Volumetric Light Beam - Volumetric Lighting solution compatible with Mobile, VR and WebGL

Being able to create volume light shapes using a cookie would be AMAZING. I would love to see that, especially since I use cookies to shape so much of the light in my scenes already.

 

Thank your for your suggestion, it's duly noted

I would like to give you more details about how I plan to support this feature. Like you can imagine, this is not an easy one to implement, so I would like to be sure that it would satisfy your needs, and expose to you also the possible limitations.

First, I plan to support box/rectangle beam shapes, since it could fit in many common cases, and it's much easier to implement on my side

Then I plan to support 'custom beam shapes'. Instead of just a circular/conic beam (or a box/rectangle beam with the feature above), the user would be able to specify a custom shape by providing a cookie texture (so basically a black and white image). The plugin would then procedurally generate the proper mesh from the texture data. I think this is the most user-friendly way to do so, since this user would anyway need to provide a cookie texture to its spotlight in many cases. It's also less cumbersome than asking the user to provide a handmade custom mesh, or than to draw a vectorial shape directly in Unity via a custom editor.

One strength and unique benefit of using this asset are the performance. In order to allow this feature to work on mobile and VR platforms without any problem, this would have some limitations. As long as the shapes drawn in the cookie texture are contained inside 1 "blob", without "holes", it would be fine. For example, supporting this kind of shapes would be fine:


However, supporting shapes with "holes", or containing separated pieces, like a grid, would be much more complicated, because I would need to generate multiple shapes or to perform some raymarching algorithm inside the shader (which would kill the peformance):


Here are my questions:
- Would it be fine for you if the plugin would only support "simple" shapes (like the star or the fan) and not "complex" shapes?
- Also would you be happy with this kind of workflow? (specifying the custom shape via a black & white cookie texture).

I would also like to thank you for your review on the AssetStore
I am always happy to hear helpful critics and suggestions to help me improve the asset. May I ask you what kind of features you would like/need (besides 'custom beam shapes') to make the asset more "all-encompassing" compared to some others solutions?

Thank you!

 

This all makes sense to me! And as someone who's developing in VR specifically, performance is a big priority, so I appreciate the desire to keep the product as efficient as possible.

Starting with square and rectangles would be a huge help, and I think at least for now if people want holes or grid shapes they can create them by stacking multiple beams side-by-side, no? A cookie would certainly be very easy to work with, at least for me.

When I said "all-encompassing" I was mostly referring to the plugins that just make volumetric light a general effect applied evenly to an entire scene via lights or highlights, often simulating "scattering," such as https://github.com/SlightlyMad/VolumetricLights or https://www.assetstore.unity3d.com/en/#!/content/67665 ... but I personally have found those products extremely difficult to understand or use, and yours offers a lot more specific art direction, so I'm glad I found it.

Thanks as always for listening, and continuing to develop the product.

 

Thanks for your useful feedback

You are right, it would be possible to use multiple beams to generate more complex shapes.

I totally get your point, since I had the same experience with other assets: I had troubles to tweak them to get the rendering and effect I wanted. That was one of my main incentive to develop this custom package.

I'll let you know once I will have made progress on the custom beam shape feature. I am really looking forward to showing you the new exciting stuff to be included in the next updates!

 

Super excited to announce the Volumetric Light Beam 1.3 update, available now!! It features major changes and improvements asked by many users:

★ New 'Volumetric Dust Particles' feature (Unity 5.5 or above)
You can add volumetric dust particles to any light beam in 1 click, to simulate highly detailed dustlight / mote effects.


The new 'Volumetric Dust Particles' component lets you easily customize the rendering, density and movement of the particles in the inspector, which are rendered using Shuriken.


Since this feature relies on particles, depending on the number of particles you want to display, this could be performance intensive. That's why a culling system is built-in in this component: when enabled, the particles which are too far away from the camera (and so which would be barely visible) won't be rendered, saving on performance.


★ New 'Attenuation Equation' feature
Using the 'Volumetric Light Beam' inspector, you can now choose between several attenuation formula, used to compute how the beams fade between 'Fade Start Distance' and 'Range Distance'. This gives you more control on how the light beams scatter in the environment. You can choose between:
- Linear: Simple linear attenuation (legacy value).
- Quadratic: Quadratic attenuation. Since this is the way light behaves in real life, this usually gives more realistic results than the linear attenuation.
- Blend: Custom blending mix between linear and quadratic attenuations (can be customize using a slider).

