Unity Rendering Paths: Forward vs. Deferred Rendering

    In Unity, rendering is the process of transforming your 3D game assets into the 2D images that players see on their screens. It's a crucial aspect of game development or real-time 3D software development, impacting both performance and visual quality. Unity offers two primary rendering paths: Forward Rendering and Deferred Rendering. In this blog, we'll explore the differences between these two approaches and their implications for your unity development. Bonus! we have also some Forward and deferred rendering FAQs at the end of this post.

Forward Rendering

1. Basic Workflow:

    In Forward Rendering, each object in the scene is rendered individually in a single pass. This means that each object's geometry, materials, and basic shading are processed in one go.

Forward Flow (Ref: Unity for Technical artist - ebook by Unity).

 

2. Multiple Passes for Effects:

    While Forward Rendering initially renders each object in one pass, additional passes are used for various effects such as lighting, shadows, and post-processing. If you have multiple lights, each light may require its own rendering pass to calculate lighting contributions.

3. Efficiency and Quality:

    Forward Rendering is efficient for scenes with a small number of lights and simple materials. It's suitable for projects targeting lower-end hardware or mobile devices. The quality of rendering in Forward Rendering depends on the number of lights and the complexity of materials. It may require compromises in scenes with many lights and intricate shading requirements.

Deferred Rendering

Deferred Flow (Ref: Unity for Technical Artist - ebook by Unity).

1. Initial G-Buffer Pass:

    In Deferred Rendering, the process is different. The initial geometry pass (often called the G-buffer pass) renders objects and records information about their positions, normals, materials, etc., into separate buffers (the G-buffer).

 

2. Separate Passes for Effects:

    Lighting, shading, and other effects are applied in subsequent passes, using the information stored in the G-buffer. These passes are more efficient when dealing with multiple lights and complex materials.

3. Efficiency and Quality:

    Deferred Rendering is generally more efficient in scenes with a large number of dynamic lights and complex materials. It minimizes overdraw and offers better performance in these scenarios.

It can also provide better quality in scenes with numerous lights and complex shading requirements, as it can handle a high volume of lights more efficiently.

FAQ: Rendering Paths in Unity

1. Which rendering path should I choose for my project?

The choice between Forward Rendering and Deferred Rendering depends on your project's specific requirements and constraints. Forward Rendering is ideal for simpler scenes with fewer lights, while Deferred Rendering excels in complex scenes with numerous dynamic lights and complex materials.

2. How to change rendering path settings in Unity?

To change the rendering path in Unity, follow these steps:

1. In the Unity Editor, go to Edit > Project Settings > Graphics.
2. In the Graphics settings, you can go to Tier Settings.
3. In tier 3 (which is used for Desktop), you can select either Forward or Deferred rendering. Please see the below given picture.

Note: Consider your target platform, scene complexity, available hardware, development team's expertise, and visual style when choosing the rendering path that best suits your game's needs.

3. What is a G-buffer?

A G-buffer is a set of buffers used in Deferred Rendering to store information about the objects in the scene. It typically includes data such as positions, normals, materials, and more, which is then used in subsequent rendering passes for lighting and shading calculations.

4. What's the difference between the initial pass in Forward and Deferred Rendering?

Both Forward and Deferred Rendering use an initial pass to render objects, but the key difference is in how this information is utilized. In Forward Rendering, the initial pass includes more basic shading, while in Deferred Rendering, the initial pass creates a G-buffer to store data for later, more complex shading passes.

5. Why is G-Buffer called G-Buffer?

A G-buffer, short for "Geometry Buffer," is a set of intermediate framebuffers or render targets used in the Deferred Rendering pipeline. These buffers store various pieces of information about the scene's geometry and materials for each pixel, often including:

• Position: The world space position of each pixel.
• Normal: The surface normal at each pixel.
• Material: Information about the material properties at each pixel, such as color, reflectivity, etc.
• Depth: The depth information, which can be used for depth-based effects and occlusion.

The term "G-buffer" arises from the fact that it primarily stores data related to the geometry of the scene. It allows the renderer to decouple the initial geometry pass from subsequent lighting and shading passes, improving efficiency and enabling more complex effects.

6. What happens during the initial pass in Deferred Rendering?

During the initial pass in Deferred Rendering, the focus is not on rendering the final appearance of objects but on gathering and saving information about the objects in the scene. This pass is responsible for rendering the objects' geometry and material properties into the G-buffer.

The initial pass in Deferred Rendering does the following:

• Renders the scene's geometry, but it doesn't directly calculate the lighting, shadows, or final shading.
• Records information about each pixel's position, normal, material properties, and depth into the G-buffer.
• Sets the stage for subsequent passes, which use the G-buffer's information to calculate lighting, shading, and other effects more efficiently.

In essence, the initial pass in Deferred Rendering is more about saving data that is used in later rendering passes to achieve complex lighting, shading, and other visual effects with improved performance.

6. Why g-buffer is not available in forward rendering

G-buffer is a concept specific to Deferred Rendering and is not used in Forward Rendering.

In Forward Rendering, as objects are rendered, shading calculations are performed in a single pass, and there is no intermediate G-buffer to store geometry and material information. The rendering process in Forward Rendering is more immediate, where each object is processed for its final appearance during the same pass.

G-buffer is a key component of the Deferred Rendering approach and serves as an intermediate storage for various data that is later used in separate passes for lighting, shading, and other post-processing effects. Forward Rendering, on the other hand, doesn't involve this level of data separation and uses a simpler, more direct approach for rendering objects.

             In conclusion, both Forward Rendering and Deferred Rendering are valuable tools in the Unity developer's arsenal, each offering specific advantages and trade-offs. By understanding the differences between these rendering paths and considering the FAQ, you can make an informed decision to create games that look great and run smoothly on your chosen platform.