nyanmisaka/ffmpeg-rockchip

FFmpeg with async and zero-copy Rockchip MPP & RGA support

What it solves

This project enables full hardware-accelerated video processing on Rockchip-based devices (such as those using the RK3588 chip). It allows the standard FFmpeg command-line tool to offload heavy video decoding, encoding, and filtering tasks to the device's specialized hardware, rather than relying on the main processor.

How it works

It integrates Rockchip’s proprietary Media Process Platform (MPP) and 2D Raster Graphic Acceleration (RGA) drivers directly into FFmpeg. By using these hardware interfaces, the software achieves "zero-copy" performance, meaning video data is moved between the decoder, filters, and encoder without unnecessary and slow memory copying.

Who it’s for

This is designed for developers and power users working with Rockchip-based hardware who need high-performance video transcoding, such as 8K streaming or complex video manipulation, while keeping CPU usage low.

Highlights

  • Supports hardware decoding for high-resolution formats including 8K 10-bit H.264, HEVC, VP9, and AV1.
  • Provides hardware-accelerated encoding for H.264, HEVC, and MJPEG.
  • Includes RGA-based filters for efficient scaling, cropping, transposing, and blending of images.
  • Enables advanced features like de-interlacing and frame-parallel (async) encoding.
  • Utilizes zero-copy DMA (Direct Memory Access) to maximize performance and efficiency.

Related

  • Project
    sirioberati/Genjustsu-Open-Source-Workflow

    A local UI and agent-assisted workflow for replacing video subjects using depth maps, SAM 3 masks, and Seedance, featuring a full restoration of original source audio.

  • Project
    Francis-Rings/StableAvatar

    StableAvatar is an open‑source diffusion‑based video transformer that generates arbitrarily long, high‑quality avatar videos driven by audio. It introduces a time‑step‑aware audio adapter, audio native guidance, and a sliding‑window fusion strategy to keep lip‑sync and identity stable over unlimited duration. The repo provides pretrained 1.3 B checkpoints, inference scripts (single‑GPU, multi‑GPU, Gradio, ComfyUI), and full training/LoRA code, with detailed data‑preparation instructions.

  • Project
    NikoDemon80/ComfyUI-H3-Motion-Context

    Adds custom nodes to ComfyUI that let you chain MiniMax‑H3 video‑plus‑audio generations. The previous clip is passed as a latent (no decode‑encode), preserving exact pixel data and continuing the audio waveform. Includes Save/Load Latent, Trim, Chain, and optional drift‑reduction nodes, enabling seamless multi‑clip videos with consistent motion and sound.

  • Project
    pireel/pireel

    Pireel Studio is an open-source AI video editor that lets users create videos using a combination of manual timeline editing, chat-based descriptions, and AI agents.