Search code examples
videoffmpeggpuflvtranscoding

Is there any way to speed up video transcoding using Nvidea GPU?


I am looking to transcode about 500gb worth of .mp4 video files to .flv format. I was wondering if there is any way my to gtx 970's could speed up this process. Otherwise i'm stuck with using a single i7 4790k.

Is there any type of NVIDEA gpu conversion program available?

Any suggestions appreciated.


Solution

  • There are multiple applications out there that can take advantage of Nvidia NVENC, a hardware-based Silicon Intellectual Property Core (SIP Block) on your GPU, and here are starters:

    1. ffmpeg: You can build and configure ffmpeg with --enable-nvenc option to turn on Nvidia NVENC support. This will require an installation of both the Nvidia CUDA runtime and the Nvidia NVENC SDK. See this for more instructions:

    (a). This Github gist will build a static ffmpeg binary on Ubuntu: https://gist.github.com/Brainiarc7/3f7695ac2a0905b05c5b

    (b). See additional build instructions: https://github.com/Brainiarc7/ffmpeg_libnvenc/blob/master/README.md

    1. Freeware on Windows: You might want to look into freeware such as StaxRip, MediaCoder and the all-round champion Hybrid. Use Google (I will not add links to their download pages as they'll be outdated over time).

    2. Commercial software: You can use the excellent ConvertXtoVideo by VSO Software. This supports simultaneous video transcode pipelines in hardware , even in cases where you have multiple GPUs (such as Nvidia Optimus systems where you may use both Intel QuickSync and Nvidia's NVENC simultaneously). You may also use ConvertXtoHD for bluray authoring. Supports the same feature set.

    For video production, there's an Nvidia NVENC plugin accelerator for both Adobe Premiere and Adobe Media Encoder (2014 and above). See the thread here for more details:https://forums.adobe.com/thread/1243687

    Thanks and regards,

    Brainiarc7

    Notes: The list is non-exhaustive. You may therefore find other alternatives whose implementations work better than, or are equal to, the options provided in this solution.