The Viper is a tricky ship for shadows since the cockpit is so small and the shadows so close. There isn't a whole lot that can be done with EVSM shadows enabled and disabling them can result in other artifacts or even client crashes. That said, if you want to try it, you can change <FilterQuality>3</FilterQuality> to "2" for "Profile_General".
The Viper is a tricky ship for shadows since the cockpit is so small and the shadows so close. There isn't a whole lot that can be done with EVSM shadows enabled and disabling them can result in other artifacts or even client crashes. That said, if you want to try it, you can change <FilterQuality>3</FilterQuality> to "2" for "Profile_General".
These shadow artifacts (captured using my v20 tables) are an issue that has been annoying me ever since EVSMs were introduced with Odyssey (r-click and open in new tab then fullscreen the image with F11 for more clarity):
I'm speaking of the darkened square that corresponds to Frustum0 as well the excessive haloing that happens within range of this frustum when directional shadows overlap. This mostly apparent at very shallow angles, but is visible often enough for me to find it quite distracting. In 3.8 and prior versions of the game, or Odyssey itself with EVSMs disabled (which is not practical for reasons listed earlier in this thread) these issues could be fixed with careful management of depth biases, but Odyssey's EVSMs ignore these.
Removing or pushing out the closest frustum reduces the obvious transition, but the net effect is worse. The whole first Frustum is still overly dark and produces serious haloing. This is the same area, using the same tables, but with the 5m frustum removed:
For reasons still unknown to me, this issue is only present with the first frustum, no matter what it's end distance. So, I had the idea of using Frustum0 as a dummy cascade with an end distance of zero. Results of this are pretty good:
That dummy cascade eliminates the darkened square completely and significantly reduces (though doesn't eliminate) situations where prominent halos are visible on shadow edges. There is a loss of close in shadow detail/crispness with having the first visible cascade at 10m instead of 5m, but I think it's a worthwhile trade off.
Profile_PlanetSurface may have enough room for me to keep the dummy while maintaining a 5m Frustum1, but I'll have to test this later. I also need to evaluate if Profile_PlanetApproach and/or Profile_StationInterior can benefit from a dummy frustum. I do not think it's very useful or practical for the other profiles, given the challenges already present in maintaining quality at longer distances, but I'll probably do some experimenting anyway. No ETA on these changes.
In the meantime, I'm going to release an update with my preliminary changes to Profile_PlanetSurface as v21.
So, I figured out that, with the right depth biases, I can actually push out that 'dummy' Frustum0 to 3.5 meters, which sharpens very near shadows considerably and allows Frustum1 to be extended slightly. This is important because Frustum1 appears to be maximum distance at which the head/helmet shadow will render and the model shadows look pretty weird when the head disappears too soon.
It had nothing to do with night vision (it was the same without NV) and I think it was shadow related because it went away instantly after I had set both directional and spot shadow quality to low before the next fight (idk which one was that cured it, I just had to do something quickly because it was unplayable). I was using the simple generic Ultra settings before U18 without any issues*, it does not seem to work now.
(* I mean, I can see why you are trying to get Ultra+ level shadows, by 'no issues' I just mean the simple ingame Ultra was enough for me in the rings and it certainly never used to cause annoying flickering before)
It had nothing to do with night vision (it was the same without NV) and I think it was shadow related because it went away instantly after I had set both directional and spot shadow quality to low before the next fight (idk which one was that cured it, I just had to do something quickly because it was unplayable). I was using the simple generic Ultra settings before U18 without any issues*, it does not seem to work now.
(* I mean, I can see why you are trying to get Ultra+ level shadows, by 'no issues' I just mean the simple ingame Ultra was enough for me in the rings and it certainly never used to cause annoying flickering before)
This looks like the flickering shadow bug that was originally introduced in U12, then fixed in U14 (and apparently reintroduced with U18). It's got nothing to do with the shadow cascades themselves and there doesn't seem to be anything that can be done to fix it on our end, short of using shadow tables that do not cast directional shadows past a certain distance.
