Ever wonder what a video file actually is, underneath the play button? Let’s build it up from scratch — no jargon, just first principles.
1. The bit
Everything in a computer starts as binary — a 0 or a 1. That’s the entire alphabet a machine speaks. Every piece of digital media, no matter how rich, eventually reduces to a long string of these.
2. The pixel
Group enough bits together and you can describe a color. A pixel is just a tiny colored dot, usually represented as three numbers — Red, Green, Blue (RGB) — each typically stored in 8 bits (a value from 0–255). More bits per channel means more color precision. This is called bit depth — 8-bit color is standard, while 10-bit or 12-bit (“HDR”) gives smoother gradients and richer color, at the cost of more data.
3. The image (a single frame)
Arrange thousands of these pixels in a grid — say, 1920 wide by 1080 tall — and you get a single image, or frame. That’s over 2 million pixels, each needing 3 bytes of color data (1 byte per channel). A single frame, uncompressed, is already about 6MB.
4. The video
A video is simply a rapid sequence of these frames, played back at a fixed rate — typically 24, 30, or 60 frames per second (the frame rate). Play them fast enough, and the human eye stops seeing a slideshow and starts perceiving continuous motion.
The problem this creates
At 30 frames per second, 1080p resolution, fully uncompressed:
- 1 frame ≈ 6MB
- 1 second of video ≈ 180MB
- 1 minute ≈ ~10.8GB
- A 2-minute ad ≈ ~21GB
Raw video is simply too large to store, edit, or send anywhere practically. Every video you’ve ever streamed, downloaded, or recorded on your phone has been through a process designed to solve exactly this problem: compression.
That’s where Part 2 picks up — how video codecs like MPEG throw away what your eyes won’t miss, and carefully keep what they will, using the concepts of I-frames, P-frames, and B-frames.
This is Part 1 of a series on how video actually works under the hood — from raw bits to the compressed, streamable files we use every day.
Note: Drafted with help from Claude and Gemini — ideas, opinions, and the GPU bills are mine 😊