In-depth analysis of video compression algorithm for hard disk recorders
A video encoding and decoding chip
The video codec chips currently used in the past and presently in the DVR field mainly include:
1. ASIC (real hard compression): The chip is a dedicated video codec chip, compression and decompression technology code is fixed in the chip, before the advent of the MPEG-4 era, video compression is mainly based on ASIC However, due to the long cycle of ASIC technology products and the inability to apply the rapid progress of algorithms, ASIC chip products are becoming less and less.
2. DSP (soft compression in principle): DSP is a high-speed signal processor IC, DVR developers in the development of the DSP platform has the following advantages:
First, the user has more freedom in development, supports multiple personalized developments, and can meet the new requirements constantly put forward by the market. It can improve product performance and enhance product competitiveness at the first time.
Secondly, the DSP has strong processing power, and can simultaneously realize the compression processing of multiple channels of audio and video signals on one DSP. At the same time, in order to meet the needs of applications in a timely manner, it also provides many video-specific functions, which makes it possible to further reduce the cost of the product. ;
Third, the development cycle is short, to achieve rapid technical updates and product replacement;
Disadvantages of DSP: Due to the limitations of the DSP architecture, it is very difficult to implement low-cost real-time compression in the face of highly-complex H.264 algorithms.
Second, the compression algorithm
There are many video compression algorithms such as MPEG-1, MJPEG, wavelet algorithm, MPEG-2, MPEG4, H.261, H.263, H.264 and H.265. In the DVR field, the compression algorithms used by DVR mainly include MJPEG. , wavelet algorithm, MPEG-4, H.264, I will briefly introduce it.
1) MJPEG
The MJPEG (MotionJPEG) compression technology standard is derived from the JPEG image compression technology and is a simple intra-frame JPEG compression. It compresses each frame of the video with a small compression ratio and a large number, usually 325×288 resolution per hour per channel. Video recording requires hard disk space of about 1G, which not only consumes large hard disks, but also is difficult to use for network video surveillance. Unless real-time is used, the advantage of this compression algorithm is better video quality. Each frame of picture quality is better when remote monitoring, and some embedded DVR or DVR cards still use this compression algorithm. Such as South Korea's PIC's PICASO video surveillance card
2) Wavelet algorithm
The wavelet algorithm is a video compression based on wavelet transform. The technique is to make the time domain resolution and frequency domain resolution of the image signal reach the highest at the same time. The kernel uses on-the-fly compression and decompression. There is no correlation between frames in the video. Its compression ratio can reach 70:1 or higher. Compression complexity is about 3 times that of JPEG. It is suitable for compressing high-definition large-screen movies. , But hard disk consumption is too large, with 352X288 video, each channel is generally about 350M per hour, the network transmission effect is not good, the current market using wavelet algorithm board Korean KODICOM card, this type of card by MPEG-4 hard disk video The impact of the card also tends to be eliminated.
3) MPEG-4
The MPEG-4 standard is an object-oriented compression method. Instead of simply separating the image into blocks like MPEG-1 and MPEG-2, the objects (objects, characters, and background) are separated from each other according to the image content. Perform intraframe and interframe coding compression, and allow flexible allocation of code rates between different objects, allocate more bytes to important objects, allocate fewer bytes to secondary objects, and greatly improve compression Than to make it get better results at lower bit rates. The transmission rate of MPEG-4 is 4.8~64kbit/s. The storage space occupied by the MPEG-4 is relatively small. The average 352X288 video recording is about 150M per hour, and the quality of MPEG-4 video recording from VCD to DVR can be Tone.
It is worth noting that the same is the MPEG-4 algorithm. There is a big difference between the companies. As far as I know, there are Microsoft (MS) MPEG-4, DIVX MEPG-4, and XVID MPEG-4. Each DVR company has completely adopted the aforementioned MPEG-4, and some of its own MPEG-4 algorithm optimization for DVR, which MSMPEG-4 kernel is not open to the public, can only be applied with WindowsMedia streaming media technology As early as in 2004 Shenzhen wave grain had used MSMPGE-4, due to the technology closed, this MPEG-4 application has been limited, wave grain has given up using this MEPG-4, and later video coding experts and hackers to crack Microsoft's MPEG-4, and made some improvements, came out with DIVX 3.11 MPEG-4, and soon DIVX MPEG-4 was red, almost became the industry standard. Rota, one of the creators of DivX technology, also applied for DivX. Legalized, the original intention of DivX was to get rid of Microsoft's technology blockade and open source, attracting a lot of software video experts to participate. Later, DXN suddenly closed the DivX source code, released its own product DivX4, after another video expert group based on OPENDIVX, developed a new MPEG-4 encoded XVID, its name is almost the opposite of DIVX, this name is Represents this video expert group's anger and hatred toward DXN. Technically speaking, Xvid is basically close to the latest version of DXN's DivX5, or beyond, and it can greatly improve the quality of DivX5. Compression time is considered to be the fastest MPEG4 encoding at the moment. In addition, Xvid has also learned from the past and published according to the GPL, which means that it is illegal for anyone to make a product without open source.
