Behind Google's Fused Video Stabilization, combining optical image stabilization and electronic image stabilization, which it uses on the Pixel 2 and Pixel 2 XL
Context & Ripple Effects
Google's video-stabilization work has been building across its stack: reviewers found the original Pixel's stabilization 'works very well' but flagged quirks like choppy focal length adjustment, and earlier in 2017 Google shipped stabilization at the app level via Google Photos v2.13. The Research Blog post now explains the Pixel 2 generation's answer: fusing optical image stabilization with electronic image stabilization rather than relying on either alone.
The post matters because it documents Google treating video as a software-plus-hardware problem years before that approach matured into the Pixel 8's Video Boost, which pushes video processing into Google's datacenters entirely.
First-order effects
- Pixel 2 and Pixel 2 XL owners get steadier handheld video directly from the camera, since the fusion lets OIS handle high-frequency shake while EIS compensates for the residual motion OIS can't absorb.
Second-order effects
- Rival phone makers face pressure to pair their own optics with motion-modeling software, because a fused pipeline raises the baseline for what counts as acceptable handheld footage.
Third-order effects
- If the pattern holds, video quality increasingly decouples from sensor and lens size and follows compute — the same logic that later carried Pixel video into cloud-based processing with Video Boost.
The trend: Smartphone video is shifting from a hardware-specified feature to a computational pipeline, with each Pixel generation moving more of the quality burden from optics into software and, eventually, the cloud.