Video Coding Standards and Algorithms — Evolution & Patent Analysis
What is Video Coding Format?
Video coding format is defined as a format for storing and transmitting digital video content (such as in a data file or bitstream) from one system to the other system. It usually employs a standardized video compression technique, which is based on discrete cosine transform (DCT) coding and motion correction. For this purpose, a certain set of video coding specifications/documents that specify and store the technical details of these formats are available. These documents are accepted as technical standards by standardization groups, including International Standards Organization (ISO) and the International Telecommunications Union (ITU), and are hence referred to as video coding standards. In the past, a number of video coding formats have been documented and further standardized, including H.120, H.261, Motion JPEG (MJPEG), MPEG-1 Part 2, H.262 / MPEG-2 Part 2 (MPEG-2 Video), DV, H.263, MPEG-4 Part 2 (MPEG-4 Visual), Motion JPEG 2000 (MJ2), Advanced Video Coding (H.264 / MPEG-4 AVC), Theora, VC-1, and Apple ProRes. Further, these video coding standards are further classified into three key groups which include DPCM (Differential pulse-code modulation), DCT (Discrete cosine transform), and DWT (Discrete wavelet transform) (Discrete wavelet transform).
Video Coding Standards: A brief summary
Starting from the 1980s, a number of video coding standards have been introduced. The table presents the brief of such video coding standards along with key remarks.
Algorithms Used for Video Coding Standards
In the past, a number of video coding standards have been published, which include H.120, H.261, MPEG-1 Part 2, MPEG-2 Part 2 (MPEG-2 Video), H.263, Motion JPEG 2000 (MJ2), H.264 / MPEG-4 AVC, Theora, H.265, AV1, and Versatile Video Coding (VVC / H.266). These standards follow a certain algorithm and thus are classified into different categories.
Algorithm wise Video coding standards
1. DCT (Discrete Cosine Transform):
The source image is divided into 8x8 pixel blocks. Further, the DCT is applied to each block from left to right and top to bottom. In response to this transformation, all block elements are compressed and then quantized by dividing by a few specific costs. The array of compressed blocks that represents the photo is saved in a much less amount of space. A DCT function returns a DCT coefficient matrix, including information in the frequency domain. The DCT coefficients are then quantized by dividing by a quantization matrix to reduce garage space. The block length cost also has an impact on the pleasure and compression ratio.
2. DWT (Discrete Wavelet Transform):
DWT is simple to implement, reduces computing time, and eliminates irrelevant source facts. To create 2D wavelets, the picture is divided into sections with high frequency and low frequency runs. The two subsequent sub images contain both high- and low-frequency vertical statistics. Further, each sub pixel is vertically convolved with the wavelet and the scaling characteristic, resulting in two new separations. Thus, a single-stage wavelet transformation consists of a filtering operation that can decompose a 2D signal into four frequency bands. The generalized block diagram for DWT image/video compression/De-compression shows the source image divided into multiple frames followed by DWT transformation, quantization, encoding and outputting of the compressed frames.
Patent Analysis
The patent data in this article shows information related to video coding standards, including the patent filing trend across the globe and the top-rated assignees.
The number of applications filed each year across the world. It is exciting to know that the patent filing trend jumped to a new level of more than 1000 applications in the year 2019–2021. However, in upcoming times, it is expected to grow as the research and development in this field are still ongoing. Apart from the top companies, many other companies are also indulged in the research process, including Canon, Samsung, Nokia, Huawei, Qualcomm, LG, etc. Henceforth, the trend in patent filing is expected to rise to a new level in the upcoming years.
The top assignees in the field of video coding are presented. Out of all, Sony, with a total number of 2546 patents, holds the majority of shares, followed by Canon and LG Electronics. The other top companies/assignees that contribute to this area of research includes Qualcomm, Samsung, Huawei, Panasonic, Sharp, Intel, Toshiba, etc. Out of many other companies, the top most companies are working with the modified DCT algorithm and the advance video encoders. Such advancement has led to technological development in view of enhanced user experience for existing video services and appropriate performance levels for new media services over 5G networks. Besides the major US companies, a number of top companies like Alibaba, Kuaishou and Hikvision, as well as Korea-based Wilus Group are also working on the development of similar technology.
Conclusion
The future scope of video standards is very bright in view of the upcoming and trending technology. This includes the deep-learning based video coding, such as reducing the compression complexity, and power consumption. Thus, this will allow higher efficiency and better output. Apart from this, the other key elements related to video coding standards include — visual quality assessment, especially in view of PSNR (Peak Signal-to-Noise Ratio), Artificial intelligence-based encoding algorithms, Hybrid encoding/decoding algorithms, etc.
Besides this, at the international level, the video encoder market is estimated to reach from $2.3B to $3.3B USD by 2027, with a CAGR of 7.6% from 2022 to 2027. This is due to the introduction of high-efficiency video coding standards for video encoding, the simplicity of connecting analogue cameras employing video encoders to a network, and the usage of cloud services to store vast amounts of data have all contributed to the market’s rise.
References
- https://einstein.informatik.uni-oldenburg.de/rechnernetze/dpcm.htm
- https://www.tutorialspoint.com/digital_communication/digital_communication_differential_pcm.htm
- https://mil.ufl.edu/nechyba/www/eel6562/course_materials/t5.wavelets/intro_dwt.pdf
- http://chenlab.ece.cornell.edu/Publication/Deepak/bookchap.pdf
- https://towardsdatascience.com/the-overview-of-video-coding-technique-f650aed6c6a4
- F. Moreno and D. Aledo, “The DLMT hardware implementation. A comparative study with the DCT and the DWT,” IECON 2012–38th Annual Conference on IEEE Industrial Electronics Society, 2012, pp. 1591–1596, doi: 10.1109/IECON.2012.6388531
- F. Moreno and D. Aledo, “ Comparative Analysis between DCT & DWT Techniques of Image Compression,” Journal of Information Engineering and Applications, Vol. 1, №2, 2011, pp. 9–17.
- https://www.globenewswire.com/news-release/2022/08/17/2500017/0/en/The-video-encoder-market-is-estimated-to-reach-USD-3-3-billion-by-2027-from-USD-2-3-billion-in-2022-with-a-CAGR-of-7-6-from-2022-to-2027.html
This blog was originally published here:- https://www.copperpodip.com/post/different-types-of-video-coding-standards-and-algorithms-used-over-the-time