4) H.264
Prior to the adoption of the H.264 video coding standard, it was called International Standard 14496-10 (MEPG-4 Part 10). Therefore, H.264 is not as advanced as MPEG-4. They are actually the same thing. The previously used MPEG-4 is a standard version of the previous MEPG-4 part 10. The current name relationship is:
H.264BaselineProfile corresponds to MPEG-4SP
H.264MainProfile corresponds to MPEG-4ASP
H.264 ExtendedProfile corresponds to MPEG-4 ARTSorFGS
H.264BaselineProfile corresponds to MPEG-4 Studio
As the latest video coding technology, H.264, has a very big prospect in the security industry. The H.264 standard adopts many new technologies to increase the compression ratio and reduce the code stream, mainly using high-precision, multi-mode prediction techniques.
The H.264 standard has designed the video coding layer VCL and the network abstraction layer NAL structure for the needs of network transmission. The network abstraction layer provides a "network friendly" interface, so that the video coding layer can be effectively applied in various systems. The H.264 standard is designed to address the needs of network transmission. The tool for error elimination is designed to facilitate the transmission of compressed video in error and packet loss environments, thus ensuring the effectiveness of video transmission.
The H.264 algorithm is divided into four layers: BaselineProfile, MainProfile, ExtendedProfile, and HighProfile. Compared with the previous algorithm, the H.264BaselineProfile is about twice as large as the previous MPEG-4 in terms of algorithm complexity, and H.264MainProfile is MPEG. Above 10 times of -4, HighProfile has the highest complexity; in terms of compressed bitstream, BaseProfile is lower than MPEG-410-20%, and MainProfile is lower than half of MPEG-4.
As you can see, implementing the H.264 MainProfile algorithm can reduce the video code stream by half, and the CIF format video can be continuously transmitted on real-time ADSL in real time. Even if the network has fluctuations, it will not be affected. At the same time, because the H.264 algorithm has a network coding layer and a network transmission layer structure, the adaptability for ordinary broadband networks and wireless networks is greatly enhanced, and the quality of transmission will not be affected due to the error of the network.
However, the current chip processing capability cannot implement a complete H.264 algorithm. Now only part of the functions in the H.264 algorithm can be implemented. Therefore, it takes a process to experience the advantages of H.264.
The video codec chips currently used in the past and presently in the DVR field mainly include:
1. ASIC (real hard compression): The chip is a dedicated video codec chip, compression and decompression technology code is fixed in the chip, before the advent of the MPEG-4 era, video compression is mainly based on ASIC However, due to the long cycle of ASIC technology products and the inability to apply the rapid progress of algorithms, ASIC chip products are becoming less and less.
2. DSP (soft compression in principle): DSP is a high-speed signal processor IC, DVR developers in the development of the DSP platform has the following advantages:
First, the user has more freedom in development, supports multiple personalized developments, and can meet the new requirements constantly put forward by the market. It can improve product performance and enhance product competitiveness at the first time.
Secondly, the DSP has strong processing power, and can simultaneously realize the compression processing of multiple channels of audio and video signals on one DSP. At the same time, in order to meet the needs of applications in a timely manner, it also provides many video-specific functions, which makes it possible to further reduce the cost of the product. ;
Third, the development cycle is short, to achieve rapid technical updates and product replacement;
Disadvantages of DSP: Due to the limitations of the DSP architecture, it is very difficult to implement low-cost real-time compression in the face of highly-complex H.264 algorithms.
Second, the compression algorithm
There are many video compression algorithms such as MPEG-1, MJPEG, wavelet algorithm, MPEG-2, MPEG4, H.261, H.263, H.264 and H.265. In the DVR field, the compression algorithms used by DVR mainly include MJPEG. , wavelet algorithm, MPEG-4, H.264, I will briefly introduce it.
1) MJPEG
The MJPEG (MotionJPEG) compression technology standard is derived from the JPEG image compression technology and is a simple intra-frame JPEG compression. It compresses each frame of the video with a small compression ratio and a large number, usually 325×288 resolution per hour per channel. Video recording requires hard disk space of about 1G, which not only consumes large hard disks, but also is difficult to use for network video surveillance. Unless real-time is used, the advantage of this compression algorithm is better video quality. Each frame of picture quality is better when remote monitoring, and some embedded DVR or DVR cards still use this compression algorithm. Such as South Korea's PIC's PICASO video surveillance card
2) Wavelet algorithm
The wavelet algorithm is a video compression based on wavelet transform. The technique is to make the time domain resolution and frequency domain resolution of the image signal reach the highest at the same time. The kernel uses on-the-fly compression and decompression. There is no correlation between frames in the video. Its compression ratio can reach 70:1 or higher. Compression complexity is about 3 times that of JPEG. It is suitable for compressing high-definition large-screen movies. , But hard disk consumption is too large, with 352X288 video, each channel is generally about 350M per hour, the network transmission effect is not good, the current market using wavelet algorithm board Korean KODICOM card, this type of card by MPEG-4 hard disk video The impact of the card also tends to be eliminated.
