Video Coding

Содержание и основные понятия

By table of contents
Alterna chapter 21.2
Video Coding
Permits, frame rate, compression, codecs, interfaces, HDR.
01
Resolution and frame rate
From SD to 8K: standards, sides and frequency of personnel
GEF
OGE No. 7.9
Video resolution
The resolution is the number of pixels horizontal and vertical. It determines the details of the picture. The standard (SD) is 720×480 (NTSC) or 720×576 (PAL). HD Ready is 1280×720. Full HD is 1920×1080. 4K UHD is 3840×2160. 8K is 7680×4320.
Name
Authorization
Peaksele
Ratio
Application
480p (SD)
720×480
345 600
3:2
Old DVD, Internet video
720p (HD)
1280×720
921 600
16:9
YouTube, web video
1080p (FHD)
1920×1080
2 073 600
16:9
Blu-ray, strimming, games
1440p (QHD)
2560×1440
3 686 400
16:9
Monitors, game monitors
4K UHD
3840×2160
8 294 400
16:9
TV 2018+, films, content
2K (DCI)
2048×1080
2 211 840
256:135
Film Projection
8K UHD
7680×4320
33 177 600
16:9
Professional, future
Frame Rate
Frame rate = number of frames per second (fps)
24 fps
Cinematographer: The standard from 1927 creates a "kinematographic" feeling.
25 fps
PAL (Europe, Russia) Television standard.
30 fps
NTSC (USA, Japan) Television Standard.
50/60 fps
Modern video, sports, game content, swimming.
120/144 fps
Gaming, VR, high-end monitors.
240+ fps
Slow shot, scientific video.
Progressive vs too rigid
Progressive (p)
- every shot is full: 1080p.
Interlaced (i)
- Each frame is divided into two "fields" (even and odd lines). 1080i = two half-cams. Less visible blinking with the same beam, but can give artifacts when moving.
🎞
Progressive vs Chercestered
1080p - Progressive
All lines each frame
1080i - Chilli
Half-cuts: odd + even lines
Show fields
02
Video compression: I, P and B-frames
How cross-staff compression works: key and diverse people
GEF
EEG No. 9
Intra-frame vs Inter-frame
Intra-frame (I-frame)
===Career===The full frame is compressed as a separate image (JPEG-like).
Inter-frame (P/B-frames)
- only stored
differences
The result is 5 to 20 times more effective than I-frame.
I
I-frame (Keyframe)
Full frame. Keeps all the information, the biggest one, but it's closing independently.
P
P-frame (Predicted)
Differentials forward. Keeps changes to the previous I or P frame. Compensation for traffic. Less than I-frame by three or five times.
B
B-frame (Bidirectional)
Differentials in both directions. Could refer to the previous and next I/P frames. Maximum compression (5-10 times less than I). The difficulty of decoding.
Example: GOP (Group of Pictures)
Typical GOP:
I B B P B B P B B I
- GOP length = 10 frames, I-frame every 10 frames. At 30 fps, that means full frame every ~0.33 seconds.
More B-frames between I \ above compression, but more delay (lateny) in decoding. For stream video, usually: IBBP or IPPP.
🎬
GOP: Group of Pictures — Video Flow Structure
60 fps
Press "Reproduce" to see the sun rise and sunset against mountains and trees
Size of frame:
I
100%
P
~30%
B
~10%
GOP: I B B P B B P B B I
Motion Compensation
Moving objects are tracked by blocks (macroblocks). For each block, P/B-frame is stored.
Traffic vector
This allows movement to be compressed much more efficiently than simply subtracting people.
03
Codes and File Formats
Containers and Codecs: What to Open
GEF
PIF
Container vs Codek
Container
(file format) is a "box" that stores audio, video, subtitles, metadata together.
Codes
- Compressor algorithm / video unpacking or audio inside the container. One container may contain different codecs.
Containers
📂 MP4 (MPEG-4 Part 14)
Container
Universal, video H.264, H.265, AAC, YouTube, iTunes, cameras.
📂 MKV (Matroska)
Container
Open, free, all codecs, commonly used for HD reeps.
📂 MOV
Container
Apple QuickTime, Professional Video, Final Cut Pro.
📂 AVI
Container
The old Microsoft format. Support for different codecs, but limited.
