Version: 2021.1
Occlusion culling additional resources
Dynamic resolution

CullingGroup API

CullingGroup offers a way to integrate your own systems into Unity’s culling and LODThe Level Of Detail (LOD) technique is an optimization that reduces the number of triangles that Unity has to render for a GameObject when its distance from the Camera increases. More info
See in Glossary
pipeline. This can be used for many purposes; for example:

  • Simulating a crowd of people, while only having full GameObjectsThe fundamental object in Unity scenes, which can represent characters, props, scenery, cameras, waypoints, and more. A GameObject’s functionality is defined by the Components attached to it. More info
    See in Glossary
    for the characters that are actually visible right now
  • Building a GPU particle system driven by Graphics.DrawProcedural, but skipping rendering particle systemsA component that simulates fluid entities such as liquids, clouds and flames by generating and animating large numbers of small 2D images in the scene. More info
    See in Glossary
    that are behind a wall
  • Tracking which spawn points are hidden from the cameraA component which creates an image of a particular viewpoint in your scene. The output is either drawn to the screen or captured as a texture. More info
    See in Glossary
    in order to spawn enemies without the player seeing them ‘pop’ into view
  • Switching characters from full-quality animation and AI calculations when close, to lower-quality cheaper behaviour at a distance
  • Having 10,000 marker points in your sceneA Scene contains the environments and menus of your game. Think of each unique Scene file as a unique level. In each Scene, you place your environments, obstacles, and decorations, essentially designing and building your game in pieces. More info
    See in Glossary
    and efficiently finding out when the player gets within 1m of any of them

The API works by having you provide an array of bounding spheres. The visibility of these spheres relative to a particular camera is then calculated, along with a ‘distance band’ value that can be treated like a LOD level number.

Getting Started with CullingGroup

There are no components or visual tools for working with CullingGroups; they are purely accessible via script.

A CullingGroup can be constructed using the ‘new’ operator:

CullingGroup group = new CullingGroup();

To have the CullingGroup perform visibility calculations, specify the camera it should use:

group.targetCamera = Camera.main;

Create and populate an array of BoundingSphere structures with the positions and radii of your spheres, and pass it to SetBoundingSpheres along with the number of spheres that are actually in the array. The number of spheres does not need to be the same as the length of the array. Unity recommends that you create an array that is big enough to hold the most spheres you will ever have at one time, even if the initial number of spheres you actually have in the array is very low. Using a larger array allows you to add or remove spheres as needed without the computationally expensive process of resizing the array at runtime.

BoundingSphere[] spheres = new BoundingSphere[1000];
spheres[0] = new BoundingSphere(Vector3.zero, 1f);
group.SetBoundingSpheres(spheres);
group.SetBoundingSphereCount(1);

At this point, the CullingGroup will begin computing the visibility of the single sphere each frame.

To clean up the CullingGroup and free all memory it uses, dispose of the CullingGroup via the standard .NET IDisposable mechanism:

group.Dispose();
group = null;

Receiving results via the onStateChanged callback

The most efficient way to respond to spheres changing their visibility or distance state is to use the onStateChanged callback field. Set this to a function which takes a CullingGroupEvent structure as an argument; it will then be called after culling is complete, for each sphere that has changed state. The members of the CullingGroupEvent structure tell you about the previous and new states of the sphere.

group.onStateChanged = StateChangedMethod;

private void StateChangedMethod(CullingGroupEvent evt)
{
    if(evt.hasBecomeVisible)
        Debug.LogFormat("Sphere {0} has become visible!", evt.index);
    if(evt.hasBecomeInvisible)
        Debug.LogFormat("Sphere {0} has become invisible!", evt.index);
}

Receiving results via the CullingGroup Query API

In addition to the onStateChanged delegate, the CullingGroup provides an API for retrieving the latest visibility and distance results of any sphere in the bounding spheres array. To check the states of a single sphere, use the IsVisible and GetDistance methods:

bool sphereIsVisible = group.IsVisible(0);
int sphereDistanceBand = group.GetDistance(0);

To check the states of multiple spheres, you can use the QueryIndices method. This method scans a continuous range of spheres to find ones that match a given visibility or distance state.

