vmaf 安装和使用

https://streaminglearningcenter.com/blogs/installing-and-using-netflix-vmaf-master.html

Prepping the Files

Note that vmafossexec.exe only works with YUV inputs so you’ll have to convert your reference and distorted files to YUV format. FFmpeg is the tool of choice and you can modify this command string for your conversion.

ffmpeg -y -i distorted.mp4 -pix_fmt yuv420p -vsync 0 distorted.yuv
ffmpeg -y -i reference.mp4 -pix_fmt yuv420p -vsync 0 reference.yuv

Obviously, you should substitute the actual extension for MP4 if necessary, and if FFmpeg doesn’t decode the file find a different decoder to do so. Note the pixel format used since you’ll input that into your VMAF-Master command string.

Here’s the command string I used during my recent project. Note that both distorted and reference files are in the same folder which also contains the batch file so there are no paths to either file. Here I’m calling the default model, computing VMAF, PSNR, SSIM, and MS SSIM, and storing the results in a file named distorted.csv, an XML-formatted file with a .csv extension.

C:\vmaf-master\x64\Release>vmafossexec yuv420p 1920 820 01-4.0m.yuv 01-2.0m.yuv C:\vmaf-master\model\vmaf_v0.6.1.pkl --psnr --ssim --ms-ssim
Start calculating VMAF score…
Exec FPS: 8.178312
VMAF score = 39.276684
PSNR score = 21.952593
SSIM score = 0.673477
MS-SSIM score = 0.703873

C:\vmaf-master\x64\Release>vmafossexec yuv420p 1920 820 01-4.0m.yuv 01-1.5m.yuv C:\vmaf-master\model\vmaf_v0.6.1.pkl --psnr --ssim --ms-ssim
Start calculating VMAF score…
Exec FPS: 8.109136
VMAF score = 37.880867
PSNR score = 21.920033
SSIM score = 0.674751
MS-SSIM score = 0.704895

C:\vmaf-master\x64\Release>vmafossexec yuv420p 1920 820 01-4.0m.yuv 01-1.0m.yuv C:\vmaf-master\model\vmaf_v0.6.1.pkl --psnr --ssim --ms-ssim
Start calculating VMAF score…
Exec FPS: 7.393184
VMAF score = 34.884087
PSNR score = 21.853622
SSIM score = 0.674992
MS-SSIM score = 0.705316

  • 0
    点赞
  • 0
    收藏
    觉得还不错? 一键收藏
  • 1
    评论

“相关推荐”对你有帮助么?

  • 非常没帮助
  • 没帮助
  • 一般
  • 有帮助
  • 非常有帮助
提交
评论 1
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值