Build A Better Skydome - Terragen 2 ?
#181
Posted 22 January 2023 - 04:13 PM
edit -- I searched and found it--- scottb613 uploaded some water textures to the library. "ORTS Water Textures"
Path is >>>
#182
Posted 22 January 2023 - 04:54 PM
Do you think you could look into adding 2d Lightning effects for ORTS using the information found here?
https://gamedevelopm...s--gamedev-2681
This would greatly add to the desert rail routes, since the lightning can be very intense in such a hot try environment.
Thanks
Robert
#183
Posted 22 January 2023 - 06:52 PM
1. I am very glad we're talking about ways to implement a more realistic sky!
2. On the subject of thunder and lightning - I already have a thunderstorm SMS available in navs_env.zip - at the moment, you have to substitute them for the rain SMS files, but I consider them up for grabs (public domain) if the community wants to use those environment sounds as a basis for this. At the moment, there are two thunder streams, one with distant thunder effects and one with close thunder effects, both set to go off at random. There are wind and rain effects in other streams. I had the idea that loading up stormy weather could load up storm_in.sms and storm_ex.sms if present, and the rain ones if not.
3. On the subject of 3D clouds - yes, it's an art-intensive process, but I'd be more than happy to build some shapes and I'm sure others would be as well. What if we use an extensible system whereby all cloud shapes in a folder are loaded at random without respect to an individual filename? Then the number of 3D clouds can grow organically as more are built. You could start with one or two and then the amount of variety increases as the community builds more. Scott, since you've been rendering skydomes, would you consider the idea of rendering textures for individual cloud sprites? Then I could use the sample FS2004 clouds as a guide for building at least a couple cloud shapes.
If we did sprite clouds, what would the shapes/hierarchy look like? Maybe the top node can be a dummy, and each sprite can be a simple 2D plane with the name "spritexxx" where "xxx" is a number? Maybe the sim could recognize that node name as a sprite?
I also have some general observations:
1. In a rail simulator, a lot of the problems with 3D clouds that exist in a flight simulator aren't really a major consideration because you're not flying through them. I am not sure we'd actually need sprite-based clouds at all. Let's say each cloud shape consisted of a flat base that covers the whole cloud (kind of like the current cloud dome), with fixed geometry constructed like the FS2004 sprites fleshing out the space above it? When the cloud is directly overhead, you'd just see the bottom, and as it passes further away, you'd start to see the top-level geometry. Considering how OR batches drawcalls, I suspect this might be more efficient as multiple clouds could effectively become a single draw as opposed to a bazillion sprites. Then 2D distant clouds like the Railworks example above can fill in the clouds in the distance.
2. I think we only really need 3 cloud layers - we can have a flat layer for stratus clouds that basically functions like the current cloud dome, then a layer of 3D cumulus clouds, then a top cloud dome with cirrus clouds.
If we wanted to experiment with maybe some basic spriteless 3D cloud shapes - I could probably build a couple in not too much time using the FS2004 sample cloud artwork just as a proof-of-concept. A whole bunch of sample cumulus clouds and the textures came with the FS2004 gamepack:
Just for the fun of it, I built a quick-and-dirty cumulus cloud using my "fixed geometry" idea - it's 6 planes:
Viewed from below here. Not the greatest cloud shape by any means, but it was made in about 3 minutes - if the cloud bases obscured the upper-level geometry enough, it'd probably look fine from the ground. As a size comparison, I went ahead and dropped my 70-tonner in the scene:
Food for thought - if we can get a half-decent set of textures going, I think we could build up some pseudo-3D clouds without too much time and effort.
Same cloud viewed from the side - the tiny black dot at the bottom center of the image is the 70-tonner. The cloud base measures about 15,000 feet across - given how big the cloud shapes are, we can cover a lot of real estate with a few simple models, a ring of distant clouds like in the Railworks example above, and a flat layer of cumulus "film" below. Add in a stratus layer and a cirrus layer - and you've got a dynamic sky without using a whole lot of resources like sprites would!
#184
Posted 22 January 2023 - 09:49 PM
Log file attached.
What exactly is the OR code now looking at from the ENV folder that it was not looking at before?
