🚀 超低延迟大数据量多媒体流传输 🚀 Ultra-Low Latency High-Volume Multimedia Streaming
受够了本地NDI的延迟？HyperFlow-SHM使用GStreamer共享内存技术，在进程间传输无压缩视频，几乎实现零延迟。就像复制文件一样，但适用于实时视频。 Tired of NDI lag on localhost? HyperFlow-SHM uses GStreamer Shared Memory to transmit uncompressed video between processes with virtually ZERO latency. Like copying a file, but for live video.
作为感知引擎的媒体传输单元，支持多种设备间的媒体流传输 As a media transmission unit for perceptual engines, supporting media streaming between multiple devices
VPE (Virtura Perceptual Engine) 是一套针对空间计算时代的工程与设计规范。HyperFlow-SHM作为VPE感知引擎的核心组件，是沉浸式信号系统 (Immersive Signal System) 的重要组成部分，旨在指导团队从“艺术创作”迈向“工业级生产”。
VPE (Virtura Perceptual Engine) is a set of engineering and design specifications for the spatial computing era. HyperFlow-SHM, as a core component of the VPE perceptual engine, is an important part of the Immersive Signal System, guiding teams from "artistic creation" to "industrial-grade production."
我们不再设计界面 (Interface)，我们设计本能 (Instinct)。当屏幕消失，世界向你敞开，文本不再是用来阅读的，而是用于生成世界的 DNA。
We no longer design interfaces; we design instincts. When screens disappear and the world opens up to you, text is no longer for reading, but for generating the DNA of the world.
- 🚀 零延迟：仅受内存读写速度限制
- 🚀 Zero Latency: Only the speed of memory reads/writes
- 💾 原始数据：支持I420/RGBA无损传输
- 💾 Raw Data: Supports I420/RGBA lossless transmission
- 🐍 Python实现：纯Python编写，易于修改
- 🐍 Pythonic: Pure Python implementation, easy to modify
- 🎨 GUI支持：基于PyQt5的接收器，带有嵌入式视频窗口
- 🎨 GUI Support: PyQt5-based receiver with embedded video window
- 🔌 多视频源：支持测试图案、摄像头和桌面捕获
- 🔌 Multiple Sources: Support for test patterns, cameras, and desktop capture
- 🔄 自动重连：管道故障时自动重连
- 🔄 Auto Reconnect: Automatic reconnection on pipeline failure
- 🔊 音频传输支持（即将推出）
- 🔊 Audio Transmission Support (Coming Soon)
- 🌐 跨设备串流支持
- 🌐 Cross-Device Streaming Support
- 🤖 实时AI信号传输潜力
- 🤖 Real-time AI Signal Transmission Potential
重要提示：使用MSVC版本，不要使用MinGW！ IMPORTANT: Use MSVC version, not MinGW!
- 运行时包：gstreamer-1.0-msvc-x86_64-1.22.7.msi
- Runtime Package: gstreamer-1.0-msvc-x86_64-1.22.7.msi
- 开发包：gstreamer-1.0-devel-msvc-x86_64-1.22.7.msi
- Development Package: gstreamer-1.0-devel-msvc-x86_64-1.22.7.msi
安装过程中： During installation:
- 选择"Complete"安装类型
- Choose "Complete" installation type
- 确保勾选"Add GStreamer to the PATH environment variable"
- Ensure "Add GStreamer to the PATH environment variable" is checked
pip install -r requirements.txtpython scripts/env_doctor.py这将验证你的GStreamer安装和Python依赖。 This will verify your GStreamer installation and Python dependencies.
python src/sender.pypython src/sender.py --source camerapython src/sender.py --source desktoppython src/receiver.py在接收器窗口中： In the receiver window:
- 输入套接字路径（默认：
/tmp/videodata） - Enter the socket path (default:
/tmp/videodata) - 点击"Connect"开始接收视频
- Click "Connect" to start receiving video
python src/sender.py [OPTIONS]| 选项 | 描述 | 默认值 |
|---|---|---|
--socket-path |
共享内存套接字路径 | /tmp/videodata |
--source |
视频源类型（test/camera/desktop） | test |
--width |
视频宽度 | 1920 |
--height |
视频高度 | 1080 |
--framerate |
视频帧率 | 30 |
| Option | Description | Default |
| -------- | ------------- | --------- |
--socket-path |
Shared memory socket path | /tmp/videodata |
--source |
Video source type (test/camera/desktop) | test |
--width |
Video width | 1920 |
--height |
Video height | 1080 |
--framerate |
Video framerate | 30 |
HyperFlow-SHM/
├── src/
│ ├── sender.py # 视频发送端（无头）
│ ├── receiver.py # GUI视频接收器
│ ├── pipeline.py # GStreamer管道封装
│ └── utils.py # 工具函数
├── scripts/
│ └── env_doctor.py # 环境诊断工具
├── requirements.txt # Python依赖
├── README.md # 本文档
├── ARCH_DESIGN.md # 架构设计文档
└── AUDIO_SUPPORT.md # 音频支持说明
HyperFlow-SHM/
├── src/
│ ├── sender.py # Video sender (Headless)
│ ├── receiver.py # GUI video receiver
│ ├── pipeline.py # GStreamer pipeline wrapper
│ └── utils.py # Utility functions
