> For the complete documentation index, see [llms.txt](https://docs.convai.com/api-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.convai.com/api-docs/zh/cha-jian-yu-ji-cheng/convai-unity-sdk/features/vision/custom-frame-sources.md).

# 自定义帧源

实现 `IVisionFrameSource` 要发布任何自定义视频管线——视频文件、自定义渲染纹理或屏幕捕获工具——而无需修改发布层。一旦你的组件放入场景中， `ConvaiVisionPublisher` 就会自动发现并进行流式传输。

### 接口契约

`IVisionFrameSource` 是通过……进行视频流传输所需的最小契约 `ConvaiVisionPublisher` 以及通过……进行实时画面显示 `VisionDebugPreview`.

```csharp
public interface IVisionFrameSource
{
    bool IsCapturing { get; }
    long FrameCount { get; }
    (int Width, int Height) FrameDimensions { get; }
    float TargetFrameRate { get; }
    string SourceId { get; }
    RenderTexture CurrentRenderTexture { get; }
    bool IsFrameReady { get; }
    event Action FrameReady;
    void StartCapture();
    void StopCapture();
}
```

你的实现必须是一个 `MonoBehaviour`. `ConvaiVisionPublisher` 使用 `GetComponent` 和 `GetComponentsInChildren`来发现帧源，而这只适用于 Unity 组件。

### Y 轴翻转要求

`CurrentRenderTexture` 必须采用 **自上而下方向** （Y 轴相对于 Unity 默认的自下而上进行了翻转）。LiveKit 和标准视频格式都期望图像顶部的 Y=0。跳过此步骤会导致接收端看到上下颠倒的视频流。

在以下写入过程中应用翻转： `Graphics.Blit` 从源纹理写入输出 `RenderTexture`:

```csharp
// sourceTexture：你的原始 Unity RenderTexture（自下而上）
// _outputRt：你通过 CurrentRenderTexture 暴露的 RenderTexture（自上而下）
Graphics.Blit(sourceTexture, _outputRt, new Vector2(1f, -1f), new Vector2(0f, 1f));
```

该 `scale.y = -1` 和 `offset.y = 1` 这两个参数组合起来会翻转垂直轴。将 `_outputRt` 设置为 `CurrentRenderTexture`.

{% hint style="warning" %}
v4.4.0 移除了在发布期间运行的 LiveKit 纹理回读中的一个内部双重垂直翻转。在 v4.4.0 之前，自定义 `IVisionFrameSource` 还需要额外补偿一次翻转，除了上面所示的翻转之外，以抵消该内部翻转并得到正确的自上而下图像。现在请移除任何此类额外翻转。上面所示的单次翻转是 v4.4.0 之后应用的唯一方向修正。保留额外翻转会重新引入上下颠倒的视频流。请参见 [发布策略](/api-docs/zh/cha-jian-yu-ji-cheng/convai-unity-sdk/features/vision/publishing-and-policies.md) 完整迁移说明。
{% endhint %}

### 最小实现

以下骨架实现了所有必需成员，并正确处理 `FrameReady` 。请填入 `CaptureFrame` 方法中的实际捕获逻辑。

```csharp
using System;
using Convai.Runtime.Vision.Sources;
using UnityEngine;

public class MyCustomFrameSource : MonoBehaviour, IVisionFrameSource
{
    [SerializeField] private int _width = 1280;
    [SerializeField] private int _height = 720;
    [SerializeField] private float _targetFps = 15f;
    [SerializeField] private string _sourceId = "custom";

    private RenderTexture _outputRt;
    private long _frameCount;
    private float _captureInterval;
    private float _nextCaptureTime;

    // IVisionFrameSource

    public bool IsCapturing { get; private set; }
    public long FrameCount => _frameCount;
    public (int Width, int Height) FrameDimensions => IsCapturing ? (_width, _height) : (0, 0);
    public float TargetFrameRate => _targetFps;
    public string SourceId => _sourceId;
    public RenderTexture CurrentRenderTexture => _outputRt;
    public bool IsFrameReady => _frameCount > 0;
    public event Action FrameReady;

    public void StartCapture()
    {
        if (IsCapturing) return;

        _outputRt = new RenderTexture(_width, _height, 24, RenderTextureFormat.ARGB32)
        {
            name = $"CustomFrameSource_{_sourceId}"
        };
        _outputRt.Create();

        _frameCount = 0;
        _captureInterval = _targetFps > 0f ? 1f / _targetFps : 1f / 15f;
        _nextCaptureTime = Time.realtimeSinceStartup;
        IsCapturing = true;
    }

    public void StopCapture()
    {
        if (!IsCapturing) return;

        IsCapturing = false;

        if (_outputRt != null)
        {
            _outputRt.Release();
            Destroy(_outputRt);
            _outputRt = null;
        }
    }

    private void Update()
    {
        if (!IsCapturing) return;

        float now = Time.realtimeSinceStartup;
        if (now < _nextCaptureTime) return;

        _nextCaptureTime = now + _captureInterval;
        CaptureFrame();
    }

    private void OnDestroy() => StopCapture();

    private void CaptureFrame()
    {
        // 替换为你的实际源纹理
        RenderTexture sourceTexture = GetYourSourceTexture();
        if (sourceTexture == null) return;

        // 将 Y 轴翻转到输出 RenderTexture 中
        Graphics.Blit(sourceTexture, _outputRt, new Vector2(1f, -1f), new Vector2(0f, 1f));

        _frameCount++;
        FrameReady?.Invoke();
    }

    private RenderTexture GetYourSourceTexture()
    {
        // 从你的自定义管线返回 RenderTexture
        return null;
    }
}
```

