Official New Post-processing Stack (Pre-Release)

Well it's not an urgent task at all... When you get to fussing over a few checkboxes, life must be incredibly good.

 

It's pretty good, but it could be perfect if only those darn CHECKBOXES WERE IN ORDER!

(╯°□°）╯︵ ┻━┻

(I'm kidding, the checkboxes are fine)

(although, Unity in general, does suffer from UI inconsistencies here and there)

 

I'll have to go back and give V2 a try then. I switched back to V1 since that was what was used in the PS4 version, however now running SinglePass rendering get a black screen at start with color grading enabled, that only happens the first time the game is run. Not the second time. Doesn't make for a good first impression to a user

 

There's a new commit that changes NaN propagation that might fix the screen blowing up when using TAA/Bloom etc.
https://github.com/Unity-Technologies/PostProcessing/issues/337#issuecomment-363077825
https://github.com/Unity-Technologies/PostProcessing/commit/b40105bcb9003522fc3c9e189c85f00950f3f354

Update: it didn't work for us.

 

LOL. Guess I learned my lesson on giving user feedback to this particular forum.

 

The lesson being you got some developer feedback, so it must be good

 

I'm not advocating for making the whole Unity UI look like the particle system. My main point is the grey circular buttons in V2 are not consistent really with anything in the Unity UI, and the boxes in V2 were more intuitive and consistent with the Unity UI in general.

 

I'm with you on this. The tool would be better if changed. It seems pretty obvious, not sure why there'd be resistance.

The rebuttal offered above is kind of silly, saying that if the interface isn't good, well then your life must be great, so why worry.

I suppose a similar line of reasoning would be that maybe your life is really bad, so why worry.

In fact, why worry about making any changes or improvements to anything, because your life is either too good or too bad for it to matter?

I'm actually not going to worry about it. lol

That's because I won't use bad tools. My advice is that until they improve it, skip it.

 

Hi, any plans of adding atmospheric effects, volumetric fog etc... !?

 

I think they are focused on making physical camera now, thats why pp stack doesnt have any updates for some time. Btw, srp team working on the sky and volumetric lighting in hd pipeline. As for me, I would like to see this effect implemented in stack http://bankslab.berkeley.edu/publications/chromablur/

 

Thanks for the feedback !

As for the "chromablur" - wow. I always though how cool would be if chromatic aberration could be somehow applied to the whole image, simply because it somehow improves realism !

 

Edit, nvm. Solved my own problem.

 

I'm making a game that is about taking photos using post processing stack v1 atm. Is there a way to calculate if an object at a given distance is in focus if we know all the Depth of Field settings plus resolution etc? I've been digging through loads of camera theory and the depth of field shaders and scripts over the last two days but I can't work it out.

 

I've never tried anything like that, and I'm not 100% I get what you're looking for. But you're trying to see if a object is physically in range to be clear for the player? If so all you need is to look at the distance of your object and cross reference that with your Depth of Field's Focal Point. Right off the top of my head the easiest way I would think to do that would be to have a raycast that checks the object it's hitting and it's distance. And since you know what the distance is for the focal point you can cross reference that with your object and determine if it's in range.

 

I, too, am making a game about taking photos, but my issue is slightly different. By any chance, does PPSv2 solve the issue where shadows are immune to motion blur? Right now shadows are very sharp regardless of how blurry the object is. See below:

 

Wow that shadow sharpness is really hard to ignore once you notice it. I'm also using motion blur to simulate shutter speed. I'll try and explain my focus problem a little better. Basically the focus distance isn't always going to match object we are trying through the camera settings to focus on. This is because I cap the focus distance, as many camera lenses will simply have everything in focus beyond a certain distance, for the sake of clarity the max I have atm is 5(m). The current default lens I have setup in game simulates a fairly standard zoom lens with 18-55mm focal length and a max zoom of approximately 3x (the zoom is actually determined by the focal length as in a real lens). Similarly it has an aperture range from f/3.5 to f/5.6, presently completely under user control but it seems that it should creep up with the focal length unless its a really good quality lens.

