数字视频与高清电视算法及接口详解

5星 · 超过95%的资源需积分: 9 122 浏览量更新于2024-07-22 2 收藏 3.59MB PDF 举报

"Digital Video and HDTV Algorithms and Interfaces" 是一本深入探讨数字视频与高清电视（HDTV）处理技术的专业书籍，适合对视频处理技术感兴趣的读者。这本书属于 Morgan Kaufmann 出版社的计算机图形学和几何建模系列，由 Charles Poynton 编著。在现代数字媒体技术中，数字视频和高清电视占据了重要地位，它们涉及到图像采集、编码、传输和显示等多个环节。本书详细阐述了这些环节中的关键算法和技术接口，为读者提供了一个全面的视角来理解和应用相关知识。 1. **数字视频基础**：书中可能涵盖了数字视频的基础概念，包括像素、分辨率、帧率、位深度等基本元素，以及数字视频的压缩标准，如MPEG、H.264等。 2. **视频信号处理**：讨论了视频信号的采样、量化、编码过程，以及如何通过滤波、去噪、增强等技术改善视频质量。 3. **高清电视（HDTV）技术**：详述了HDTV的规格标准，如分辨率、宽高比等，并可能涉及到高清信号的传输、接收和显示技术。 4. **接口与协议**：书中可能包含关于视频接口（如HDMI、DVI、SDI等）的介绍，以及数据传输协议（如AES/EBU、SMPTE等）的工作原理。 5. **视频处理算法**：详细介绍了视频处理中的各种算法，如运动估计与补偿、帧间预测、色彩空间转换、图像缩放、视频稳定等。 6. **图像处理**：可能会涵盖图像处理的基本技术，如边缘检测、图像分割、色彩空间转换等，以及这些技术在视频处理中的应用。 7. **计算机图形学结合**：书中的内容可能也涉及到计算机图形学的一些概念，如光照模型、纹理映射、3D渲染等，这些技术常用于视频特效和后期制作。 8. **实时处理与交互**：可能讨论了如何实现高效实时的视频处理，以及在交互式系统中如何设计用户友好的视频接口。 9. **实例分析与应用**：通过实际案例分析，展示了所讲述的理论和技术如何在现实项目中得到应用。 10. **最新进展与未来趋势**：可能还包括对当前技术发展趋势的分析，如4K、8K超高清视频，以及虚拟现实（VR）、增强现实（AR）等新兴领域。 "Digital Video and HDTV Algorithms and Interfaces" 是一本深入的教程，它不仅提供了理论基础，还涵盖了实际应用和技术接口的细节，对于想要深入了解数字视频和高清电视技术的读者来说是一本宝贵的参考资料。

10 DIGITAL VIDEO AND HDTV ALGORITHMS AND INTERFACES

Spatiotemporal domains

A sequence of still pictures captured and displayed at

a sufficiently high rate

–

typically between 24 and 60

pictures per second

–

can create the illusion of motion,

as I will describe on page 51. Sampling in time, in

combination with 2-D (spatial) sampling, causes digital

video to be sampled in three axes

–

horizontal, vertical,

and temporal

–

as sketched in Figure 1.13 above. One-

dimensional sampling theory, to be detailed in Filtering

and sampling, on page 141, applies along each axis.

At the left of Figure 1.13 is a sketch of a two-dimen-

sional spatial domain of a single image. Some image

processing operations, such as certain kinds of filtering,

can be performed separately on the horizontal and

vertical axes, and have an effect in the spatial domain –

these operations are called separable. Other processing

operations cannot be separated into horizontal and

vertical facets, and must be performed directly on

a two-dimensional sample array. Two-dimensional

sampling will be detailed in Image digitization and

reconstruction, on page 187.