The visual difference between each result is much more obvious in Unity​

In the inspector, there is a new 'Attenuation Equation' available.
Also, the 'Fading Distance' properties have been slightly changed in the inspector:
- 'Fading End Distance' is now 'Range Distance' (to be more consistent with the Light's 'Range' property)
- 'Fading Start Distance' is now easily tweakable via a convenient Slider (which goes between 0.0 and 'Range Distance')



★ Fix lights beams not properly rendered with orthographic cameras
This also fixes the light beams not properly visible in the editor scene view when using orthographic projection.


Here the updated showcase video highlighting some of these new features:

I sincerely hope you'll like these improvements!
Please consider writing a review on the AssetStore page if you are happy with the asset and the support

Thanks!

 

Hi guys,

A new update has been released on the store: Version 1.31
This is not a major update, but it brings some useful changes:
- Performance improvements.
- UI improvements (realtime infos in 'Volumetric Dust Particles' inspector).
- Noise3D workaround because of a bug in Unity Editor (not happening in build) when WebGL is the current platform.

I am currently working on more important features. I'll be able to share with you these exciting new stuff in the next days and weeks. Stay tuned

 

Happy to announce the availability of a detailed documentation about the plugin!

As the asset becomes bigger with more and more features, it was really necessary to write a manual/guide to clearly explain how to use it.
Useful tool-tips and hints are already visible directly in the Unity Editor, giving more information about each properties. But relying only on these couldn't scale well for the future and what's coming next

Even if I try to keep the asset as easy-to-use as possible, I still want to offer enough flexibility. This documentation will be a great place to highlight the features and the strength of the asset, but also its limitations to be extremely clear about what it can and cannot do.

I tried to make it as interesting as possible, with a lot of tips, images and animated gif. I sincerely hope you'll find it useful.

>> Volumetric Light Beam Documentation <<

 

Hi guys,

To celebrate the new year, I am pleased to announce that the Volumetric Light Beam 1.32 update is available on the AssetStore, and it includes a long-awaited feature: Dynamic Occlusion

The Dynamic Occlusion feature gives awareness to your light beams: they react to the changes in the environment. Light Beams can be blocked and occluded by geometry.
This new feature uses raycasting to compute a fast and efficient approximation of the occlusion. Therefore, the occlusion won't be perfectly precise, but will provide satisfying results in most cases while keeping high performance even on low-end platforms.



Please have a look at the detailed documentation to learn more about this new feature, how to use it, but also about its limitations.

In addition, the 1.32 update includes some minor UI improvements.

I hope you will find this feature useful

I am already working on the next update, which should make the 3D Noise feature available on mobile platforms. Stay tuned!

 

Thanks for the new features!! it's really great!

 

Hi guys,
I made a short and anonymous survey to learn more about your usage of the asset and your needs, in order to help me design better and more relevant features. So if you are a current user, or if you are interested in purchasing this asset in the future, please consider taking 1 min to fill the survey: https://goo.gl/forms/Hd1xU5QNrUMN6rYp1

Thank you very much in advance!

 

Hi guys,
I am working on very long-awaited features: rectangle-shape beams and light scattering simulation using cookie textures.
And I wanted to tease you a bit with the following video. This is still very WIP and it requires a lot of work, but I find the results already pretty awesome!

Stay tuned!

 

Hi!
I really like the look of you asset, but I have run into a problem where it looks good in the editor, but when building, noise is lost and the beam is much sharper. See attached screenshots.

 

Hi Johnny, sorry about this inconvenience.

- On which platform are you running the build?
- When running the build on the platform, could you execute the following line and tell me what it prints:
Debug.Log(SystemInfo.graphicsShaderLevel);

Thanks

 

The game is run on the same Windows 10 PC as I have Unity on so the shaderLevel should be the same

 

Please send me an email techsaladunity@gmail.com
I can provide you a WIP version with a possible fix, and I'll give some direction to find the issue.

Thanks

 

Hi guys,
Just wanted to clarify some stuff related to the issue @Johnny_Jr encountered, which is about instantiating and modifying VolumetricLightBeam via script.

With the current version 1.32, the beam is generated as soon as the component VolumetricLightBeam is attached to a GameObject.
So in short, if you want to change properties just after instantiating a VolumetricLightBeam component, you have to call UpdateAfterManualPropertyChange (or GenerateGeometry if you have changed geomSides and/or geomCap properties) after the changes to take them into account.

