Unity3d tutorial notes - unity initial 03

Here is finishing unity The third day of tutorial notes ！ Another day full of harvest ！！！

One sentence per day ： Be a gentle person , Never be humble or arrogant, clear and kind

plug-in unit ：Instant OC

[ link ：https://pan.baidu.com/s/1A8IUPgpk8rGgzE7NnbY0DA?pwd=ufl2

Extraction code ：ufl2（ Free access to bloggers' online disks Instant OC plug-in unit , Don't forget Click the blog post to thank you ）]

03DAY

Catalog

Make a small map of the game scene

Render pipeline ：

Vertex Processing ：

Primitive Assembly ：

Rasterize ：

Pixel processing ：

cache ：

Frame cache ：

The depth of the cache z-buffer：

Occlusion Culling Occlusion culling ：

meaning （ advantage ）：

shortcoming ：

Applicability ：

step ：

LOD

meaning （ advantage ）：

attribute ：

step ：

Applicability ：

shortcoming ：

Three aspects of performance optimization ：

---

Make a small map of the game scene

Make the model of the camera look like an object , Just put them in a unified parent object , Operate on the parent object ——> The camera can have multiple ,Audio Listener（ Ears ） There can only be one

 

（ Error prompt of two cameras ）——> Depth adjustment value ——> Make a plane on the character , Used as a small map logo ,Main Camera Don't want to see the plane ,MipCamera Don't want to see people ： New layer , use Culling Mask hide

                                  

 

          

 （ During actual development ： The map camera doesn't let you see the terrain , Rendering costs a lot , The artist prepares a picture and puts it on the ground ,camera Just look at the character logo , Performance optimization ）——> Remove the blank part of the small map ：MipCamera—>Clear Flags—>Depth Only（ Do not render blank space ）

Render pipeline ：

Graphic data in GPU After arithmetic processing , The process of finally outputting to the screen

CPU： a central processor

GPU： The graphics card , Make a rendering of

 

Draw Call（unity5 It's called Batches）: Every time the engine prepares data and notifies GPU The process of . Popular speaking , The number of times the graphics card is called to render objects per frame

（Batches The bigger it is , Render more , To reduce Batches）

 

Vertex Processing ：

Accept （GPU Send it ） Model vertex data ; Coordinate transformation

Primitive Assembly ：

Assembly surface ： Connect adjacent vertices , Draw triangular faces

 

Rasterize ：

Calculate the pixels on the triangular plane , And provide reasonable interpolation parameters for the later coloring stage

 

Pixel processing ：

Shade each pixel region , Write to cache

cache ：

A storage pixel data Memory block , The most important caches are frame cache and depth cache

Frame cache ：

Store the of each pixel color , namely Rendered image . The frame buffer is often in the graphics card , The graphics card constantly reads and outputs to the screen

The depth of the cache z-buffer：

Storing pixels In depth information , namely The distance from the object to the camera .

When rasterizing, the depth value of each pixel is calculated , If the new depth value is more than the existing depth value near , be Pixel color Written to Frame cache , And replace the depth cache .

Occlusion Culling Occlusion culling ：

meaning （ advantage ）：

Before the object is sent to the rendering pipeline , Remove objects that cannot be seen in the camera's perspective , This reduces the amount of rendering data per frame , Improve rendering performance （ advantage ）

plug-in unit ：Instant Occlusion Culling

[ link ：https://pan.baidu.com/s/1A8IUPgpk8rGgzE7NnbY0DA?pwd=ufl2

Extraction code ：ufl2（ Free access to bloggers' online disks Instant OC plug-in unit , Please click the blog post to thank you ）]

 

shortcoming ：

Judge whether the object is covered ,CPU Using extra overhead

Applicability ：

In the scene , Objects are densely distributed and there are a large number of blocked objects

step ：

1. Create layers

2. Make a layer for the game （ Remove the participating occlusion ） And labels （ Will automatically attach IOClod Script ）

3. Add Collider to the object Collider Components

4. Camera additional script IOCcam.

LOD

meaning （ advantage ）：

Multiple levels of detail （Level Of Detail）

LOD Technology refers to the importance of the position of the nodes of the object model in the display environment , Determine the resource allocation for object rendering , Reduce the number of faces and detail of unimportant objects , Thus, efficient rendering operations may be obtained

It reduces the distance between the camera and the model to select high 、 in 、 Low model （Lod_0 Fine mold 、Lod_1 Medium model 、Lod_2 Simple model ）

attribute ：

lod 1distance： When the distance between the camera and the object is less than the current distance , Use Lod_0 Model

lod 2distance： The distance between the camera and the object is greater than Lod 1 And less than the current distance , Use Lod_1 Model , Greater than the current value , Use Lod_2 Model

step ：

1. Create layers
2. Create an empty object and put the model （Loa_0、Loa_1、Loa_2） Add to
3. Assign layers and labels to parent objects
4. Add collider components to parent or child objects
5. . Camera additional script IOCcam

Applicability ：

Scene , There are high modes and the high modes will vary far and near

shortcoming ：

CPU Consume 、 Take up memory （ The original one , Now three ）、（ Art works need to make three models ）

 

Three aspects of performance optimization ：

CPU Use 、 Rendering 、 Memory usage

 （window——>Profiler）