Dynamic backgrounds and fog don't seem to work as from 1.11.2.7

[ginga81]

Description

Select a route that uses dynamic backgrounds and lighting, and load your preferred vehicles.

Reproduction

If the issue occurs in multiple routes/ trains, please provide one or two samples.
In order to reproduce the issue and debug it, it's helpful to have the following:

Tohogu-Shinkansen_fromSendai-v0.06.02(EN)
Touhogu-Shinkansen-Hayabusa27_fromSendai_v0.07a(EN)(summer,summer-rain,winter)

Route

http://midnightexpginga81.web.fc2.com/touhoku_shinkansen/route_data/Touhogu-Shinkansen-Routedata-fromSendai_v0.06.02.zip

Object and Sound
https://drive.usercontent.google.com/download?id=15PudZowh0zbFmTX92DO4r4uT944DDYIZ&export=download&authuser=0

Train

any

Logs

This issue do not crash.
So no log.

Related information

My PC's spec:
OS:Ubuntu 24.04 LTS 64bit
CPU:AMD® Ryzen 7 2700x eight-core processor × 16
Memory:31.3GB
Graphic:Radeon RX 9060 XT 16GB model
AMD Radeon Graphics (radeonsi, gfx1200, LLVM 20.1.2, DRM 3.61, 6.14.0-37-generic) (0x7590)

[ginga81]

To 1.11.2.6, it works.
Dynamic lighting is perhaps works.
When I run, I felt.
But only felt, perhaps also not work...

[leezer3]

It's the fix from #1195

Before this, the background distance was accidentally fixed at 600m for dynamic + objects.
It's not actually a bug- You've got fog obscuring the background.

(Drop the viewing distance to 600m and your background will re-appear)

[ginga81]

This weekend we will be releasing an updated timetable for the evening to night time with the same dynamic lighting and fog settings. Should I announce everyone to set their drawing distance within 600m when driving the timetable with dynamic lighting?

[ginga81]

If it can be support a drawing distance longer than 600m, I will stop the release new version until fix.

[leezer3]

So, this has opened an interesting issue.

Looking further into this, in the documentation, it states the following:

For the fog calculations, the background image is assumed to be at 600 meters distance from the camera, regardless of the actual viewing distance.

However, this isn't the case. We actually check against the BackgroundImageDistance (viewing distance):
https://github.com/leezer3/OpenBVE/blob/master/source/OpenBVE/Graphics/NewRenderer.cs#L197

I've confirmed this is the case going back to Michelle's last version, and it would appear to be a very old bug.

According to the documentation (again...):

objects after the ending distance appear in the fog color

This is why your background is appearing as the fog color.
I'm very reluctant to change the current behaviour- It seems considerably more logical, and isn't tying us to what I presume is a BVE2 / BVE4 issue when using differing fog distances.

---

I think the 'best' way to deal with this is likely to be to allow the setting of the fog blending distance of the background, but this would be via the XML.
Might well need an errata note too, and possibly a hacks entry if we find any legacy files which don't take kindly to this fog behaviour.

[ginga81]

I see.
So the 'bug' was in versions up to 1.11.2.6.
And 'fix' from 1.11.2.7.
When I thought about it from a document's perspective, I understand that it was correct to say that it was invisible.
If that's the case, I think it's contradictory that the 600m limit is only applied when a DynamicBackground is used.
I think it should be white when it's over 600m, regardless of whether there is a DynamicBackground or not.

[leezer3]

Not quite- The 'bug' is still present, it's just that the dynamic backgrounds behaviour 'matched' the documentation, but that was by accident :(

The build from today leaves the 'bug' intact, but adds the following property to dynamic backgrounds:

<FogDistance>Distance</FogDistance>

Distance should be set to the relative fog blending distance you wish to use for the background.
Setting this to 600 will reproduce the effect you had previously.

---

Other than this, I'm proposing to delete the 600m fog statement from the documentation as 'incorrect' and adding it to the errata instead, as something that never worked properly, and that 'fixing' it would actually seem to make things worse.

[ginga81]

If that's the case, I'm very relieved.
Actually, I suddenly got attention after the Netflix movie about Shinkansen last year, and every time I write about my work on Twitter, I get 10 times the response than before everytime.
Among them are many people who are hearing about OpenBVE for the first time.
10 times more means 90% of people are probably new to it.
In that situation, it would have been very difficult to people to 'Please use the old version for the evening diagram'.
So I'm very grateful.
I'll re-create the data using fogdistance, verify it, and then publish the data after the official version is released.
That way, I thought even first-time players would be able to play easily, just by asking them to install the latest OpenBVE.

[ginga81]

Thank you very much!
After adding the XML, it now displays in the latest build.
However, since there are many first-time users, I will release it after the official version is released.

[ginga81]

The new parameter have been incorporated and are ready to be published.
Once they are reflected in the major version, I will be published at the same time.
I think that asking 'you to play with the new daily build' will probably be difficult for 90% of new users, so...

[leezer3]

OK, I've pushed 1.12.0.0 into pre-release.
I'm still working on a bunch of not yet properly implemented MSTS stuff, but that can wait for a newer build.

Will put that up on the main site probably tomorrow. (Really need to do a decent video for that...)