Code (CSharp):

1. var beam = gameObject.AddComponent<VLB.VolumetricLightBeam>();
2. beam.noiseEnabled = true;
3. beam.noiseIntensity = 1.0f;
4. beam.UpdateAfterManualPropertyChange(); // take changes into account

Note that you can also enable the trackChangesDuringPlaytime property to take any changes into account in realtime (like if you want to animate the beam) at a minor performance cost.

Anyway, I am currently working to improve these things, like being able to modify beam's properties just after instantiation without having to manually call a function. The documentation definitely needs more details about all that also.

 

Hi guys,

The 1.33 update has been released on the AssetStore, featuring:
- Add '3D Noise' support on mobile devices. Now the limitation is the following: 3D Noise feature requires shader capabilities equal or higher than Shader Model 3.5 / OpenGL ES 3.0. Any mobile devices released after 2012 should support it.
- Fix '3D Noise' bug not appearing in editor's scene view with WebGL set as the current platform.
- Improve scripting API: you can now change any properties just after instantiating a new beam from script, without having to call any other methods (Therefore the previous post is not relevant anymore).
- Improve compatibility with Unity 2018 beta.
- Fix issues when compiling shaders with STEREO_MULTIVIEW_ON.

Have a nice weekend

 

Great, thank you for all the work to make this useful asset even better. I guess you made it in a way that when newer features are disabled it almost has no negative effect on the performance on low end devices (compared to the first versions with less features)?

 

Absolutely!
The new "CPU" features such as Dust Particles and Dynamic Occlusion are implemented separately on dedicated Components, so they have no effect when not used.
The new "GPU" features are implemented via shader variants for the same reason.

Really glad you appreciate the new updates and features

I am also working on the rectangle-shape beams and light scattering simulation using cookie textures which will massively improve the asset capabilities. But it takes time since these new features require almost a complete rewrite and refactor. Stay tuned

 

Hey there,
I just bought your asset in the hopes that it will solve issues I'm having with other "post process" fog effects.

You are stating it should work with VR and Single Pass Stereo, but unfortunately NOT on mobile VR (android) with single pass stereo.
The issue seems to be with the depth buffer, same issue almost all other fx have. When building your demo scene for mobile vr and sps, there are nasty depth clipping errors, and some volumes, like the big green beam on the right, disappear when directly looking at it.
Tested on Moto Z Daydream (here even the depth buffer is completely offset, meaning the "volume hold out mask" of the cylinders is not on the cylinders but anywhere, and moves strangely when moving the head.
Tested with unity 2017.1 and 2017.3
With samsungs 6 the depth buffer seems mostly correct, but there are still some heavy visual artifacts visible, like depth planes slicing through geometry and effects.

I would really appreciate if you could investigate, or I bought this in vain once again :/

 

Hi,

Thank you for your purchase. I am sorry about the issue you are experiencing.
Based on your comment, it seems that you are experiencing depth precision issues, which is quite common on mobile (also know as "banding") because depth buffer is usually encoded in a low precision buffer on these platforms. There is a note about this in the documentation: http://saladgamer.com/vlb-doc/comp-lightbeam/#opaque-geometry

You can try the following:
- First disable the Soft Intersections with Opaque Geometry. Simply set the "Opaque Geometry" property to 0 in the inspector. With this feature disabled, there is no depth sampling anymore and you should not experience any visual glitch anymore.
- If you want to re-enable it again, you will have to lower the distance between the near and far clip distance of your camera. Basically set the near distance as big as possible, and the far distance as low as possible.

Please let me know how it goes.
Thanks

 

Thanks for the very quick reply!
There definitely seems to be an issue with daydream or my specific Moto Z phone. Depth buffer just is not correct.

However, turning Soft Intersections to 0 helps.
I couldn't see an improvement lowering the cameras far clipping plane (in the demo scene to something like 0.1 near and 35 far) It still shows a lot of slicing when moving the head (Correction: it helps a lot increasing the near clipping plane, not sure how practical this is).

Also turning of soft intersections is not the look I wanted to achieve :/
Is there no way to improve the look on mobile with soft intersections?
Thanks,
Phil

 

Hi Phil,

I will investigate to check if there are better ways to sample depth buffer on mobile.
Unfortunately, I am quite pessimistic about that, since it's a common issue on mobile. And regarding the feature, there is no magic: it relies on the depth buffer, so if the depth buffer has a poor precision, the feature will produce poor results.

