编译FFmpeg（支持CUDA加速，支持h264）
2026-06-03
1、检查GPU相关配置
nvidia-smi
Wed Feb 12 16:48:50 2025
+---------------------------------------------------------------------------------------+
| NVIDIA-SMI 535.216.01 Driver Version: 535.216.01 CUDA Version: 12.2 |
|-----------------------------------------+----------------------+----------------------+
| GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|=========================================+======================+======================|
| 0 NVIDIA GeForce GTX 1650 Off | 00000000:01:00.0 Off | N/A |
| 46% 51C P2 N/A / 75W | 369MiB / 4096MiB | 36% Default |
| | | N/A |
+-----------------------------------------+----------------------+----------------------+
+---------------------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=======================================================================================|
| 0 N/A N/A 2473 G /usr/lib/xorg/Xorg 49MiB |
+---------------------------------------------------------------------------------------+
2、依赖安装编译
1、编译安装 ffnvcodec
➜ ~ git clone https://git.videolan.org/git/ffmpeg/nv-codec-headers.git
➜ ~ cd nv-codec-headers
➜ ~ sudo make install
sed 's#@@PREFIX@@#/usr/local#' ffnvcodec.pc.in > ffnvcodec.pc
install -m 0755 -d '/usr/local/include/ffnvcodec'
install -m 0644 include/ffnvcodec/*.h '/usr/local/include/ffnvcodec'
install -m 0755 -d '/usr/local/lib/pkgconfig'
install -m 0644 ffnvcodec.pc '/usr/local/lib/pkgconfig'
➜ ~ export PKG_CONFIG_PATH="/usr/local/lib/pkgconfig"
➜ ~ pkg-config --modversion ffnvcodec
12.0.16.1
2、配置编译安装FFmpeg
sudo ./configure --prefix=/usr/local/ffmpeg --enable-shared --disable-static --disable-doc --enable-gpl --enable-libx264 --enable-cuda --enable-cuvid --enable-network--enable-nvenc
sudo make -j # 有多少物理核心写多少,可以加快编译速度
sudo make install
# 系统配置
sudo vim ~/.bashrc
#在末尾,添加如下内容:
export LD_LIBRARY_PATH=/usr/local/ffmpeg/lib:$LD_LIBRARY_PATH
#保存,关闭文件,运行命使配置生效:
source ~/.bashrc
#生成静态链接文件,如果不执行此步骤,可能系统检测不到安装的ffmpeg软件,因此需要运行如下命令:
cd /usr/bin
sudo ln -s /home/XXX/ffmpeg-4.1/ffprobe ffprobe
sudo ln -s /home/XXX/ffmpeg-4.1/ffmpeg ffmpeg
sudo ln -s /home/XXX/ffmpeg-4.1/ffplay ffplay
# 检查安装
ffmpeg -hwaccels
# 以下为安装成功的显示结果
ffmpeg version 7.0.2 Copyright (c) 2000-2024 the FFmpeg developers
built with gcc 11 (Ubuntu 11.4.0-9ubuntu1)
configuration: --prefix=/usr/local/ffmpeg --enable-shared --disable-static --disable-doc --enable-gpl --enable-libx264 --enable-cuda --enable-cuvid --enable-nvenc
libavutil 59. 8.100 / 59. 8.100
libavcodec 61. 3.100 / 61. 3.100
libavformat 61. 1.100 / 61. 1.100
libavdevice 61. 1.100 / 61. 1.100
libavfilter 10. 1.100 / 10. 1.100
libswscale 8. 1.100 / 8. 1.100
libswresample 5. 1.100 / 5. 1.100
libpostproc 58. 1.100 / 58. 1.100
Hardware acceleration methods:
cuda
3、实际实现GPU硬解码代码
// 初始化和配置一个 H.264 视频编码器,并将其与 RTSP 输出流关联
int PushOpencv::open_codec(int width, int height, int den) {
int ret = 0;
avformat_network_init();
// const AVCodec *codec = avcodec_find_encoder(AV_CODEC_ID_H264); // CPU初始编码器
const AVCodec *codec = avcodec_find_encoder_by_name("h264_nvenc"); // GPU硬件加速编码器
if (!codec)
{
throw std::logic_error("Can`t find h264 encoder!"); // 找不到264编码器