Basically with all these settings you can reasonably easily get quite a wide or narrow DoF that isn't necessarily dictated by the focus distance. The issue I have is if there are bunch of photo target points within the picture at different distances how do I tell which ones are in focus (assuming the objects in question are relatively narrow). I found a heap of tables listing near and far focus planes for various lens configurations but I can't find any that behave like the Unity DoF. Its further complicated by the way the DoF changes according to the screen resolution in Unity. When I take a photo in game I super sample it up to the desired output resolution and the DoF will always look a bit different to the way it looked in the camera viewer. I think the answer might be in understanding how the Post Processing stack calculates Circle of Confusion and how that changes according to the resolution. I found some logic about CoC in the Post processing stack DoF code but I couldn't get the math right or there is more to it.

 

Hello,
I am rendering two video textures on planes in a car interior for ui simulation. Shader used for this is the avpro unlit shader.
Is it somehow possible to exclude these from tonemapping and bloom in PP V2.
Means. Bloom and Tonemapping works for car interior but is not affecting the two video textures for my ui displays?

 

My team is also using 2017.1.1f1, and I am not prepared to update them just to test the beta v2 stack, but I do need to workaround the black rectangles/screen TAA bug others have run into. I have v1 working fine, HDR+deferred+TAA+bloom+Color Grading

What version of v2 is stable for people on 2017.1.1f1 (Windows and Xbox)? I pulled the latest today to get the Xbox compile fix, and Bloom and Color Grading seem to have no effect, only AO works, and TAA is shaking my screen. I'm getting these warnings, that I imagine mean that 2017.1.1f1 isn't really supported in this pull of v2:

• CommandBuffer: built-in render texture type 3 not found while executing Post-processing (SetRenderTarget depth buffer)
• CommandBuffer: built-in render texture type 3 not found while executing Opaque Only Post-processing (SetRenderTarget depth buffer)

I guess I will go make some hacks to clamp in TAA, but I really wanted to test out some version of v2, even if it is an older beta version.

 

Can anybody give me any kind of tips how can I solve such problem:
I have Post effects A - a game world
I have Post effects B - render texture and other world with more bloom, sepia, etc.
I want whole game to be with effect A, but WITHOUT the portal's render texture
I want the other world (other world means normal world + Vector3.right * 1000f) to be with effect B, but also Portal's render texture to be with effect B
3. Player moves through portal to other world - main camera switches from A to B
4. Player comes back from portal world to main world - main camera switches from B to A

Althrough changing post processing effects profile during game is simple, I can't manage how can I make a part of my game with different post processing effects - the render texture. When I add a post effects component to the main camera, whole game has effects, even portal's render texture. Disabling Layer in Post process layer in main camera doesn't change anything, it's like not working.

Look on the screenshot, the world has post processing, the render texture in the portal has exactly the same effect as main world which is bad, because I want the render texture to have the post processing of the other camera which is looking through the portal.

 

Can you render the portal (quad with a render texture, or something similar) after the post effects stack is done on your main camera ("A")?

 

This sounds really logical. How can I render the rendertexture after post effects? Can I somehow set an order of rendering?

 

I am not sure yet, I am investigating today.
EDIT:

• I think we are getting off topic, so further discussion warrants a new thread I think.
• If we were using 2018.1 (beta): it looks like the Scriptable Render Pipeline would be a good way to customize the order of rendering, as you can pick when combinations of renderqueue/layer are rendered. I don't know if the post effects stack is compatible with 2018.1, but v2 in github might be.
• If using a version before 2018:
  • Render the portal with a Command Buffer that will execute after post effects. OR
  • I think this would be more complicated than the command buffer, but you could write a post effect that runs last in the stack and composites the render texture with the portals into the main camera render, as rendered to separate camera and render texture, checking that depth buffer render texture generated from the portal quad camera and the depth buffer from your main camera.