I am not aware about the specificities of depth sampling in mobile VR with single pass stereo. Have you tried to use Multi Pass stereo mode to check if the depth sampling is not accurate?

You can also check in the Player Settings in the Android tab to make sure "Use 32-bits Display Buffer" is checked.
According to the Unity Documentation, it helps regarding "banding" artefacts: https://docs.unity3d.com/Manual/class-PlayerSettingsAndroid.html


In the meantime, the only workaround is to disable blending with opaque geometry.
I'll let you know if I can make some improvements.
Please let me know if my suggestions help.
Thanks

 

The issue with the wrong depth buffer is only on: mobile VR daydream (Moto Z) with Single Pass Stereo enabled.
I just noticed, that the depth buffer is correct in the left eye, but offset / wrong in the right eye.
All good with Multipass.
All good with Single Pass on Oculus Samsung S6 Phone.
This might even be a driver issue (qualcomm), I read something about that in a daydream forum, but not sure if my moto z has that issue. Cannot test with another daydream ready phone at the moment.

The 32bit display buffer was activated. But you are right, it is not related to the volume, but generally the low light scene, also I have weird issue on GearVr where the fog is smooth, but the point light is stepped / has banding, only on gearvr, not on oculus. Will have to figure this out But while testing and switching phones I thought it is related to the plugin.

So to sum up, I think all is good for now, except the sps issue on daydream. I can work with multipass for now and maybe somehow this issue will be fixed from google/unity.

Regarding the banding, maybe you could implement on your end some sort of dithering? But I guess that would only be applied to banding in the fog beam. For the common depth precision slicing/banding in low light situations, there is maybe another way to smooth things out.

Anyway thanks for your help! I

 

Thank you very much for the details. Like you said, it really sounds like a driver issue...
I will investigate if there is a way to smooth the depth sampling despite the low precision.
Thanks again

 

Hey @phili_maas, have you checked this parameter? It's in the PlayerSettings, in the Android XR Settings. You can choose the depth format per VR platform (Daydream, Cardboard...). Could you check this parameter for Daydream is at the highest value (24-bit depth) please?

Let me know if it helps

 

Hey, will you support point lights in the future? It is the main light I use and really need it to have these nice effects.

 

Thanks for your interest.
With the current technique I use, the result with point lights won’t be satisfying enough. I have currently no plan to support point lights. My main focus is to improve and add features to spotlights support while keeping the compatibility with mobile, VR, WebGL...
So to be honest, if you need point lights support, I would suggest you to look for an alternative.
Thanks.

 

I already abought the asset and will use it. But point lights would be nice. Any place where I can search for point light volumetric lights?

 

Yes I know that point lights support would be appreciated. I am actually rewriting most of the asset code to support cookie textures. Once I will be done with this feature, I will investigate if there is a way to support point lights, but it would require quite a lot of R&D and so I cannot guarantee anything.

Regarding alternative solutions, have a look at these open source projects:
https://github.com/SlightlyMad/VolumetricLights
https://github.com/Unity-Technologies/VolumetricLighting
However, other techniques usually require a higher shader level and are heavier performance-wise. So they might not support low performance platforms such as mobile. So depending on your needs, it might fit with your project.

 

Thanks pal!

 

Great asset, thank you. And the rectangle-shape beams and light scattering simulation using cookie textures tease look awesome, can't wait.

Can I make a feature request? (Or maybe it can be done but I haven't found the right settings?...) Would it be possible for the cone to render with alpha at the start of the cone? So that the beam essentially looked as if it came out of nowhere (wouldn't need a light source to look right) Here's a quick photoshopped idea of what it might look like (original on the left)

 

Hi! Thanks for your suggestion
Sure it can definitely be done. I will take it into account for the next update if possible.

 

Thanks for the great asset !!! i came to the forum to ask support for light cookie, but apparently you are already working on this ! i saw the rectangle exemple, it is a very nice effect. Can we expect the update anytime soon ? it look already working and polished !

Will it support any type of cookie for point and spot light ? Combined with the IES importer avaiable on the store, it could produce tremendous effect of realistic light ! right now i take 2 light to create the effect which is bad for performance

 

Hey! Thanks a lot for your interest. Yes I am actively working on the feature.

To answer your questions, it will support any type of cookie texture. However it will only support spotlights. I don’t have plans for point lights support for now.

And I cannot give you an ETA. It’s a major rewrite of the plugin and I want to keep the highest quality as possible. So it takes some time to test on every platforms and make sure it works perfectly with good performance.