// Allocate an array to hold the resulting sphere indices - the size of the array determines the maximum spheres checked per call
int[] resultIndices = new int[1000];
// Also set up an int for storing the actual number of results that have been placed into the array
int numResults = 0;

// Find all spheres that are visible
numResults = group.QueryIndices(true, resultIndices, 0);
// Find all spheres that are in distance band 1
numResults = group.QueryIndices(1, resultIndices, 0);
// Find all spheres that are hidden in distance band 2, skipping the first 100
numResults = group.QueryIndices(false, 2, resultIndices, 100);

Remember that the information retrieved by the query API is only updated when the camera used by the CullingGroup actually performs its culling.

CullingGroup API Best Practices

When considering how you might apply CullingGroup to your project, consider the following aspects of the CullingGroup design.

Using visibility

All the volumes for which CullingGroup computes visibility are defined by bounding spheres - in practice, a position (the center of the sphere) and a radius value. No other bounding shapes are supported, for performance reasons. In practice this means you will be defining a sphere that fully encloses the object you are interested in culling. If a tighter fit is needed, consider using multiple spheres to cover different parts of the object, and making decisions based on the visibility state of all of the spheres.

In order to evaluate visibility, the CullingGroup needs to know which camera visibility should be computed from. Currently a single CullingGroup only supports a single camera. If you need to evaluate visibility to multiple cameras, you should use one CullingGroup per camera and combine the results.

The CullingGroup will calculate visibility based on frustum culling and static occlusion cullingA that disables rendering of objects when they are not currently seen by the camera because they are obscured (occluded) by other objects. More info
See in Glossary
only. It will not take dynamic objects into account as potential occluders.

Using distance

The CullingGroup is capable of calculating the distance between some reference point (for example, the position of the camera or player) and the closest point on each sphere. This distance value is not provided to you directly, but is instead quantized using a set of threshold values that you provide, in order to calculate a discrete ‘distance band’ integer result. The intention is that you interpret these distance bands as ‘close range’, ‘medium range’, ‘far range’, and so on.

The CullingGroup will provide callbacks when an object moves from being in one band to being in another, giving you the opportunity to do things like change the behaviour of that object to something less CPU-intensive.

Any spheres that are beyond the last distance band will be considered to be invisible, allowing you to easily construct a culling implementation that completely deactivates objects that are very far away. If you do not want this behaviour, simply set your final threshold value to be at an infinite distance away.

Only a single reference point is supported per CullingGroup.

Performance and design

The CullingGroup API does not give you the ability to make changes to your scene and then immediately request the new visibility state of a bounding sphere. For performance reasons, the CullingGroup only calculates new visibility information during execution of culling for the camera as a whole; it’s at this point that the information is available to you, either via a callback, or via the CullingGroup query API. In practice, this means you should approach CullingGroup in an asynchronous manner.

The bounding spheres array you provide to the CullingGroup is referenced by the CullingGroup, rather than copied. This means you should keep a reference to the array that you pass to SetBoundingSpheres, and that you can modify the contents of this array without needing to call SetBoundingSpheres again. If you need multiple CullingGroups that calculate visibility and distances for the same set of spheres - for example, for multiple cameras - then it’s efficient to have all the CullingGroups share the same bounding sphere array instance.

Occlusion culling additional resources
Dynamic resolution
Copyright © 2023 Unity Technologies
优美缔软件(上海)有限公司 版权所有
"Unity"、Unity 徽标及其他 Unity 商标是 Unity Technologies 或其附属机构在美国及其他地区的商标或注册商标。其他名称或品牌是其各自所有者的商标。
公安部备案号:
31010902002961