`FrameReady` 必须在 **Unity 主线程上**. `ConvaiVisionPublisher` 和 `VisionDebugPreview` 两者都假定所有 `IVisionFrameSource` 回调都在主线程上执行。如果你的捕获逻辑运行在后台线程上，请使用一个在 `Update`中检查的标志将事件触发切回主线程，如上面的骨架所示。

### 公开生命周期状态

实现 `IVisionFrameSourceStatusProvider` 连同 `IVisionFrameSource` 一起，以公开更丰富的生命周期状态——权限流程、延迟初始化或结构化错误信息。然后发布方就可以在无需轮询的情况下响应就绪状态变化。

```csharp
public class MyCustomFrameSource : MonoBehaviour, IVisionFrameSource, IVisionFrameSourceStatusProvider
{
    // --- IVisionFrameSourceStatusProvider ---

    public VisionSourceState State { get; private set; } = VisionSourceState.Idle;
    public VisionSourceErrorKind ErrorKind { get; private set; } = VisionSourceErrorKind.None;
    public string StatusMessage { get; private set; } = string.Empty;
    public bool HasUsableFrame => FrameCount > 0;
    public event Action StatusChanged;

    private void SetState(VisionSourceState state, VisionSourceErrorKind error = VisionSourceErrorKind.None, string message = "")
    {
        State = state;
        ErrorKind = error;
        StatusMessage = message;
        StatusChanged?.Invoke();
    }

    public void StartCapture()
    {
        SetState(VisionSourceState.Starting);
        // ... 初始化捕获 ...
        SetState(VisionSourceState.Ready);
    }

    public void StopCapture()
    {
        // ... 释放资源 ...
        SetState(VisionSourceState.Stopped);
    }

    // ... IVisionFrameSource 的其余实现
}
```

### 自动发现

一旦你的组件放入场景中， `ConvaiVisionPublisher` 会按以下顺序自动发现它：

1. 该 **源** Inspector 中的字段（显式赋值）。
2. `GetComponent<CameraVisionFrameSource>()` 在同一个 GameObject 上（内置优先级）。
3. `GetComponentsInChildren<MonoBehaviour>(true)` ——首先 `IVisionFrameSource` 在同一个 GameObject 或其子对象中找到的。

如果在步骤 3 下找到多个帧源，发布方会记录警告并选择第一个。请显式分配 **源** 字段，以避免歧义。

### 下一步

{% content-ref url="/pages/885a0e6676256e7738207107695532701ef9b1ec" %}
[Vision 调试预览](/api-docs/zh/cha-jian-yu-ji-cheng/convai-unity-sdk/features/vision/debug-preview.md)
{% endcontent-ref %}

{% content-ref url="/pages/9ee175b12da238182d707abd778abdc3a8706c80" %}
[Vision 脚本 API](/api-docs/zh/cha-jian-yu-ji-cheng/convai-unity-sdk/features/vision/scripting-api.md)
{% endcontent-ref %}


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.convai.com/api-docs/zh/cha-jian-yu-ji-cheng/convai-unity-sdk/features/vision/custom-frame-sources.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.