Video Encoding for the Web in 2026: Part 2: Methodology and Tooling
This is Part 2; see Part 1 for an introduction. Methodology My big goal here is to be able to repeatably measure how well various video encoder settings work against the types of videos that I care about. In 2025, I took a single 1080p test video1Included as deception1, below. and ran around a thousand tests against it to see how well various H.265 encoding options worked. This time, I wanted to test against multiple videos and multiple encoders to get a much wider view of encoding performance. Before I started this, I wanted to improve my tooling so I wasn’t wasting time or measuring the wrong thing.