 

what am I supposed to do with this? this is the 2.0.2 preview, but also get the same bugs as 2.00 beta 5 version. what do I do to avoid this? or am i screwed until its fixed? I am currently using 2017.3.1f1

 

I just had this same error on Unity 2017 31 !:
"Invalid pass number (1) for Graphics.Blit (Material "(Unknown material)" with 1 passes)"
after changing from Gamma to Linear. Closing and opening Unity did not solve it.
After the workaround of "Reimport All" assets, the warning is gone and instead I have this yellow one:
"
Rebuilding GUID cache: Deleting metadata Packages/com.unity.ads/Editor/Resources/iOS/builds/UnityAds.framework/Headers/UnityAds.h because the asset doesn't exist anymore.
MetaData path library/metadata/c9/c99bfe4fc4e5c4dc3917c14075acd117
"
Now back on work again!
Hope this helps others here !

 

Did a solution ever come to this because i'm running into the same issue, can't seem to fix it.

 

Did a solution ever come to this because i'm running into the same issue, can't seem to fix it.

 

Is it possible to increase the sharpeness of the SSR?
I've tried also the "Overkill" preset or a custom one but I got only blurred reflections.

I think that it will be really good to have an option that doesn't blur anything and leave to the material smoothness.
Is it something doable with SSR?

Thanks in advance.

 

1. it's based on screen resolution. The blur is the result of the optimization.
2. You can use Legacy SSR from Legacy Cinematic Imageeffects on assetstore or
https://github.com/cCharkes/StochasticScreenSpaceReflection

Legacy SSR in low preset and proper settings has higher performance compare to Stochastic

https://forum.unity.com/threads/new-post-processing-stack-pre-release.435581/page-14#post-3225462

 

Thank you for your reply.
I've tested the Post Process Stack V1 and the SSR seems really more sharp than the V2 one and as performance it seems really equal if not better.
I hope to see something similar, as quality, in V2 because, at the moment, any kind of settings you use, it always blurred.

 

Gotta agree though, SSR in V2 seems reduced in quality. Not sure about performance.

 

Are there any tips for upgrading from V1 to V2? Or is it as simple as replacing the code and creating new V2 profiles?

 

Settings are same in most effects

only Color Grading has a bit different settings (x100)

V1 :
0-1

V2:
0-100

 

Thanks, I ended up diving in and figuring it out over the weekend.

For anyone else the basic steps of going from V1 to V2 are you have to throw out all the old V1 profiles, create new ones you assign to post process volume components which you make part of the scene instead of attached to your camera. Assign a custom layer to your object with the volume component attached, which you select in the post processing layer component you attach to the camera (which you use instead of the old V1 post processing component attached to the camera, where you used to assign the profile directly, now it is indirectly through layers).

Finally if you want your anti-aliasing settings configurable by the user, you'll need to write a little code to change the AA settings on each camera's post processing layer component as it is no longer part of the post processing profile.

 

Anyone, or @Chman is it possible to exclude objects from motion blur?
Ah Nvm

 

PostProcessLayer resets the projection matrix on the camera automatically every update. Is there a way to disable that? I have an Asset Store asset that does custom manipulation of the camera matrix for off-axis projection.
https://assetstore.unity.com/packages/tools/camera/offaxiscamera-98991

Because Postprocessing resets the camera projection matrix it is incompatible with OffAxisCamera at this time =(

EDIT: It happens regardless the Anti-Aliasing setting. (TTA use matrix jittering).

