![]() That number is increasing rapidly over the last few months as people can’t go out for entertainment during the pandemic. 4K Ultra High Definition (4K UHD) is now a standard feature in over 60% of the televisions sold in North America. In that article, I focusing upscaling 720×480 (480i) into 1920×1080 (1080p) footage.įor one thing, video standards have changed in just a few short years. I originally wrote this article about 7 years ago, when a cinematographer friend of mine asked me to upscale footage for his reel from the 1990s to HD. When I started my career in video post-production, “standard definition” was, in fact, the standard. You don’t have a camera that can shoot 4K and want your video to be 4K (or other sizes).You work with old video footage that’s in the standard format.When this command is executed, FFmpeg computes the VMAF score and sends the score to the console on the last line of its output: VMAF score: 92.419584Įt voilà! For this example the score would then be recorded as “92.419584”.Why would you want to upscale your footage? We therefore made a local build of FFmpeg 4.2.1 with support for libvmaf.a (version 0.6.1, VDK version 1.3.15) as described in VMAF's GitHub repo using master branch code.įor computing the VMAF score of a reconstructed video we then did: ffmpeg -nostats -i reference.mov -i -i mask.png -lavfi "split overlay overlay libvmaf" -f null. The precompiled builds of FFmpeg I have come across do not have VMAF support enabled. Commercially available tools like MSU Quality Measurement Tool computes VMAF, as well, but it only runs natively under Windows and on Linux via Wine - not optimal for those of us who are Mac users. Thus, a VMAF score of 70 can be interpreted as a vote between “good” and “fair” by an average viewer under the 1080p and 3H condition.Ī note to the reader here: a viewing distance of 3H means “3 times the height of the display”.Ī fairly straight-forward way of computing the VMAF is using a FFmpeg video filter. Viewers voted the video quality on the scale of “bad”, “poor”, “fair”, “good” and “excellent”, and roughly speaking, “bad” is mapped to the VMAF scale 20 and “excellent” to 100. As an example, the default model v0.6.1 is trained using scores collected by the Absolute Category Rating (ACR) methodology using a 1080p display with viewing distance of 3H. A good way to think about a VMAF score is to linearly map it to the human opinion scale under which condition the training scores are obtained. VMAF scores range from 0 to 100, with 0 indicating the lowest quality, and 100 the highest. Quoting the authors themselves from their Medium announcement post, VMAF scores are interpreted as: It is intended to replace classic scientific metrics like PSNR and SSIM which do not accurately capturing human perception. VMAF is short for Video Multi-method Assessment Fusion which is a modern video quality metric developed by Netflix that combines human vision modeling with machine learning. We decided to use VMAF as the metric for comparing algorithms. Significant consideration and experimentation was put into aligning the settings as much as possible. Some products offer more flexibility than others. ![]() Where applicable screenshots of the settings used are included below. It’s outside the scope of this post to dive deeply into the technical details of each. Each algorithm obviously has its own strengths and weaknesses in areas such as visual quality, speed, ease of use, and flexibility from a UI standpoint. Both of these were obviously included.ĭue to the proprietary nature of many of the algorithms involved, it’s impossible to make any predictions about their complexity and sophistication. and Adobe After Effect’s Detail-Preserving Upscale. For instance, we have heard the term “industry standard in the broadcast industry” used for both Grass Valley’s Alchemist Ph.C. ![]() ![]() ![]() Initially we performed some research and identified the group of candidates to pit against our algorithm. We had to limit the scope of our testing and be smart in terms of selecting video footage and competing algorithms. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |