~$ the project
sherboa engine is a free, independent project created for educational and experimental purposes.
it is designed as a user-friendly tool for comparing video quality with a focus on video engineering, streaming and media systems.
~$ powered by linux & all his friends
HTML · CSS · javascript · python · fastAPI · ffmpeg · docker · github actions · kubernetes · public cloud · terraform
~$ VMAF
the industry's preferred metric, is an open-source perceptual video quality assessment algorithm developed by netflix.
sherboa engine implements VMAF v1.0.16 - standard HD model (1080p/3H).
~$ SSIM
evaluates how visually similar a compressed or processed video is to the original.
instead of comparing pixels individually, it measures similarities in luminance, contrast, and structural details, making it more representative of human visual perception.
~$ PSNR
measures the mathematical difference between an original video and a compressed or processed version.
it is based on pixel-by-pixel error and quantifies how much distortion has been introduced during encoding.
unlike SSIM, which focuses on perceived visual similarity, PSNR focuses on signal accuracy and reconstruction error.
~$ ffmpeg
is a free and open-source multimedia framework for video processing and media analysis.
~$...and all that chess
thousands of chess games were played for this project to come to life!