3) MPEG-4
The MPEG-4 standard is an object-oriented compression method. Instead of simply separating the image into blocks like MPEG-1 and MPEG-2, the objects (objects, characters, and background) are separated from each other according to the image content. Perform intraframe and interframe coding compression, and allow flexible allocation of code rates between different objects, allocate more bytes to important objects, allocate fewer bytes to secondary objects, and greatly improve compression Than to make it get better results at lower bit rates. The transmission rate of MPEG-4 is 4.8~64kbit/s. The storage space occupied by the MPEG-4 is relatively small. The average 352X288 video recording is about 150M per hour, and the quality of MPEG-4 video recording from VCD to DVR can be Tone.
It is worth noting that the same is the MPEG-4 algorithm. There is a big difference between the companies. As far as I know, there are Microsoft (MS) MPEG-4, DIVX MEPG-4, and XVID MPEG-4. Each DVR company has completely adopted the aforementioned MPEG-4, and some of its own MPEG-4 algorithm optimization for DVR, which MSMPEG-4 kernel is not open to the public, can only be applied with WindowsMedia streaming media technology As early as in 2004 Shenzhen wave grain had used MSMPGE-4, due to the technology closed, this MPEG-4 application has been limited, wave grain has given up using this MEPG-4, and later video coding experts and hackers to crack Microsoft's MPEG-4, and made some improvements, came out with DIVX 3.11 MPEG-4, and soon DIVX MPEG-4 was red, almost became the industry standard. Rota, one of the creators of DivX technology, also applied for DivX. Legalized, the original intention of DivX was to get rid of Microsoft's technology blockade and open source, attracting a lot of software video experts to participate. Later, DXN suddenly closed the DivX source code, released its own product DivX4, after another video expert group based on OPENDIVX, developed a new MPEG-4 encoded XVID, its name is almost the opposite of DIVX, this name is Represents this video expert group's anger and hatred toward DXN. Technically speaking, Xvid is basically close to the latest version of DXN's DivX5, or beyond, and it can greatly improve the quality of DivX5. Compression time is considered to be the fastest MPEG4 encoding at the moment. In addition, Xvid has also learned from the past and published according to the GPL, which means that it is illegal for anyone to make a product without open source.
4) H.264
Prior to the adoption of the H.264 video coding standard, it was called International Standard 14496-10 (MEPG-4 Part 10). Therefore, H.264 is not as advanced as MPEG-4. They are actually the same thing. The previously used MPEG-4 is a standard version of the previous MEPG-4 part 10. The current name relationship is:
H.264BaselineProfile corresponds to MPEG-4SP
H.264MainProfile corresponds to MPEG-4ASP
H.264 ExtendedProfile corresponds to MPEG-4 ARTSorFGS
H.264BaselineProfile corresponds to MPEG-4 Studio
As the latest video coding technology, H.264, has a very big prospect in the security industry. The H.264 standard adopts many new technologies to increase the compression ratio and reduce the code stream, mainly using high-precision, multi-mode prediction techniques.
The H.264 standard has designed the video coding layer VCL and the network abstraction layer NAL structure for the needs of network transmission. The network abstraction layer provides a "network friendly" interface, so that the video coding layer can be effectively applied in various systems. The H.264 standard is designed to address the needs of network transmission. The tool for error elimination is designed to facilitate the transmission of compressed video in error and packet loss environments, thus ensuring the effectiveness of video transmission.
The H.264 algorithm is divided into four layers: BaselineProfile, MainProfile, ExtendedProfile, and HighProfile. Compared with the previous algorithm, the H.264BaselineProfile is about twice as large as the previous MPEG-4 in terms of algorithm complexity, and H.264MainProfile is MPEG. Above 10 times of -4, HighProfile has the highest complexity; in terms of compressed bitstream, BaseProfile is lower than MPEG-410-20%, and MainProfile is lower than half of MPEG-4.
As you can see, implementing the H.264 MainProfile algorithm can reduce the video code stream by half, and the CIF format video can be continuously transmitted on real-time ADSL in real time. Even if the network has fluctuations, it will not be affected. At the same time, because the H.264 algorithm has a network coding layer and a network transmission layer structure, the adaptability for ordinary broadband networks and wireless networks is greatly enhanced, and the quality of transmission will not be affected due to the error of the network.
However, the current chip processing capability cannot implement a complete H.264 algorithm. Now only part of the functions in the H.264 algorithm can be implemented. Therefore, it takes a process to experience the advantages of H.264.
Main functions:
Rear wash, feminine wash, massage wash, kid wash, Seat heating, warm air drying, Soft close, seat sensor, night light, Water temp adjust, air temp adjust, seat temp adjust, nozzle adjust, Instant water heating, IPX4 water proof, Side control panel, remote control, Energy saving mode
Extra functions:
automatic open/ close, automatic flushing, with Super flushing system by own research and development
Intelligent Items,Smart Cover,Automatic Toilet,Intelligent Bidet
XIAMEN JIEENTE IMPORT AND EXPORT CO.,LTD , https://www.xiamenjieente.com