Blog

H.265 COMPRESSION CAMS; THE NEW PHASE OF DVR COMPRESSION

For years, the most efficient DVR compressor available was the H.264. The H264 proved to be an excellent tool for recording surveillance videos. Most of the high quality videos especially those that we stream online were done with the H264 compression. The H264 works by comparing the motion of the encoded image vs the stationary objects, keeping the stationary object in a frame to serve as the reference while it records the motion of the moving object.

The H264, advanced video encoding provides high quality videos at a much lower bit rate than prior standard video protocols like the MPEG-2 and MPEG-4 part 2. The H264 unfortunately does not have a standardized format so many companies have to hold patents for their security camera system, leading to separate generic files for the H264 compressed videos.

The infinite truth is that with improvements in technology, older gadgets and systems will fade away. That fate befell the H264; It is being replaced by the H265 or High efficiency video coding and MPEG-H part 2 which has upgraded the standard. Both codecs work by comparing different parts of a video frame to sift the static and the active frames, then replacing them with a short info that describes the original pixels. What then are the advantages of the H265?

SUPERIOR COMPRESSION

In the compression lies one of the key differences between the HVEC and the H264. The H265 has the ability to expand the size of these areas into bigger or smaller blocks called CTU (Coding tree units). The HVEC can get to compression levels of almost twice that of the H264. It can cover sizes from 4 by 4 to 64 by 64 compared to the H264 that covers just 16 by 16 as its maximum block size.

This improved CTU segmentation, a better motion compensation and glaring prediction need a higher level of signal processing ability for video compression, but then it has a lower impact on the amount of computation needed for decompression. With the motion compensation prediction, the HEVC can reference blocks of pixels to another area within the same frame (intra prediction) or in a different frame (inter prediction).

IMPROVED INTRAFRAME PREDICTION

Video compression depends on predicting motion between frames. Without a change in pixel, a video codec will rather reference the pixel than reproducing it as this saves space. Thus improved motion prediction allows for a more efficient file size and better compression quality. The HVEC also allows single frames of video to be compressed more effectively.

To do this, it describes its pixels with a maths function instead of pixel values. The use of these maths functions means that there is less space used and a further reduction in the file size. All these are possible because of the H265’s intra-frame prediction function that allows for a whopping 33 directions of motions compared to just 9 directions for H264. This brings another advantage, as with the HVEC, a video with about the same visual quality will take up just half of the space that is required by the H264.

MULTIPLE CONCURRENT PROCESSING

The new standard uses slices and tiles, which can be decoded independently from the rest of the frame. With this change, the decoding process can be divided across multiple threads that are run concurrently, taking advantage of more effective decoding opportunities that are available on multi-core processors. As video resolutions continue to improve, there is a need for such high levels of efficiency to decode a video at a streamable pace on lower-end devices.

H265 supports advanced extensions like AVX/AVX2 and FMA3/FMA4. Parallel processing is enabled as the individual rectangular regions that split the image are independent. There is also the Wavefront parallel processing (WPP) feature that the H265 possesses that the H264 does not.  This decision tree allows more productive and effective compression.

HARDWARE SUPPORT

The H265 codec boasts a far more superior hardware support than its predecessor. The latest Intel processors support it. These processors have special instruction sets for encoding and decoding H265 videos, and it is expected that newer models will follow suit. This means that the HVEC has huge advantage over other high resolution video codecs with respect to speed and consistency. When you also add the fact that its predecessor was very popular with the hardware developers, it is easy to see why there is a continuation of the effective support for the H265.

With the significant difference in computational overhead required to encode and decode high-resolution H265 videos, there could be a major support drift between hardware and software supported implementations for the H265.

SOFTWARE SUPPORT

The support is not restricted to Intel hardware alone. H265 videos can be encoded at up to 8k UHD which is the equivalent of 8192 pixels by 4320 pixels. There has been further improvement on the bit reduction of the bit levels with the H265. There is a reduction to 57% for 1080p and 64% for UHD or 4K videos relative to its predecessor. To put how far they have gone in improving the codec, only a handful of cameras currently produce 8k videos, and only a few monitors can have an output of such high resolution. H265 has gone steps ahead and with the prominence of HD in today’s standard systems, 4k and then 8k videos should eventually become very popular.

GENERAL

The longevity of the H265 and its build to last for newer models stands it out. It allows users to record very high-resolution videos surveillance at higher frames. This means that videos can be stored for longer retention times and this is at a lower system cost.

To make the most of the qualities of the H265 codec, it is important to use GPU based hardware accelerated video surveillance NVR’s.