Product image optimization at Mister Spex
summary
An automated unsupervised pipeline converts supplier photos to a 16-bit PNG intermediate, applies a two-step ImageMagick resize, then iteratively compresses with progressive JPEGs and tuned chroma subsampling (preferring 4:2:0), measuring structural dissimilarity (DSSIM) and using an adaptive entropy-based quality adjustment to pick the lowest quality that stays below the visual threshold. PNGs use PNGQuant index-color conversion when transparency is needed. The system runs on thousands of SKUs, pushes results to S3/CloudFront, and reduces typical product-file size to about 68% while keeping acceptable visual fidelity.
AI-written summary · about
chapters (7)
- 0:00Start
- 3:41Product Images
- 18:10Resizing
- 20:02Image Complexity
- 23:14Retina Images
- 26:35Shrinking Process
- 28:56Chroma Subsampling
related talks
44mImage Optimization for eCommerce websites
37mImage Formats, Performance and Cognitive Load
06mOvercoming the Challenges of Image Delivery
30mReimagining the image pipeline at Squarespace
52mThe JPEG is dead! Long live the JPEG!
17mEfficient & Frugal Image Data Processing Pipelines
03mDelivering fast and beautiful images and video
41mDeliver Perfect Images At Any Size