├── scripts/
│ └── env_doctor.py # Environment diagnostic tool
├── requirements.txt # Python dependencies
├── README.md # This file
├── ARCH_DESIGN.md # Architecture design document
└── AUDIO_SUPPORT.md # Audio support documentation
| 特性 | HyperFlow-SHM | NDI |
|---|---|---|
| 延迟 | ~1ms | ~50-100ms |
| CPU使用率 | 低 | 高 |
| 带宽 | 无 | 高 |
| 压缩 | 无（无损） | 有损/无损 |
| 网络依赖 | 否 | 是 |
| Feature | HyperFlow-SHM | NDI |
| --------- | --------------- | ----- |
| Latency | ~1ms | ~50-100ms |
| CPU Usage | Low | High |
| Bandwidth | None | High |
| Compression | None (Lossless) | Lossy/Lossless |
| Network Dependent | No | Yes |
Source -> Caps -> ShmSink
ShmSrc -> Caps -> VideoConvert -> AutoVideoSink
- VJ表演：应用间零延迟视频路由
- VJ Performances: Zero-latency video routing between applications
- 虚拟制作：绿幕工作流的高速视频传输
- Virtual Production: High-speed video transport for green screen workflows
- 交互式媒体艺术：实时视频处理管道
- Interactive Media Art: Real-time video processing pipelines
- 游戏直播：低延迟捕获和显示
- Game Streaming: Low-latency capture and display
- 舞台表演：多设备间实时视频流传输
- Stage Performances: Real-time video streaming between multiple devices
- 音乐现场：VJ与音乐同步的低延迟视频传输
- Music Live: Low-latency video streaming synchronized with music
- 沉浸式大空间：多投影、多设备协同
- Immersive Large Spaces: Multi-projection, multi-device collaboration
- 交互装置：实时视频流反馈
- Interactive Installations: Real-time video stream feedback
- 头盔到屏幕：VR头盔到显示器的实时传输
- Helmet to Screen: Real-time transmission from VR helmet to display
- 屏幕到头盔：显示器到VR头盔的实时传输
- Screen to Helmet: Real-time transmission from display to VR helmet
- 多设备串流：多设备间的双向视频流传输
- Multi-device Streaming: Bidirectional video streaming between multiple devices
- Apple Vision Pro适配：基于GStreamer的跨平台特性
- Apple Vision Pro Adaptation: Based on GStreamer's cross-platform features
HyperFlow-SHM基于GStreamer的跨平台特性，能够在不同操作系统和设备间实现高效的视频流传输。GStreamer作为成熟的多媒体框架，支持Windows、macOS、Linux等多种平台，为跨设备支持提供了坚实的技术基础。
基于macOS/iOS的GStreamer支持，HyperFlow-SHM可以轻松适配Apple Vision Pro，实现从传统设备到空间计算设备的无缝视频流传输，为沉浸式体验提供低延迟的媒体传输解决方案。
HyperFlow-SHM的低延迟架构使其非常适合实时AI信号传输。在空间计算时代，AI生成的实时信号需要快速传输和处理，HyperFlow-SHM的零延迟特性能够满足这一需求，为AI/Agent集成提供可靠的传输基础。
-
找不到GStreamer
-
GStreamer not found
- 确保安装了GStreamer MSVC
- Ensure GStreamer MSVC is installed
- 检查
gstreamer\1.0\msvc_x86_64\bin是否在PATH中 - Check that
gstreamer\1.0\msvc_x86_64\binis in your PATH
-
找不到共享内存插件
-
Shared memory plugin not found
- 安装GStreamer开发包
- Install the GStreamer Development package
- 用
gst-inspect-1.0.exe shm验证 - Verify with:
gst-inspect-1.0.exe shm
-
Python导入错误
-
Python import errors
- 确保安装了PyGObject：
pip install PyGObject - Ensure PyGObject is installed:
pip install PyGObject - 在Windows上，您可能需要安装GTK运行时
- On Windows, you may need to install GTK Runtime
- 确保安装了PyGObject：
- 查看控制台输出获取GStreamer错误信息
- Check the console output for GStreamer error messages
- 使用
env_doctor.py诊断环境问题 - Use
env_doctor.pyto diagnose environment issues
- 音频传输支持
- Audio transmission support
- Spout/Syphon桥接
- Spout/Syphon bridge
- Linux/macOS兼容性
- Linux/macOS compatibility
- 配置文件支持
- Configuration file support
- 多流支持
- Multiple stream support
- 网络传输扩展（跨设备、跨网络）
- Network transmission extension (cross-device, cross-network)
- Apple Vision Pro适配
- Apple Vision Pro adaptation
- VR/AR设备支持
- VR/AR device support
- AI/Agent集成
- AI/Agent integration
欢迎贡献！请随时提交Pull Request。 Contributions are welcome! Please feel free to submit a Pull Request.
MIT License
- GStreamer项目提供了优秀的多媒体框架
- GStreamer Project for the excellent multimedia framework
- PyQt5提供了GUI工具包
- PyQt5 for the GUI toolkit
- 感谢所有贡献者和用户
- All contributors and users
- VPE (Virtura Perceptual Engine) 团队的理念指导
- The conceptual guidance from the VPE (Virtura Perceptual Engine) team
gstreamer video-streaming shared-memory zero-latency python pyqt5 windows multimedia realtime vj virtual-production spatial-computing perceptual-engine ai-signaling apple-vision-pro