Comprehensive analysis of video compression technology and MPEG noise reduction technology


Theoretically, the digital TV (DTV) picture quality is better than the traditional analog TV, without ghosting, snowflakes, flutter and color distortion. Moreover, as can be demonstrated by analog TV signals, the biggest flaw is that there are many spots on the picture, and the picture is not delicate enough due to insufficient response to high-frequency signals. In short, the bandwidth is not enough. The more detailed the image, the higher the resolution and the greater the bandwidth required.

A long time ago, the US official allocated every 6MHz bandwidth in the available spectrum to each channel of the ABC to provide analog TV signals. This limitation on video bandwidth and its corresponding display standard (NTSC color space) decided The characteristics of traditional TV sets have been determined, and the quality of TV pictures has been determined for decades.

With the advent of digital television, broadcasters have seen opportunities to more fully utilize their allocated bandwidth. Indeed, from their point of view, the most prominent advantage of digital TV is that it allows more channels to be transmitted in the same bandwidth, and can also support subsequent high-definition television programs (HDTV).

Lengthy data

Bandwidth limit

HDTV has very high technical requirements. The traditional NTSC signal that propagates analog signals must use a minimum bandwidth of 4.2 MHz within a 6 MHz bandwidth of a channel, and scan 525 lines at a field frequency of 29.97 Hz. After digital quantization, encoding and compression, the signal can be recorded on a DVD with a bit transmission bit rate from 2Mbits / s to 10Mbits / s (supporting self-adaption), with an average of 4Mbits / s. In comparison, a typical HDTV has 5 times the resolution of an analog TV. Therefore, under the same conditions, the transmission data rate should be 5 times that of the analog signal to achieve the same performance.

Whether it is traditional over-the-air broadcasting (OTA), cable TV set-top boxes, or satellite TV, they are subject to bandwidth constraints when transmitting signals. In the limited bandwidth, they also need to add bandwidth-consuming services, including interactive broadcasting, Premium channels and TV program tables, etc.

So, how can we solve the problem? Using compression technology is one way.

Digital video compression causes distortion

The most commonly used digital video compression algorithm is MPEG-2. From the existing satellite TV transmission, cable digital TV transmission to aerial digital broadcasting, MPEG-2 has been widely adopted internationally in various applications.

MPEG-2 first removes temporal redundancy through motion compensation, then divides a frame of images into 8x8 pixel lattices, and uses DCT (discrete cosine transform) to remove spatial redundancy within each lattice. After the DCT is completed, the compression is completed through quantization and recombination, then variable length coding is performed, and finally Huffman coding. The entire compression process greatly reduces the bit rate (> 10: 1 compression ratio).

However, the reduction in bit rate also brings problems, because the encoding loses some original video information, which may cause serious negative effects, so MPEG-2 is called lossy encoding. It discards image information that is considered visually minor. The greater the compression, the greater the difference between the encoded image and the original image. Image quality and fidelity now depend on the selected (or usually applied) compression level. Because it is directly related to the available bandwidth, we must ask ourselves when will excessive video compression not occur?

Visible distortion

The bandwidth limitation and excessive image compression in digital signal transmission make the compressed image completely different from that seen in the analog world.

In general, analog image degradation (or noise) often appears in the form of Gaussian noise. The advantage of this noise is that it retains basic content and is not easily noticed because of human visual defects. We often see simulated images that are a little blurry and uncomfortable, but this does not make people feel obviously disgusted.

Digital noise follows a different distribution pattern, and more importantly, its special form makes people feel unnatural. When MPEG-2 encoding (or any DCT-based codec) is used to its limit, there are two main ways of distortion: Mosquito noise and Blocking artifacts.

Mosquito noise

On clear, colorful backgrounds, mosquito noise around prominent objects, computer-simulated objects, or scrolling characters is most noticeable. It looks like some kind of hazy thing or glitter around the high-frequency boundary between the object and the background (a sharp jump formed between the foreground object and the background), and sometimes it is even mistaken for a surround. Unfortunately, this small effect can also be seen on shapes closer to nature like the human body. The VIRIS project team (Video Reference Injury System) defines mosquito noise as “distortion that accompanies the edges of moving objects, manifested as a layer of flying objects and / or blurry air bubbles around the object (like mosquitoes surrounding people Head and shoulders fly) ".

When the image is reconstructed and some data is discarded due to the use of inverse cosine transform, mosquito noise appears.

"Mosquitoes" can also be found in other parts of an image, for example, mosquito noise can also appear at specific texture boundaries or at granular objects. The result is a bit similar to random noise. The noise seems to be mixed with texture or particles, and looks like the original features of the picture.

The square effect, worthy of the name, shows annoying and unnatural squares in the image. Sometimes it appears as a large block, it is a kind of image distortion, and is caused by the block coding structure.

When the encoding is maximized, each pixel lattice is averaged fairly coarsely, making it look like a large pixel. The calculation of each pixel lattice is different, so that there is a clear boundary between each lattice.

This effect is more pronounced when the object or camera is moving fast. The best example is that during the NFL (National Football League) broadcast, athletes holding the ball looked like the Mario brothers in the old Nintendo game console.

Pre-smooth

Although pre-smoothing is not one of the image compression processing algorithms, it has been used to eliminate this digital distortion.

Broadcasters and content providers have become increasingly aware of the shortcomings of their distribution systems, and some of them have adopted a rather controversial solution to existing bandwidth limitations: pre-smoothing.

By eliminating the high-frequency part of the image before the signal is input to the channel, the encoder has more time to process its task, and the resulting image is less affected by the square effect and mosquito noise. On the other hand, this kind of over-filtering to a certain extent also loses all subtle changes and textures in the original image.

For example, a football player with a week of beard now looks as if his beard is shaved cleanly (even if he is stationary), while the stadium looks like a large green carpet.

It can be proved that although some people think that pre-smoothing is also good, it is completely an irreversible algorithm. Once the details are processed, people can no longer rebuild them.

However, the square effect and mosquito noise did disappear.

Body Soap

private label body soap ,30 years experience guangzhou cosmetics factory supply ,OEM/ODM service is welcomed .
Guangzhou Athena Cosmetics Manufacturing Co.,Ltd is a subsidiary of YALAN international group which is located in Guangzhou,china.it is the production center of YALAN international group branding products,specialized in the development and production if cosmetics.ATHENA is qualified for import and export.
controls the quality in full accordance with GMP, ISO9001 quality system good manufacturing practice for cosmetics and key points of cosmetic production licence inspection.

body soap,body whitening soap,body soap bar,cheap body soap

Athena(Guangzhou) Cosmetics Manufacturer Co., Ltd , https://www.ydgzathenaskincare.com