Skip to content

videocodec.md

Latest commit

 

History

History
303 lines (189 loc) · 6.21 KB

videocodec.md

File metadata and controls

303 lines (189 loc) · 6.21 KB

Video Compression Standards for Embedded Camera Systems

A Technical Whitepaper on H.264 (AVC) and H.265 (HEVC)

Version: 1.0
Date: 01-2026
Audience: Embedded engineers, camera architects, firmware developers

1. Introduction

Video compression is a core component of modern embedded camera systems, enabling real-time video transmission and storage under constrained bandwidth, storage, and power conditions.

Two dominant international standards are:

  • H.264 / AVC (Advanced Video Coding)
  • H.265 / HEVC (High Efficiency Video Coding)

This document provides:

  • Core concepts of both codecs
  • Technical comparison
  • Advantages & limitations
  • Standards and RFC references
  • Practical considerations for embedded camera platforms

2. Video Codec Fundamentals

2.1 What is a Video Codec?

A video codec is a method to:

  • Compress raw video data for transmission/storage
  • Decompress encoded video for playback

Primary goals:

  • Reduce bitrate
  • Maintain acceptable visual quality
  • Enable real-time processing on limited hardware

Raw video example (uncompressed):

1920 × 1080 × 8bit × 3 × 30fps ≈ 1.49 Gbps

→ Not feasible without compression.

2.2 Block-Based Hybrid Video Coding Model

Both H.264 and H.265 are based on the hybrid block-based coding architecture:

Image

Image

Main components:

  1. Block partitioning
  2. Prediction (intra / inter)
  3. Transform & quantization
  4. Entropy coding
  5. Loop filters

3. H.264 / AVC Overview

3.1 Standard Definition

  • Standard name: ITU-T H.264 / ISO/IEC 14496-10
  • First published: 2003
  • Widely deployed in IP cameras, CCTV, streaming

3.2 Key Characteristics

  • Macroblock size: 16×16

  • Prediction types:

    • Intra (spatial)
    • Inter (temporal, motion compensation)

  • Entropy coding:

    • CAVLC (baseline)
    • CABAC (main/high)

Image

3.3 Profiles Commonly Used in Cameras

Profile Use case
Baseline Low-end MCU, low latency
Main Consumer IP cameras
High High-quality recording / NVR

3.4 Advantages

  • Mature ecosystem
  • Low hardware complexity
  • Broad hardware decoder support
  • Stable real-time performance

3.5 Limitations

  • Lower compression efficiency
  • Higher bitrate at high resolutions
  • Less optimal for 4K+ video

4. H.265 / HEVC Overview

4.1 Standard Definition

  • Standard name: ITU-T H.265 / ISO/IEC 23008-2
  • Published: 2013
  • Designed for high-resolution video (4K / 8K)

4.2 Key Innovations

  • Coding Tree Units (CTU): up to 64×64
  • More flexible partitioning
  • Improved motion vectors
  • Advanced intra prediction
  • Enhanced entropy coding

Image

Image

4.3 Compression Efficiency

  • ~40–50% bitrate reduction vs H.264 at same quality

  • Ideal for:

    • 4K cameras
    • Low-bandwidth networks
    • Cloud storage optimization

4.4 Advantages

  • Superior compression efficiency
  • Better subjective quality at low bitrate
  • Designed for UHD content

4.5 Limitations

  • Higher encoder complexity
  • Increased CPU / hardware requirements
  • Licensing concerns in some markets
  • Slightly higher encoding latency

5. H.264 vs H.265 Comparison

Feature H.264 H.265
Max block size 16×16 64×64
Compression Baseline ~50% better
CPU usage Lower Higher
Latency Lower Slightly higher
Embedded support Excellent Platform-dependent
4K efficiency Moderate Excellent

6. How to Distinguish H.264 and H.265 in Practice

6.1 By Stream Metadata

  • SDP:
a=rtpmap:96 H264/90000
a=rtpmap:96 H265/90000

6.2 By NAL Unit Type

  • H.264 start code: Image

H264 obtains the NALU Type by code & 0x1F,

Image

00 00 00 01 67

  0110 0111 (0x67)
& 0001 1111 (0x1F) 
-----------
  0000 0111 (0x07) = 7(DEC) => (SPS)
  • H.265 NAL header structure differs (2-byte header)
    Image

H265 obtains the NALU Type by (code & 0x7E) >> 1.

Image

  0010 0110 (0x26)
& 0111 1110 (0x7E)
-----------
  0010 0110 (0x26) >> 1 = 0001 0011 (0x13) = 19 (DEC) => (IDR)

6.3 By Bitrate Observation

  • Same resolution & quality:

    • H.265 bitrate ≈ 50–60% of H.264

7. RFC and Standard References

7.1 RTP Payload Formats

7.2 SDP & Transport

7.3 Codec Standards

8. Embedded Camera Design Considerations

When choosing a codec:

  • Available hardware encoder (ISP / VPU)
  • Target resolution & framerate
  • Network conditions
  • Power & thermal constraints
  • Decoder compatibility (NVR, mobile app)

Typical recommendation:

  • 1080p / low latency → H.264
  • 4K / cloud storage / WAN → H.265

9. Conclusion

H.264 remains the most compatible and robust choice for embedded cameras, while H.265 provides significant bandwidth and storage savings for high-resolution and next-generation systems.

A well-designed camera platform often supports both codecs, allowing flexible deployment across different scenarios.