model changes.
2. Ability for ground polys to not have to be flat and to automatically adjust with terrain changes.
3. Truesky improvements
- red colours at dawn/dusk are overblown
- clouds appear in a grid pattern
- clouds still look too low res and pixelated at times
- there is a strong banding effect like an inversion layer that appears almost everywhere in OC conditions
- clouds do not impact on ground lighting conditions well enough
- Weather injections mean you sometimes get super fast streams of clouds
- Still get artifacting around objects and terrain
- Precipitation perhaps needs making more visible
- Ability for devs to accurate extract weather conditions - currently I see that Trueglass effects in aircraft often don't match the visible weather conditions when EA is on
4. Refactored nav database, with Navigraph updateability
5. Expanded autogen capability - more complex autogen models, improved support for speedtrees without devs being crippled by exorbitant licencing charges which mean accomplishable projects are uneconomical to deliver

LM have done a superlative job with performance increases, especially with v5 and the transition to DX12. None of the above should take away from that at all. I'd hope that it would be possible to work with NVidia and to exploit some of their new and emerging technologies as well.

great, that would be my list too.
Any confirmation of LM when or if P3D 5.2 will arrive?