Open sl-service-account opened 4 years ago
Dan Linden commented at 2019-08-28T18:27:20Z
Hi polysail, I don't see the same lighting on EEP that is in your screenshot. Would you repro this bug and paste the Help > About Second Life info into the Environment field of this Jira?
polysail commented at 2020-02-08T06:13:52Z, updated at 2020-02-08T06:14:10Z
Okay. So, after quite some time, I've tracked down some more info on this bug. It is NOT actually an EEP specific bug. It's a peculiar, ( and highly intermittent ) "state" that all of the viewers, EEP, Non-EEP and the current Release of Firestorm occasionally get "stuck in". I've done my absolute best to try all manner of things to try and get a reliable repro, however I have not been successful in getting a reliable trigger to throw the viewer into this defective state, it just seems to happen sometimes when logging in.( Great.. one of those bugs (sigh)... I know... )
The undesirable state makes projector lamps suddenly fail to cast shadows at certain camera distances while zooming in and out of the scene. That is to say, if the camera is 10 meters away from the avatar, projector lamps will render and interact with scene geometry correctly, they will cast shadows, and display their projected texture. However, as the camera is zoomed out, to say 15 meters sometimes, it just hits an ( seemingly arbitrary ) threshold where this behavior ceases, and the light goes clean through the geometry, flooding the area in full-strength zero-falloff value light. Zooming the camera back in ... sometimes... fixes the situation, but oftentimes not. This "state" persists, even if local projector lights seem to work, ones in he distance might not, etc etc, and the projected textures rarely seem to function correctly.
Restarting the viewer and logging in again usually restores the viewer to correct behavior, unless of course it doesn't...
I've gone through an entire area of our home sim while I was unwittingly in this "altered" viewer state, where the projector lamps had no falloff, and adjusted all my lights down to "sane" values, only to discover upon relog, ( which usually clears the bad state ) that I had now dimmed all my lights to be only barely visible once the falloff-calculations went back to functioning.
polysail commented at 2020-02-08T06:26:04Z
Rewrote this Jira to reflect the behavior exists across all current viewers, including, but not isolated to EEP.
Dan Linden commented at 2020-02-10T20:31:33Z, updated at 2020-02-10T20:32:17Z
The first video (https://gyazo.com/78229c86239f4e51730dea6db246fc48) is an EEP specific bug when switching from forward rendering to deferred rendering, SL-12005 "[EEP] Projector light brightness steps down as camera approaches it". This bug has been fixed internally and will ship in EEP RC 6.4.0.536347 or later.
I see the light turning on and off in the second video (https://gyazo.com/ac8c3aac5684facb94a71474e65b6b2c). Are you doing that or is it happening by itself? Which viewer is that?
polysail commented at 2020-02-11T00:41:30Z
I'm not actually 100% sure about the source of the flickering, as I didn't personally shoot that video, however, the video is taken on Firestorm Release. But the flickering is not the valuable part of the information. The important part is that you can see how the lighting cones have harsh edges, and do not interact with the building geometry at all. It's as if the geometry isn't even there.
I have personally seen it flickering between shadows being correctly cast from the projector, and not being cast from the projector. However, it is possible that the EEP fix will have addressed this? I will continue to update this if I can get this viewer 'state' again and can collect more info.
Dan Linden commented at 2020-02-12T00:55:03Z
I see shadows cast at the location of https://gyazo.com/ac8c3aac5684facb94a71474e65b6b2c with both the SL Release viewer and EEP viewer. Perhaps shadows were not enabled in that video. If you can find a repro in the SL Release viewer or EEP viewer we can look at it further.
Thank you! Dan
polysail commented at 2020-02-12T04:17:07Z
That's what I'm trying to tell you Dan, when the viewer goes into this "state" the shadows aren't rendered, even if they ARE enabled. That's the bug. Next time I get a viewer into this bizarre state I'll do my best to document it again. But I can promise you, the viewer in that video DOES have shadows enabled, they're just not working because of ... well that's the entire point of this JIRA.
What just happened?
This is not Actually an EEP Bug, but it does present itself on the EEP Viewer reliably:
https://gyazo.com/78229c86239f4e51730dea6db246fc48
You can see the complete lack of interactions with geometry here :
https://gyazo.com/ac8c3aac5684facb94a71474e65b6b2c
What were you doing when it happened?
Playing with projector lights. Logging in and out of the viewer seems to reset this "bad state" but if this state is triggered, by some means ( I have no idea what it is sorry ), it persists until logout.
What were you expecting to happen instead?
I was expecting a nice even diffuse glow that had a soft either linear or inverse square falloff, instead I got the appended screenshots.
Other information
Attachments
Links
Related
Original Jira Fields
| Field | Value | | ------------- | ------------- | | Issue | BUG-227557 | | Summary | Odd lighting effects - Bad Viewer "State" causes Projector Lighting Cones occasionally have no falloff effect. | | Type | Bug | | Priority | Unset | | Status | Needs More Info | | Resolution | Unresolved | | Labels | whirly-eep | | Created at | 2019-08-28T15:41:53Z | | Updated at | 2020-06-15T15:23:24Z | ``` { 'Build Id': 'unset', 'Business Unit': ['Platform'], 'Date of First Response': '2019-08-28T13:27:20.361-0500', "Is there anything you'd like to add?": 'EEP is not ready for release.', 'ReOpened Count': 0.0, 'Severity': 'Unset', 'System': 'SL Viewer', 'Target Viewer Version': 'viewer-development', 'What just happened?': 'Everyone is talking about "EEP is getting ready to be released, we think we\'ve solved all the graphics issues."\r\n\r\nYet none of these lighting issues in Part 1 & 2 seem to have been addressed, and they are literally game-breakingly bad.', 'What were you doing when it happened?': 'Literally just spinning my camera around and looking at things. Anything with point lights really.', 'What were you expecting to happen instead?': 'I was expecting a nice even diffuse glow that had a soft either linear or inverse square falloff, instead I got the appended screenshots.', } ```