Why Video Compression Matters: Decoding Formats, Codecs, and Streaming Essentials

Since 2015, live‑streaming apps like Inke, Huajiao, Kuaishou, Douyu, and others have exploded, prompting developers to tackle the technical challenges of building streaming platforms.

Background

Early web‑based live streams relied on ActiveX controls for IE, then Flash with Adobe Flash Media Server or RealLive Stream Server. Limited bandwidth and technology kept the market small until mobile internet and 4G popularized video capture on smartphones, leading to a surge in live‑streaming features across many apps.

File Extensions Overview

Common video containers include .wmv, .avi, .divx, .flv, .mp4, .mov, .mkv, .rmvb, .ogg, .webm, among others. These extensions represent container formats that bundle various codec tracks; the extension alone does not reveal the underlying encoding.

Video Compression Technology

Video data is massive before compression, requiring high‑capacity storage and bandwidth. Modern smartphones (e.g., iPhone 7) support 4K/30 fps, 1080p/60 fps, and slow‑motion modes, but raw footage still demands efficient compression.

Key compression considerations include:

Real‑time vs. non‑real‑time processing

Symmetric vs. asymmetric encoding

Compression ratio

Lossless vs. lossy compression

Intra‑frame vs. inter‑frame comparison

Bitrate control

Live streams (e.g., YY) aim for true real‑time delivery with no perceptible delay, while “pseudo‑live” streams introduce buffering delays of several seconds.

Higher frame rates (e.g., 60 fps) provide smoother motion, but insufficient frame rates cause noticeable stutter, especially when frames containing critical audio‑visual sync information are lost.

Raw data size grows quickly with resolution and frame rate: 720p @ 25 fps ≈ 263 MB/s, 1080p @ 25 fps ≈ 593 MB/s, illustrating the massive storage and bandwidth demands of high‑definition video.

Compression Standards Overview

Major video codecs and their typical use cases:

H.261 – video conferencing over ISDN (≈ p × 64 kb/s)

MPEG‑1 – CD‑ROM quality video (≈ 1.5 Mb/s)

MPEG‑2 – digital TV and HDTV (≥ 2 Mb/s)

H.263 – low‑bitrate PSTN video (< 33.6 kb/s)

MPEG‑4 – object‑based coding, interactive media (variable bitrate)

H.264 (AVC) – low‑bitrate HD, surveillance, video‑conference (10‑100 kb/s)

The MPEG compression pipeline consists of five steps: resolution reduction, motion estimation, discrete cosine transform (DCT), quantization, and entropy coding. Converting RGB to YUV exploits the human eye’s lower sensitivity to color detail, further reducing data.

Frames are categorized as I‑frames (intra‑coded), P‑frames (predictive based on previous frames), and B‑frames (bidirectional prediction using both past and future frames). H.264 enhances this model with GOP structures to mitigate packet loss and improve bandwidth efficiency.

Conclusion

The article introduced the principles of video compression and codec evolution. As bandwidth constraints ease, future compression will focus more on eliminating video redundancy rather than merely reducing size. Today, bandwidth remains a critical cost factor for live‑streaming services, making efficient video compression essential.