Christopher Chedeau - Video Editing in the Browser

Discover the future of video editing in the browser with Christopher Chedeau, exploring the challenges and opportunities of decomposing images into sinusoids, compressing video, and leveraging WebAssembly and native code.

Key takeaways

Using JavaScript, it’s possible to decompose an image into a sum of many sinusoids, just like FFT.
In video editing, keyframes are used as reference points.
WebAssembly has a high overhead, so it’s only useful for specific parts of an application.
It’s harder to compress images when they contain many pixels with similar colors.
WebAssembly and native code use different memory management strategies, so it’s hard to bridge the gap between the two.
Implementing compression algorithms in JavaScript requires careful tuning to avoid overflow and performance issues.
The process of decoding video can be slow if the image is too complex.
Decomposing a video into a sum of simple images can allow for more efficient compression.
Compressing images is easier if they contain many repetitive patterns, making it possible to compress them quickly.
JavaScript libraries like FFT.js can be used for complex mathematical transformations.
Web codecs are already being implemented in browsers to enable browser-based video editing.

Christopher Chedeau - Video Editing in the Browser

More talks