Thank you for your understanding. I’ll make sure to keep this thread up to date!

 

I have no worry for the quality !
point light is good for "light on wall" that normally trigger light being itself which is the case for 99% of wall-mounted light. It would be really great for you to support this. Wall light can be in fog too and can need ambiance too it fits in your asset and if you don't do it, who will ?

i have know idea how all this work, but just idea that come through my head :
Right now, as a cheap workaround, wall-mounted point light can be supported by having a spot set in front of the point light facing the point light (for wall mounted light) and the volumetric cone can be pointed in the same direction as the light (opposite of the wall) with a very wide 179.9 angle. As most point light are either attach to a wall or ceiling you could have a "half-sphere" volume instead of a cone, have two of them on both side to form a circle, allow us to select if we need both (floating light bulb) or only 1 (ceiling and wall) have the projection for cookie to face center of sphere. With 1 full sphere it would be problematic since volume would go through the surface which would be mostly unwanted.

 

Thanks for your suggestion.
I am aware that a lot of people would be really interested in point light support.
But it would really require a lot of R&D to achieve a good result with the technique I currently use for spotlights. The thing is, with spotlights, the beam of light is usually very focused, and so it's not a big deal to not handle "shadows" and light scattering. But by supporting point lights, it would be almost mandatory to also support these features to achieve a decent result. And the problem is that it really difficult to do that while keeping high performances on devices such as mobile.

So I will investigate these features, but I have honestly no idea how long it could take...
That's why I would prefer to release features such as rectangle shapes and cookie textures for spotlights first, since these features are also long-awaited, and I know that they are definitely doable in a reasonable amount of time

 

i wish you good luck with that challenge !!! one feature at a time, people will certainly continue to buy. Maybe having a shader for mobile and one for higher end device to not limit yourself with mobile capability and performance.

 

Why is the light cone in your demo also illuminating the ceiling when you look through the cone is this a bug?

 

Hi. Sorry I don’t get it. Could you post a screenshot of the issue please?
Thanks

 

Hi guys,

The 1.34 update has been released, aiming for a stronger and more reliable API. Here is the changelist:
- 'trackChangesDuringPlaytime' property can now be changed in real-time.
- All properties are now properly clamped when changed during playtime from script and/or animator.
- Editor UI improvements.

For the next update, I plan to do some refactoring in order to prepare for more massive changes. Basically, I plan to merge the 'outside' and 'inside' alpha properties. There will be only 1 alpha value to control for both the inside and the outside rendering of the beam.

Would you be fine with this change?
Or do you really need to have 2 distinct properties (the inside alpha value different from the outside alpha value)?
Please let me know!

Thanks

 

In your Image and in the Videos the Light is Iluminating the ceiling which should be impossible for this light.

 

Thanks for the clarification.
Are you talking about this area?


Because the light is NOT illuminating the ceiling. This area looks brighter because it's the volumetric light beam which is in front of the camera... But the ceiling is not illuminated: if the camera goes beyond the light beam (like it does in the video), you can see that the ceiling is dark.
Thanks

 

Hi,

I want to buy it, but I really need box beam. when do you add it?
can I use cone shape to achieve box shape?

Thanks

 

Hi!
Thanks a lot for your interest. No you cannot achieve a box beam using the current cone shape. That's why I am developing a specific feature for it
This feature is still in development. I know a lot of developers are expecting it! I can't give you an ETA because I don't want to communicate a release date without being sure I could keep it. But it will be released and I will make the announcement on this thread.
Thanks again.

 

hello,

I bought this plugin and Im happy to use it, but I have maybe silly question - where can I set sorting layer for the light beam? I cannot find the option.

 

Thanks! And glad you are happy with it
Regarding your question, do you mean the sorting layer to adjust the rendering with 2D sprites? If yes, it's not implemented yet but this feature is part of the roadmap. Actually I plan to release this feature very soon.
Thanks

 

This made me look as well, and I can see places behind the volume getting illuminated without any light source. The explanation is confusing as there is no way the ceiling (in the example) can receive any light and illuminate it. This feels like a bug or limitation

 

The ceiling is not pitch black because of ambient lighting and light bounces.
Anyway, if there are some lightmapping issues, it's absolutely not related to the plugin, since the actual lighting (lightmap baking, GI...) is still done by Unity. This plugin only adds some procedural volumetric shapes to improve the visual quality.
Hope it is more clear

 

Thanks,

I've bought your package btw, hope you add this feature soon.