Attached File(s)
-
OpenRailsLog.txt (12.92K)
Number of downloads: 129
#185
Posted 23 January 2023 - 09:50 AM
#186
Posted 24 January 2023 - 02:37 PM
railguy, on 22 January 2023 - 09:07 AM, said:
The sky and clouds will be chosen based on weather parameters; we may well be adding more presets than the simple clear/rain/snow options and any new options will need to be supported elsewhere (activity, weather file)
Genma Saotome, on 22 January 2023 - 11:58 AM, said:
There might be scope for this later on, after I've analysed the current sky colour issues, but I am not expecting to implement a full sky/cloud dome replacement from content like the Railworks example (that's what the MSTS environments is for)
Instead, content may get the option of overriding the colour choice for e.g. clear sky, sunset, etc., but I want to emphasise that the detail has not been figured out yet
Jovet, on 22 January 2023 - 02:55 PM, said:
Yes, there is hope. I've love for it to happen but currently I'm looking up to the sky :)
SP 0-6-0, on 22 January 2023 - 04:54 PM, said:
https://gamedevelopm...s--gamedev-2681
This would greatly add to the desert rail routes, since the lightning can be very intense in such a hot try environment.
This is really interesting, thank's for sharing! :sign_thanks:
I did come across the Köppen climate classification and similar things recently, which would potentially give us two things:
- An inhabited Earth-wide classification of the climate
- A taxonomy for routes to specify a climate
We already have some basic code to check for latitude/longitude inside a "desert zone" to change the built-in behaviour of the weather, so it might be good to expand that based on a climate classification system
ErickC, on 22 January 2023 - 06:52 PM, said:
Thanks, it is really useful to have options like this!
ErickC, on 22 January 2023 - 06:52 PM, said:
It seems reasonably likely at this point that we'll be wanting to do this even for textured skies; e.g. some kind of folder/naming convention for picking individual cloud textures (or parts of cloud textures?) depending on the layer and weather conditions
Definitely needs more investigation though into what we need exactly
ErickC, on 22 January 2023 - 06:52 PM, said:
That's an interesting point, we might only need some bits to be 3D (or have the appearance of 3D)
ErickC, on 22 January 2023 - 06:52 PM, said:
These do seem pretty good for such a quick setup!
I'm going to try out some flat options first but we may nevertheless end up with 3D clouds
PerryPlatypus, on 22 January 2023 - 09:49 PM, said:
Log file attached.
What exactly is the OR code now looking at from the ENV folder that it was not looking at before?
James Ross, on 17 January 2023 - 01:22 PM, said:
It looks like your issue is different from the other thread, so I will prepare an additional fix for the MRL Mullan Pass crash (and any other route with similar data)
#187
Posted 26 January 2023 - 02:55 PM
PerryPlatypus, on 22 January 2023 - 09:49 PM, said:
I don't have the route, so I can't be 100% sure, but I believe this crash is fixed in Unstable Version U2023.01.26-2251
#188
Posted 29 January 2023 - 02:04 PM
Glad you found the link about the 2d lightning interesting. I will be looking forward to lightning in the future.
A couple other observations, A long time ago Robert Murphey had started to look into supporting more of the original MSTS .Env files. The MSTS compatible mode in ORTS still produces a weird cartoonish looking sky. What is the current status of those improvements to support more of the MSTS .env files?
Second,
Regarding rain and environmental effects. I'd like to see rain hitting the windows and the wipers removing the rain drops. I'd also like to see rain flowing down the windows in passenger view modes.
Maybe some of the needed coding can be adapted from the info here? Windscreen Rain
Rain 2
Raindrop effect pass view.
The rain on the passenger car windows is at the 2:06 mark.
Robert
#189
Posted 01 February 2023 - 11:51 AM
James Ross, on 26 January 2023 - 02:55 PM, said:
Sorry for the late reply. Yes, I have confirmed no crashes on Mullan now, thanks!
I did notice that the super-hyper bright Headlights and Sphere of Light on certain locomotives even in the middle of the day, turning the ground to a blinding white color, is back (not sure when that popped up, but it's not present in the latest Stable version). I am assuming that's nothing you've done related to the environment work per se, but since it plays into the whole lighting engine I thought I'd mention it.
#190
Posted 02 February 2023 - 02:38 PM
SP 0-6-0, on 29 January 2023 - 02:04 PM, said:
I am not currently planning any work on the MSTS environment support, but others are welcome to have a go - and maybe the updates I am making to the OR environment code will help
SP 0-6-0, on 29 January 2023 - 02:04 PM, said:
This is definitely a desirable feature, although I don't know if it would be possible to implement for existing models - it may need specific data adding to properly identify the window surfaces and how the wiper clears it
Something I'll keep in mind for the future
PerryPlatypus, on 01 February 2023 - 11:51 AM, said:
Great, thanks for checking!
PerryPlatypus, on 01 February 2023 - 11:51 AM, said:
I don't believe any changes I've made here would do that, but good to highlight it, and you're right that it'll be included in the brightness / colours review after I am done with the wind