VERTICAL

TEMPORAL

SPATIAL

HORIZONTAL

(TRANSVERSE)

Figure 1.13 Spatio-

temporal domains

CHAPTER 1 RASTER IMAGES 11

Lightness terminology

In a grayscale image, each pixel value represents what is

loosely called brightness. However, brightness is defined

formally as the attribute of a visual sensation according

to which an area appears to emit more or less light. This

definition is obviously subjective, so brightness is an

inappropriate metric for image data.

See Appendix B, Introduction to

radiometry and photometry, on

page 601.

Intensity is radiant power in a particular direction;

radiance is intensity per unit projected area. These

terms disregard wavelength composition. But in color

imaging, wavelength is important! Neither of these

quantities is a suitable metric for color image data.

Luminance is radiance weighted by the spectral sensi-

tivity associated with the brightness sensation of vision.

Luminance is proportional to intensity. Imaging systems

rarely use pixel values proportional to luminance; values

nonlinearly related to luminance are usually used.

Illuminance is luminance integrated over a half-sphere.

Lightness

–

formally, CIE L*

–

is the standard approxi-

mation to the perceptual response to luminance. It is

computed by subjecting luminance to a nonlinear

transfer function that mimics vision. A few grayscale

imaging systems have pixel values proportional to L*.

Regrettably, many practitioners of

computer graphics, and of digital

image processing, have a cavalier

attitude toward these terms. In the

HSB, HSI, HSL, and HSV systems,

B allegedly stands for brightness,

I for intensity, L for lightness, and

V for value. None of these systems

computes brightness, intensity,

luminance, or value according to

any definition that is recognized in

color science!

Value refers to measures of lightness apart from CIE L*.

In image science, value is rarely

–

if ever

–

used in any

sense consistent with accurate color. (Several different

value scales are graphed in Figure 20.2 on page 208.)

Color images are sensed and reproduced based upon

tristimulus values, whose amplitudes are proportional to

intensity but whose spectral compositions are carefully

chosen according to the principles of color science. As

their name implies, tristimulus values come in sets of 3.

The image sensor of a digital camera produces values,

proportional to radiance, that approximate red, green,

and blue (RGB) tristimulus values. (I call these values

linear-light.) However, in most imaging systems, RGB

tristimulus values are subject to a nonlinear transfer

The term luminance is often care-

lessly and incorrectly used to refer

to luma; see below. In image

reproduction, we are usually

concerned not with (absolute)

luminance, but with relative lumi-

nance, to be detailed on page 206.

12 DIGITAL VIDEO AND HDTV ALGORITHMS AND INTERFACES

function

–

gamma correction

–

that mimics the percep-

tual response. Most imaging systems use RGB values

that are not proportional to intensity. The notation

R’G’B’ denotes the nonlinearity.

See Appendix A, YUV and luminance

considered harmful, on page 595.

Luma (Y’) is formed as a suitably weighted sum of

R’G’B’; it is the basis of luma/color difference coding.

Luma is comparable to lightness; it is often carelessly

and incorrectly called luminance by video engineers.

Nonlinear image coding

Vision cannot distinguish two luminance levels if the

ratio between them is less than about 1.01

–

in other

words, the visual threshold for luminance difference is

about 1%. This contrast sensitivity threshold is estab-

lished by experiments using the test pattern such as the

one sketched in Figure 1.14 in the margin; details will

be presented in Contrast sensitivity, on page 198.

Consider pixel values proportional to luminance, where

code zero represents black, and the maximum code

value of 255 represents white, as in Figure 1.15.

Code 100 lies at the point on the scale where the ratio

between adjacent luminance values is 1%: The

boundary between a region of code 100 samples and

a region of code 101 samples is likely to be visible.

As the pixel value decreases below 100, the difference

in luminance between adjacent codes becomes increas-

ingly perceptible: At code 25, the ratio between adja-

cent luminance values is 4%. In a large area of smoothly

varying shades of gray, these luminance differences are

likely to be visible or even objectionable. Visible jumps

in luminance produce artifacts known as contouring or

banding.