EDIT 2: I found the sinner. Line 279 in PostProcessingLayer.cs
https://github.com/Unity-Technologi...stProcessing/Runtime/PostProcessLayer.cs#L279

Code (CSharp):

1. // Resets the projection matrix from previous frame in case TAA was enabled.
2. // We also need to force reset the non-jittered projection matrix here as it's not done
3. // when ResetProjectionMatrix() is called and will break transparent rendering if TAA
4. // is switched off and the FOV or any other camera property changes.
5. m_Camera.ResetProjectionMatrix();

EDIT3: Please make modification of the projection matrix optional. Or make it possible to set reference to a function like it is done for the jittered projection matrix in TAA.

 

Also having this problem - post effects disappear when pressing play.

 

Hi, found a bug in post processing v2
It is happening in 2018.1 b13 with v2.0.3 added using the Package Manager.
The issue is that if you move an object is scene in play mode it remains blurred until you toggle it off and on again. Not sure if this is expected or an issue - but seems to me as an incorrect behavior !

 

Just a simple question...I hope with a simple answer.
Is it possible to exclude one or more geometries from the ScreenSpace Reflection effect (V2) ?
Thanks in advance

 

Has anyone tried the DepthOfField for V2? Does it work good like the legacy Depth Of Field image effect?
The V1 of Unity Post Processing stack DepthOfField is just far to difficult to get anything reliable. Something just seems very wrong with the way it works. I reverted back to the legacy DoF because I'm getting too frustrated trying to figure out what the V1 settings should be to simulate what the legacy DoF could do.

 

It's not possible to get samething like legacy one
You can just have ~good settings here:

Code (CSharp):

1.         dof.aperture.value = 0.9f;
2.         dof.focalLength.value = 2f;
3.         dof.kernelSize.value = KernelSize.Medium;
4.         dof.focusDistance.value = dofDistance;
5.  

modify dof distance

 

I've mentioned before that it needs simplified, but certain people seem to want to work in the current unusable way.

The focus puller here might be useful - https://github.com/keijiro/PostProcessingUtilities

https://forum.unity.com/threads/new-post-processing-stack-pre-release.435581/page-10#post-3079888
https://forum.unity.com/threads/new-post-processing-stack-pre-release.435581/page-15#post-3260586

 

Edit: see below

EDIT:
It is strange but after further trying to figure out the issue, I noticed that the DoF effect changes based on the window size of the player. I can't make sense of it, but this is the issue. It was unbeknownst to me until now.
Here is the game window at 964x542. No DoF is noticeable.


Now I increase the game window. Now the DoF is visibe


k
This is why I was struggling so much figuring out the settings.

Can anyone else try it out to see if the DoF changes based on the player window size?

 

Has anyone tried out resizing the player window to see if the DoF effect changes? Like I mentioned above, when the player window is shrunk down, the DoF effect is almost completely gone. I can reproduce this every-time in the editor playmode and when running a standalone

 

Same here..

 

There's still a few problems in WebGL after all this time:

• Depth of Field not supported.
• Ambient Occlusion in Ambient/Indirect only not supported.
• SSR not supported.
• Auto Exposure not supported.

There needs to be non compute shaders versions of each of these effects. All apart from Auto Exposure were working in the previous stack.

 

05/07/2018 Just downloaded Unity and the Post Processing stack, however the PP gives me a bunch of compiler errors. Is there a fix in sight?

 

Upgraded to v2 today and Ambient Occlusion isn't behaving very well with Orthographic cameras. It just adds a dark shadow gradient across the top of the camera and doesn't affect the rest of the screen.
Tried it on a new blank camera with the same results.

Any ideas, am I missing something simple or is it broken atm?

 

I'm testing out the LightweightVRTemplate which comes with 2018.1.0f2 and none of the post processing settings seem to be working in Play mode. They are displaying fine in the game view when I'm not playing, but as soon as the game starts all post processing disappears.

 

Same. Can't get anything to work despite following rules to the letter.

 

Hello.

Post Processing Stack asset is great but it lacks a feature. Unlike "Beautify" or the other assets, this one doesn't have a sharpening effect.

​

Please add the effect to the stack.
Thanks in advance.