}
// b 创建编码器上下文
outputVc = avcodec_alloc_context3(codec);
if (!outputVc)
{
throw std::logic_error("avcodec_alloc_context3 failed!"); // 创建编码器失败
}
// c 配置编码器参数
outputVc->flags |= AV_CODEC_FLAG_GLOBAL_HEADER; // 全局参数
outputVc->codec_id = codec->id;
outputVc->codec_type = AVMEDIA_TYPE_VIDEO;
outputVc->thread_count = 4;
outputVc->bit_rate = 1024 * 1024 * 8; // 压缩后每秒视频的bit位大小为50kb
outputVc->width = width;
outputVc->height = height;
outputVc->time_base = {1, den};
outputVc->framerate = {den, 1};
outputVc->gop_size = 5;
outputVc->max_b_frames = 0;
outputVc->qmax = 30;
outputVc->qmin = 10;
outputVc->pix_fmt = AV_PIX_FMT_YUV420P;
// // CPU进行编码配置
// av_opt_set(outputVc->priv_data, "preset", "ultrafast", 0); //preset 决定了编码器的速度与质量权衡,superfast 是一种快速模式。
// av_opt_set(outputVc->priv_data, "tune", "zerolatency", 0); //tune 的 zerolatency 配置适合实时流媒体应用,可减少编码延迟。
// av_opt_set(outputVc->priv_data, "profile", "main", 0); // Main profile 兼容性更好
// av_opt_set(outputVc->priv_data, "nal-hrd", "none", 0); // 关闭 HRD,减少延迟
// // profile 是 H.264 的一种简化模式,适合低延迟、低复杂度的编码场景::防止转换成webrtc播放卡顿
// av_opt_set(outputVc->priv_data, "profile", "baseline", 0); // 设置 baseline profile
//GPU硬解码详细配置参数
av_opt_set(outputVc->priv_data, "preset", "p3", 0); // 低延迟高质量
av_opt_set(outputVc->priv_data, "tune", "ll", 0);
av_opt_set(outputVc->priv_data, "profile", "main", 0);
av_opt_set(outputVc->priv_data, "rc", "vbr", 0); //cbr (Constant Bitrate): 恒定比特率模式。编码器将尽量保持视频流的比特率不变,适用于带宽稳定的场景。vbr (Variable Bitrate): 可变比特率模式。编码器根据视频内容的复杂度调整比特率,适用于不太受带宽限制的场景,能够在保持相同视频质量的同时节省带宽。
// av_opt_set(outputVc->priv_data, "bitrate", "4000000", 0); // 4Mbps
av_opt_set(outputVc->priv_data, "maxrate", "4000000", 0);
av_opt_set(outputVc->priv_data, "bufsize", "8000000", 0);
av_opt_set(outputVc->priv_data, "gop", "30", 0);
av_opt_set(outputVc->priv_data, "bframes", "0", 0);
av_opt_set(outputVc->priv_data, "spatial-aq", "1", 0);
av_opt_set(outputVc->priv_data, "temporal-aq", "1", 0);
av_opt_set(outputVc->priv_data, "zerolatency", "1", 0);
// d 打开编码器上下文
ret = avcodec_open2(outputVc, codec, 0);
std::cout << "avcodec_open2 success!" << std::endl;
ret = avformat_alloc_output_context2(&output, nullptr, "rtsp", url.c_str());
vs = avformat_new_stream(output, outputVc->codec);
vs->codecpar->codec_tag = 0;
// 从编码器复制参数
avcodec_parameters_from_context(vs->codecpar, outputVc);
av_dump_format(output, 0, url.c_str(), 1);
ret = avio_open(&output->pb, url.c_str(), AVIO_FLAG_WRITE);
return ret;
}
4、使用报错记录
- ffmpeg硬编码踩坑Driver does not support the required nvenc API version. Required: 12.2 Found: 12.0
如何处理在使用ffmpeg时因nvidia驱动与nv-codec-headers版本不匹配引发的问题,通过下载对应驱动版本的nv-codec-headers、编译安装并重新编译ffmpeg来解决此问题,确保硬件加速功能正常工作。
1、在官网上下载一个符合现在nvidia驱动版本的nv-codec-headers版本:
2、重新编译安装nv-codec-headers以及FFmpeg
5、参考链接:
- https://blog.csdn.net/JineD/article/details/128420330
- https://blog.csdn.net/qq_43513908/article/details/138161139
- https://blog.yearnfar.com/posts/编程/linux/nvidia-gpu加速ffmpeg/
- https://blog.csdn.net/whf_139/article/details/135515346
- https://blog.csdn.net/as812252319/article/details/115258830
- https://blog.csdn.net/qq_36397240/article/details/107745149