Linear-light codes above 100 suffer no banding arti-

facts. However, as code value increases toward white,

the codes have decreasing perceptual utility: At code

200, the luminance ratio between adjacent codes is just

0.5%, near the threshold of visibility. Codes 200 and

201 are visually indistinguishable; code 201 could be

discarded without its absence being noticed.

Y+∆Y

Figure 1.14 Contrast sensi-

tivity test pattern reveals

that a just-noticeable differ-

ence (JND) occurs when the

step between luminance

levels is 1% of Y.

∆ = 1%

100

101

255

∆ = 0.5%

200

201

∆ = 4%

2.55 : 1

Figure 1.15 The “code 100”

problem with linear-light

coding is that at code levels

below 100, the steps between

code values have ratios larger

than the visual threshold: The

steps are liable to be visible.

CHAPTER 1 RASTER IMAGES 13

High-quality image reproduction requires a ratio of at

least 30 to 1 between the luminance of white and the

luminance of black, as I will explain in Contrast ratio, on

page 197. In 8-bit linear-light coding, the ratio between

the brightest luminance (code 255) and the darkest

luminance that can be reproduced without banding

(code 100) is only 2.55:1. Linear-light coding in 8 bits is

unsuitable for high-quality images.

This “code 100” problem can be mitigated by placing

the top end of the scale at a code value higher than

100, as sketched in Figure 1.16 in the margin. If lumi-

nance is represented in 12 bits, white is at code 4095;

the luminance ratio between code 100 and white

reaches 40.95:1. However, the vast majority of those

4096 code values cannot be distinguished visually; for

example, codes 4001 through 4040 are visually indis-

tinguishable. Rather than coding luminance linearly

with a large number of bits, we can use many fewer

code values assigned nonlinearly on a perceptual scale.

If the threshold of vision behaved strictly according to

the 1% relationship across the whole tone scale, then

luminance could be coded logarithmically. For a con-

trast ratio of 100:1, about 463 code values would be

required, corresponding to about 9 bits. In video,

for reasons to be explained in Luminance and lightness,

on page 203, instead of modeling the lightness sensi-

tivity of vision as a logarithmic function, we model it as

a power function with an exponent of about 0.4.

Conversely, monitor R’G’B’ values

are proportional to reproduced

luminance raised to approximately

the 0.4-power.

The luminance of the red, green, or blue primary light

produced by a monitor is proportional to voltage (or

code value) raised to approximately the 2.5-power. This

will be detailed in Chapter 23, Gamma, on page 257.

The cathode ray tube (CRT) is the

dominant display device for tele-

vision receivers and for desktop

computers.

Amazingly, a CRT’s transfer function is nearly the

inverse of vision’s lightness sensitivity! The nonlinear

lightness response of vision and the power function

intrinsic to a CRT combine to cause monitor voltage, or

code value, to exhibit perceptual uniformity, as demon-

strated in Figures 1.17 and 1.18 overleaf.

∆ = 1%

100

101

4095

40.95 : 1

Figure 1.16 The “code 100”

problem is mitigated by using

more than 8 bits to represent

luminance. Here, 12 bits are

used, placing the top end of the

scale at 4095. However, the

majority of these 4096 codes

cannot be distinguished visually.

lg .

100

101

463 1 01 100

463

≈≈

剩余699页未读，继续阅读

sinat_25828795

粉丝: 0
资源: 1

数字视频与高清电视算法及接口详解

Digital Video And HDTV Algorithms And Interfaces.pdf

Digital Video and HDTV

Digital Video And Hdtv Algorithms And Interfaces

Digital Video and HDTV Algorithms and Interfaces

Digital Video and HDTV Algorithms and Interfaces.pdf

Digital Video and HDTV, Algorithms and Interfaces

Digital Video and HD Algorithms and Interfaces, Second Edition

Digital_Video_and_HDTV_Algorithms_and_Interfaces.r_电子书籍_Video_

audit-libs-static-2.8.5-4.el7.i686.rpm.zip

avahi-tools-0.6.31-20.el7.x86_64.rpm.zip

最新资源