Videos
🎬 H.264 (AVC)
Codek · Partner: MPEG-4 Part 10
The most common. Blu-ray, YouTube, Netflix (old), good compression, intercompatibility, Bitrate 5-25 Mbit/s for 1080p.
🎬 H.265 (HEVC)
Coden H.264
It's twice as effective as H.264 at the same quality. 4K/8K TV, iPhone video. Patent, licensing is spores.
🎬 VP9
Code · Google, open
Alternative H.265. YouTube (WebM) Free patents, comparable quality at a similar bite.
🎬 AV1
Code · Alliance for Open Media
Modern, free (royalty-free) 30% more efficient than H.265. The future of strimming (YouTube AV1, Netflix in the future).
Performance comparison (1080p, ~1 Mbit/s)
H.264: Basic quality, compatibility.
H.265: 40% better with the same bite.
VP9: ~ like H.265, available.
AV1: Best today, free but heavy for coding.
04
Video interfaces
VGA, DVI, HDMI, DisplayPort, USB-C: Capacity Comparison
GEF
Interface
Type
Max Passage
Max resolution / fps
Audio
Application
VGA (D-Sub)
Analog
~1 Gbit/s
2048×1536 / 85 Hz
None
Legacy, old monitors
DVI-D (Single Link)
Digital
1.65 Gbit/s
1920×1080 / 60 Hz
None
Legacy, Old GPU
HDMI 1.4
Digital
10.2 Gbit/s
4K / 30 Hz
Yes (8ch)
TV, monitors, game consoles
HDMI 2.0
Digital
18 Gbit/s
4K / 60 Hz
Yes (32ch)
4K TV, PS4 Pro, Xbox One X
HDMI 2.1
Digital
48 Gbit/s
8K / 60 Hz, 4K / 120 Hz
Yes (32ch eARC)
New-generation game consoles, RTX 30+, RX 6000+
DisplayPort 1.4
Digital
32.4 Gbit/s
8K / 60 Hz, 4K / 120 Hz
Yes (32ch)
Professional monitors, GPU
DisplayPort 2.0
Digital
80 Gbit/s
8K / 60 Hz, 4K / 240 Hz
Yes
Promising Displays, USB4
USB-C (Alt Mode)
Digital
Depends on mode
Depends (DP/TB)
Yes
Laptops, USB-C monitors, Thunderbolt 3/4
🔌
Comparison of interface capacity
4K 60fps HDR needs ~18 Gbit/s. DisplayPort 2.0 is the fastest (80 Gbps/s).
05
HDR in video
High Dynamic Range: 8-bit vs 10 bit vs 12bit, HDR standards
GEF
PIF
What HDR is
HDR
High Dynamic Range is an extended dynamic range. Standard SDR uses 8 bits per canal (256 levels per color). HDR uses 10 bits (1,024 levels) or 12 bits (4,096 levels).
8 bits
SDR (Standard Dynamic Range)
16.7 million colors. 256 levels per channel. Enough for most content. Used in DVD, Blu-ray (standard), most videos.
10 bits
HDR10, HDR 10+, Dolby Vision
1.07 billion flowers, floating gradients without stripes, Ultra HD Blu-ray standard, new TV, YouTube HDR, target luminance: 1000-400 ft.
12 bits
Dolby Vision, future
68.7 billion flowers, almost invisible gradients, professional filmmakers, Dolby Vision on maximum settings.
HLG
Hybrid Log-Gamma (BBC/NHK)
Broadcast HDR. Reverse compatibility with SDR.
HDR standards
HDR10
===MaxCLL===It is an open standard, 10 bits, and static metadata.
HDR10+
is a dynamic metadata (Samsung, Amazon).
Dolby Vision
- Dynamic metadata, up to 12 bits, professional cinema, Dolby licensing.
HLG
- for Broadcast, reverse compatibility.
📡
EGE No. 10: Data transmission speed
Formula:
V = v x t, where V is the data volume, v is the transmission rate, t is the time.
Units:
1 Mbit/s = 1024 Kbps = 10 24 x 1024 bits/s.
It is important:
1 byte = 8 bits, so speed in bytes/s = speed in bits/s pi 8.
📡
Objective: Data transmission rate
A file of 120 MB is transmitted through the channel at 10 Mbit/s. How many seconds will the transmiss