this MIGHT be a result of the engine scalability settings. in the top toolbar, click settings > engine scalability settings > and set everything to epic
In my level, I have a point light that was both SOURCE radius and length, larger than ZERO. It is set to movable. As you can see in the first image, it illuminates the ceiling, but when I pull the camera back, it seems to reset the size of both radius and length to ZERO, as shown in the second image.
Any ideas what is happening and how to fix it? Maybe it is a problem with 4.9 and fixed in 4.10 or 4.11
Everything is set to EPIC, and I am running a GTX 980Ti
The above problem is one of many lighting issues. They may or may not be related. See video: - YouTube
All the issues except the point light problem seem to have been resolved. I can’t confirm it without more time with everything, however, they do not occur now that I don’t run any other directX applications while UE is open.
I have figure out a bit more that is interesting regarding the point lights. If I cast shadows, then the light works as set up in the details tab, but they are being used as a fill light, and I don’t want shadows enabled. This has either got to be a bug or there is an engine setting to force the light to use it’s settings when the camera if further away.
Please check out the new video: - YouTube
Hey ,
Just curious, but have you attempted this in a different engine version as you mentioned in your original post?
Would you happen to be able to reproduce this in a blank project and provide me with some steps to get it to occur on my end?
Thanks,
Well, god news and potential bad news.
First the important stuff:
After a variety of trying different blank projects and in different engine versions, I was able to discover the offending actor, which is a BP Actor class which represents a tealight candle. in this BP is a particle system that emits light. When i disable the light in the Particle System, the Point Light in the level acts as it should.
Then I looked into the properties is of the particle system and discovered that is I enabled “Use Fixed Relative Bounds” then the Point Light would behave properly, but the Particle System light go not work. I tried changing the bounds box, but nothing seemed to reverse this.
Note that the point light has dimensions but acts as if it doesn’t:
With Particle System Fixed Bounding box enabled, the Point Light works, but the PS light does not:
In another test, with the PS light functioning and the Point Light not functioning properly, if I drag the tealight BP out of camera frame, then the Point Light returns to the proper function.
So the question now is how do I get both the Point Light in the level to function at the same time as the Particle System Light?
Other notes during this process not relative to the posted issue:
For some reason, when I had originally migrated this project to 4.10, I could not build lighting because the Swarm Agent would fail to launch. Not sure why, but I thought it had to do with Substance Plug-In. While trying to resolve the above issue, I started to slowly rebuild the level instead of purely migrating. I would build lighting along the way. Things were working, then I opened up the “migrated” level and it built lighting. I’m only mentioning this because I often will compile my own preview builds with Substance, and there has been a swarm agent problem before.
Cheers,
Sterling
Hey Sterling,
You are on the right track by modifying the bounds of the particle system. You mention you tried editing the values, but the numbers you entered might not be enough.
Try checking the ‘Use Fixed Relative Bounds’ again and changing the values to more extreme numbers. I would do this within the particle system editor (Cascade) instead of the blueprint details panel, just in case there is an override somewhere.
Below is an example of how I modified a particle system so it did not cull when the bounds were too small or set up incorrectly.
Fixed Relative Bounds
The areas in orange are the placed you want to look. The black area next to the emitter that is highlighted is where you need to click to change the bounds. Notice how my values go in the positive and negative direction to create a large bounding box while making sure the emitter is still in the center.
I would play around with these values until you get the result you are looking for, and also take a look at the occlusion settings to make sure those are being set somewhere as well.
For the other question, would you mind creating a new answerhub post for the Swarm Agent issue?
Thank you,
Hi Andrew,
Thanks for your continued help. I had tried different sized bounding boxes, but I could never get the light module working once Use Fixed Bounding Box was enabled. After I saw your last post, I tried some pretty extreme sizes just to be sure. The PS is just a simple flame for a tealight.
See my PS Setup. Don’t be confused by the Bounds and Occlusion settings in the image. I’ve tried every combination imaginable. No matter what, as long as Fixed Relative Bounds is enabled, the Light Module seems to not have any affect:
I also played with the BP Class bounds scale, but nothing I changed in any of the BP’s component properties had any affect.
I did notice in your example above, you do not have a Light Module in your emitter. Could you try adding that and make sure it works with this movable Point Light problem?
I’m simply mystified. I do have an alternative solution, but for the sake of the Engine, I’d like to know if this is actually a bug or my inability to work out the right bounding box solution.
Regards,
Sterling
Thank you for clarifying your own set up as it has helped me get some results on my end, however I am getting some irregularities with the movable point light at specific distances and not seeing the culling of the particle effect using ‘Fixed Relative Bounding Box’.
Instead of me trying to reproduce this in a project of my own, and the potential of not being able to get the issue to occur. Would you be willing to share your project with me via a google drive link or dropboxx link? Even better, if you are able to reproduce this issue in a new blank project with some simple steps, you can provide me with that instead.
All you need to do is zip up the Content, Config, and .Uproject into a folder and share that with me via an external folder sharing application like google drive or dropbox as mentioned earlier.
Thank you,
I have created a simplified clean project on Google Drive. I’m going to PM the link to you. Keep in mind that the problem exits in 4.9 - 4.11. My current workaround was to disable the light module in the Particle System, and add a point light to the tealight BP class. With this, I can keep working, but perhaps you can find a solution with what you are getting access to.
Cheers,
Sterling
Thanks for the project. I was able to clearly see the issue on my end, and definitely think it is a bug. Before I go ahead and enter it into our internal system, I am going to loop in another team member who knows the particle editor and rendering a bit more in depth and might be able to shed some light on the subject…pun intended. I will return with my findings once I have either entered a bug or found an answer to the cause of the issue.
Cheers,
Excellent!
So we were definitely on the right track according to editing the bounds scale. In your test project, if you set the particle emitter to ‘Use Fixed Relative Bounds Scale’ and set the values to Min X = -100, Y = -100, Z = -100 and Max X = 100, Y = 100, Z = 100 the flickering will go away.
Let me know if you get the same results, and if you can translate this to your project it should resolve the issue.
Cheers,
I’m not getting the results you are, or you are confused about the situation when it stops flickering. As long as the PS is in the frame of the camera, the point light behaves as if it has no source radius or length. So, when the Use Fixed Relative Bounds is enabled, it does stop flickering, but it doesn’t respect the light’s parameters.
I have made a short video demonstrating this. You will notice that it starts off correctly, but when I pan down, the point light changes to something that it shouldn’t. This is also the case when I enable the Fixed bounds option in the PS.
-Sterling
Andrew,
Thanks for focusing on this problem. Bugs are never good, but at least we nailed it down and it will hopefully be fixed at a later date.
Cheers,
Sterling
I did some further digging and found a bug that was entered a long time ago dealing with this issue specifically that was backlogged. It is directly related to a particle system with a light module next to a point light in the scene with source length. When the light actor no longer casts shadows, the source length is discarded. I am going to re-open the ticket and add some interest to the bug report.
Since you have a workaround, I would go ahead and use that until the fix is in place. Thank you for your patience while we got this figured out and let me know if you have further questions.