shadowing glitch

[disc]

I've seeing a shadowing problem since the dynamic shadows are implemented. I can't really explain it with words, so i've attached a video instead.

[James Ross (twpol)]

Does it still happen with an experimental build of V1371 or later?

[disc (disc86543)]

Yes, this video made with the latest, V1388

[Wayne Campbell (wayneinbc)]

Just as a possible clue here ... I have noticed this sort of behaviour when a scene contains objects made up of 'large' mesh triangles. If the mesh triangles in a scenery object exceed about 200 meters in length, then the shadows are effected as shown in the video.

[cjakeman (cjakeman)]

Hi Kerko,

Very convincing video. Can you tell us where to download this route and where the station building is? Is the route available for free?

Thanks,

Chris

[disc (disc86543)]

It's free, Xtracks needed for it:
https://docs.google.com/uc?id=0BwJFITblw5mXbE96QWFlbWlLNE0&export=download

This bug at Budapest-Nyugati or Nyugati
Probably because the whole station building is modeled in one piece, so it's very big.

[cjakeman (cjakeman)]

Hi Kerko,

Thanks for packaging up the XTracks which the route needs. But I still don't have the name of the route or from where to download it. Am I missing something?

Best wishes,

Chris

[Peter Gulyas (pzgulyas)]

(Just a note from an outsider: The zip in comment #6 contains the route, not the XTracks. It's name is "Alföld", it is the largest free route from Hungary. An alternative download link in case google docs isn't preferred: http://www.simtreni.com/download/index.php?subcat=7 . But in this case it must be downloaded in more parts.)

[Peter Gulyas (pzgulyas)]

I can reproduce this bug only with ShadowAllShapes turned on in registry.

[Peter Gulyas (pzgulyas)]

I could reproduce this bug in r1477 on standard route with standard rolling stock as well.

1. Create new registry key of type DWORD and set its value to 1: HKEY_CURRENT_USER\Software\OpenRails\ORTS\ShadowAllShapes

2. Start OR, and select route Northeast Corridor, activity Summer Express

3. When game started, go to camera 3

4. Zoom out with PageDown key till it is possible.

Now the weird shadow gets visible. You can also try rotating around to see it moving.

[disc (disc86543)]

Actually this happens every time at everywhere, even without shadowallshapes. It's more noticeable at morning and evenings when the shadows are longer.

[James Ross (twpol)]

I cannot reproduce using the steps in command #10 but I have seen this occasionally myself so I'll have a look.

[James Ross (twpol)]

Fixed in V1508.

[Wayne Campbell (wayneinbc)]

May be related to Bug 1160216

[disc (disc86543)]

now a similar bug appeared again. I use V 1517

See the video

[disc (disc86543)]

Actually, this is worse than before :(

[disc (disc86543)]

Since the distant mountain update, i'm seeing long flickering shadow stripes too.

[James Ross (twpol)]

Did you experience better/fixed shadowing after the fix was committed and before the Distant Mountains (DM) were added? If you're ONLY seeing the issue with DM enabled in the options, please file a new bug on that. If you're seeing the same issue with and without DM enabled, I'll investigate more here. The flickering stripes should also go in to the new bug, as it's dependent on that experimental feature. Thanks!

[disc (disc86543)]

I've made the video one day before the first release of DM, so which is in the video, that isn't in connection with DM. After the fix, the angle dependent disappearing shadows gone, but few releases later(before the DM) i saw those that are in the video.

 But the stripes are probably caused by DM.

[disc (disc86543)]

I've just reviewed the old video and compared with the current problem and as i see, it's the same problem.

[disc (disc86543)]

another video about the "camera following" shadow and the "growing tree" shadows, maybe it can help to find out the problem.

[James Ross (twpol)]

Sorry about the delay with this bug, but I don't think the video in comment #21 is actually caused by the change here. There are other issues that are caused by it, however (some of the issues in the video in comment #15, for example).

[disc (disc86543)]

these shadow are also visible on default tokyo-hakone route, selecting shinjuku - machida path, at spring, time set to 12:00. It seems the problem occurs when of the shadows of big buildings, with big faces. But doesn't with smaller houses.

[James Ross (twpol)]

The original issue here has been fixed in V1597. The shadowing issue on water seen in the 3rd video is a separate issue and should be filed as a new bug.

[disc (disc86543)]

It's "fixed" indeed... because now the shadows are now completely disappeared :D

[disc (disc86543)]

oops no, arent disappeared, but depending on time of day, the shadows can be very faint, barely visible. This happens when the time is closing to noon.

[James Ross (twpol)]

They look fine to me, but if they have got fainter it's directly as a result of the required change to fix the platform being in shadow in this bug. You can't have everything, unfortunately.

[disc (disc86543)]

It seems that the shadows have fixed color shouldn't be work as substractive material does? And why the color of shadows are depending on time? Actually in reality, the highest contrast between lit and shadowed surfaces are at noon. But now in OR the opposite happens.

[James Ross (twpol)]

You have absolutely no idea how shadows are implemented. They are very sensitive to any changes and it is impossible to satisfy everyone's desires for them at once. Maybe I'll have another free weekend to dedicate to them (I've spent *many days* working on this issue) some time but until then, this bug is fixed.

[James Ross (twpol)]

BTW, you can mostly 'fix' the lighter shadows by turning off ShadowMapBlur.

[James Ross (twpol)]

Okay, turned out to be easier to fix the faint shadows, so they should be sorted in X.1600.