science Interactive Laboratory
PROTOTYPE v1.31. The Player View Client Side
Hover over the timeline below. This simulates what a user sees.
Real-time Calculation
2. The Server Reality Asset on CDN
This is the single image file (Sprite Sheet) actually downloaded by the browser. The red box shows what the user sees.
The browser loads this one image instead of 25 separate requests.
Mapped VTT File (storyboard.vtt)
1. The Sprite Sheet
When a movie is uploaded, a transcoding server extracts a frame every few seconds. Instead of saving thousands of tiny image files, it stitches them into a single, massive grid image called a Sprite Sheet.
2. The VTT Map
A text file (WebVTT) acts like an index, telling the player: "At 00:05, look at the sprite sheet at coordinates x=160, y=0".
3. The CSS Magic
When you hover, the browser doesn't crop the image. It uses `background-position` to shift the massive sprite sheet behind a small viewing window.
Why do it this way?
Imagine a 2-hour movie with a preview every 10 seconds. That's 720 images.
block Old Way (Individual Images)
- • 720 separate HTTP requests
- • Huge overhead from headers (TCP/IP handshake)
- • Choppy UI as images load one by one
check_circle Netflix Way (Sprite Sheet)
- • 1 single HTTP request
- • Instant rendering (image is pre-cached)
- • Smooth 60fps scrubbing experience