Video content is an increasingly popular medium of communication, and more and more people are engaging in video recording and editing. As a result, the amount of video content shared daily on social media and platforms such as YouTube, TikTok, Vimeo, and others has greatly increased. This is made possible by codecs that compress videos and allow them to be viewed on our devices without major issues. Unfortunately, the understanding of codecs in practice is still quite poor, which is why a research of this master thesis was conducted. The aim of the research was to create guidelines that would make it a little easier for video creators to understand codecs and the way in which the various factors and decisions made in the recording process affect the final quality of the video presented.
This study analysed the effect of motion amount, bit depth, bitrate, frames per second, video resolution, single frame compression recording versus multi-frame compression recording, and the effect of noise on the final video image quality. Efficiency of the three most commonly used codecs: MPEG-2, H.264 and H.265, was analysed by objective analysis of video quality and by visual analysis of selected details. For the objective analysis the results of PSNR, SSIM, VMAF and EPSNR matrices were used, and with the obtained results the influence of each parameter on the final video compression quality was determined.
Obtained results clearly showed how each parameter affects the compression with each codec, and from this some guidelines were defined to help video producers to better understand the codecs and how they work.
|