Compression ratio is the simple measure of how much smaller a compressed file is compared to its uncompressed source. Written as a number like "500:1" or as a percentage like "99.8 %". A two-hour HD movie raw is around 670 GB; as a streaming H.264 file it's about 2 GB, a compression ratio of roughly 335:1. The same movie in AV1 might come down to 1 GB — a ratio close to 670:1.
Two reference points for the numbers. Lossless compression of video typically hits 2:1 to 4:1 — every bit of original information is preserved, but the savings are modest. Lossy compression (everyday H.264, HEVC, AV1) routinely hits 100:1 to 1000:1 by throwing away information the viewer can't notice. Per codec generation, the ratio improves: H.264 → HEVC saves ~50 %, HEVC → AV1 saves ~30 %, AV1 → VVC is similar. The result: each new codec generation lets you ship the same visual quality through a smaller pipe.
For a product team, compression ratio is a useful executive shorthand for "how aggressive is our delivery encoding?" but it's only meaningful in context. A 1000:1 ratio on a static talking head is unimpressive; the same 1000:1 ratio on a snow-storm action sequence is remarkable. What actually matters is the quality-versus-size trade-off — typically measured by VMAF or SSIM at a given bitrate, not by the raw ratio. Quoting compression ratios in marketing is fine; making engineering decisions on them alone is a recipe for shipping unwatchable encodes that happen to be very small.