Still much testing and tweaking I need to do, but v23 is enough of an improvement that I'll post it.
I've integrated the 3.5m cascade into every profile where it was practical to do so and adjusted the end distances of the other frustums to compensate. I've also reworked many depth biases, primarily increasing them to what I had back in my pre-Odyssey profiles, to mitigate some persistent shadow acne that was still visible on ships.
I still need to see if they ever fixed the issue where Profile_StationInterior was used in place of Profile_PlanetApproach after docking request had been granted at surface outposts/settlements. Profile_AsteroidField could also use some more tuning.
Did not notice it yesterday and the day before in the ring. I'm not sure though if I set the shadow quality settings back to Ultra, gonna check it later. (Even right after 18.00 it wasn't happening all the time, on the first occasion it started flickering after about the first 3 minutes of the fight, while in the next instance it was present from the get go. Lowering shadow quality stopped the flickering in the ring, but not in stations. It was weird.)
I saw some flickering also on space stations for a few moments when approaching the space station. It disappeared shortly after. This was new to me and has not been seen before. And I play since December 2014. I do not use the modification files.
Did not notice it yesterday and the day before in the ring. I'm not sure though if I set the shadow quality settings back to Ultra, gonna check it later. (Even right after 18.00 it wasn't happening all the time, on the first occasion it started flickering after about the first 3 minutes of the fight, while in the next instance it was present from the get go. Lowering shadow quality stopped the flickering in the ring, but not in stations. It was weird.)
I didn't think there was anything that could be done without a patch, but when I was trying to push asteroid shadow distance past their full illumination distance, the flickering seemed to have stopped, for me. All the above does is push out the distance of the last cascade from 9km to 15km; it should be essentially free, performance wise.
Same approach might potentially work with the other profiles, but general seems to have a tighter distance cap, while all the others have special considerations that make it impractical to try to max out shadow distance.
I didn't think there was anything that could be done without a patch, but when I was trying to push asteroid shadow distance past their full illumination distance, the flickering seemed to have stopped, for me. All the above does is push out the distance of the last cascade from 9km to 15km; it should be essentially free, performance wise.
Same approach might potentially work with the other profiles, but general seems to have a tighter distance cap, while all the others have special considerations that make it impractical to try to max out shadow distance.
This is awesome, thank you.
Def gonna try it the next time I can see flickering rocks.
Edit: I've just checked it, both directional and spot shadow quality is on Ultra, which means it was the same way yesterday night when I participated in 3 or 4 wingfights and there were no flickering shadows. I would have 100% noticed it if it had been present, because it's extremely annoying when it's happening mid-fight. It's nearly unplayable level of bad and you cannot really do anything to fix it while you are under fire.
I've just checked it, both directional and spot shadow quality is on Ultra, which means it was the same way yesterday night when I participated in 3 or 4 wingfights and there were no flickering shadows. I would have 100% noticed it if it had been present, because it's extremely annoying when it's happening mid-fight. It's nearly unplayable level of bad and you cannot really do anything to fix it while you are under fire.
It's just the generic ingame Ultra setting.
I have not yet installed the fix you sent, because I thought I was going to need to see the flickering first otherwise there would be no way to tell what fixed it.
Still much testing and tweaking I need to do, but v23 is enough of an improvement that I'll post it.
I've integrated the 3.5m cascade into every profile where it was practical to do so and adjusted the end distances of the other frustums to compensate. I've also reworked many depth biases, primarily increasing them to what I had back in my pre-Odyssey profiles, to mitigate some persistent shadow acne that was still visible on ships.
I still need to see if they ever fixed the issue where Profile_StationInterior was used in place of Profile_PlanetApproach after docking request had been granted at surface outposts/settlements. Profile_AsteroidField could also use some more tuning.
U18 reintroduced a bug that was previously fixed in U14 that causes certain shadows to flash/flicker. It's difficult to separate reports of flickering shadows caused by this bug from ones caused by the shadow tables (stock or otherwise) themselves, especially without video examples.
ufile.io
