Video compression has come a long way since the introduction of the first digital transmission, the Integrated Services Digital Network (ISDN), back in the 1980s. Utilising (now long forgotten) H.261, a video compression standard used mostly for video conferencing, it was introduced by the CCIR (Consultative Committee on International Radio) group, which later became known as ITU (International Telecommunication Union).
H.261 worked with CIF size video (352 x 288 pixels) and achieved sufficient good quality for video conferencing, especially when predominantly static people are just talking and hardly moving in a video conference.
Around the same time, near the end of 1980s, personal computers became more popular and a new video compression standard was introduced for converting analogue VHS and S-VHS movies to fit onto a CD media – the MPEG-1. This was proposed by the Motion Pictures Experts Group (MPEG), and the main idea was to advise a video compression that can encode movies with up to 1.5 Mb/s, sufficient streaming speed to be played back from a CD media.
The digitisation of analogue video became seriously attractive after the introduction of of CD media and larger computer hard drives. The MPEG-1 standard was the first attempt to digitise the video industry, predominantly the broadcast and multimedia. Like the H.261, MPEG-1 also worked with CIF size video and achieved sufficient good quality comparable to VHS recorded quality – i.e. up to 240 TV lines.
After the introduction of S-VHS analogue recording some time in the 1990’s, which claimed 400 TV lines of horizontal resolution, the broadcast industry had to come up with a video compression that equals or exceeds the S-VHS. Consequently, around 1993/1994, the MPEG-2 standard was proposed. This was a more advanced form of video compression than MPEG-1, and allowed for much higher picture quality. Instead of saving movies on Betamax and S-VHS video tapes, it became possible to save a full featured movie in a digital format, in MPEG-2 on a newly created DVD media. Cable television was possible, where MPEG-2 streams were used to transmit the content. MPEG-2 was designed to use more than 1.5 Mb/s, although it was backwards compatible with MEPG-1, it could go over 16 Mb/s.
The DVD quality movies were typically encoded with around 4 Mb/s, surpassing VHS and even S- VHS resolution of 400 TV lines. A typical MEPG-2 encoded high quality video was using so-called D1 resolution (or 4CIF) which was designed to offer up to 450 TV lines.
Ten years have passed since the introduction of MPEG-2, and the television industry decided on yet another huge jump – the High Definition TV (HD TV) format.
The HD is a digital video format from the source, rather than being converted from analogue, as was the case with DVD media. The HD format is the current television format and it is also known as 1080HD with 1920 x 1080 pixels. HD offers five times the pixel count of D1 resolution.
When an HD signal is produced by an HD camera, it comes out as 1.5 Gb/s or 3 Gb/s stream, depending on if it is 1080i (interlaced) or 1080p (progressive). This is huge data traffic coming out from one HD camera, impossible to imagine 20 years ago. In order to be able to transmit and store such a huge amount of video data, a new video compression was needed. Although MPEG-2 was flexible enough to cater for HD video format as well, a more efficient video compression was needed. As a result, about 10 years after the introduction of MPEG-2, the Advanced Video Codec (AVC), also known as H.264, was introduced.
The H.264 is the current most popular video compression, used for broadcasting, saving high quality movies on Blue-Ray disks, or recording HD and MP multiple CCTV cameras.
The H.264 offers at least four times the efficiency of the MPEG-2, so that a nice looking HD stream would require around 16 Mb/s with MPEG-2, but the same visual quality can be achieved with only 4 Mb/s using H.264.
AVC or H.264 is still the most popular video compression today, used in broadcasting, storing HD movies on Blue-Ray disks, and certainly in IP CCTV for recording and transmitting multiple HD and MP CCTV cameras.
H.264 offers at least four times the efﬁciency of MPEG-2, so while a nice looking HD stream would require around 16 Mb/s with MPEG-2, the same visual quality can be achieved with H.264 using only 4 Mb/s. In fact, what was a very decent compression for SD D1 video using MPEG-2, at 4Mb/s, it is the same streaming bandwidth of only 4Mb/s with H.264 that achieves a very good video quality for 1080 HD.
Now, another 10 years have passed since the introduction of the H.264.
The latest television advancements are now offering even larger video formats, so-called 4k video, with quadruple the pixel count of HD, i.e. 3840×2160 pixels. 4k is basically equal to live streaming of 8 mega pixel video. It is also known as Ultra-HD-1 resolution. Many broadcast studios, and many production houses, are already using 4k on their movie sets.
An even more impressive format called 8k is being experimented with, offering another quadruple resolution of 7680×4320 pixels, which is almost 32 mega pixels of live streaming video. This is also known as Ultra-HD-2.
When viewing 4k, and 8k video, a viewer sits closer to the display relative to the viewable details and this immerses the visual sensors completely. It is said that the viewing experience is almost three dimensional without having the 3D goggles. This was reported by many viewers watching the London Olympics in 2012 with the experimental 8k video.
The H.264 compression can be applied to 4k video too, but more efficient video compression was sought after. So, in 2013, a new video compression, H.265, also called High Efficiency Video Codec (HEVC), was introduced.
What are the key features of the HEVC/H.265?
First, and most importantly, it is twice as efficient when compared to AVC/H.264. This means, to produce the same visual quality of what H.264 would produce with 4 Mb/s, HEVC/H.265 could produce it with 2 Mb/s.
To put this another way too – with the same stream as H.264, H.265 will produce a video stream twice as nice visually and smoother to watch (if there was a way to measure nice and smooth).
In 2014, a subjective video comparison was conducted by BBC among the students at the University of West Scotland and the following score was produced:
H.265 average bit-rate reduction compared to H.264
Video format => 576i 720p 1080p 4k
HEVC/H.265 52% 56% 62% 64%
This potentially means saving a lot of hard disk space if we were to encode with H.265 but wanted to obtain the same visual quality as what we are doing today with H.264. Of course, in order to do that, the source material needs to be encoded with H.265. Certainly, this also means saving in network bandwidth, not just storage space.
The key technical reason behind the improvement of H.265 over H.264 is in the more complex intra-prediction of moving objects, using partitions of the frame versus macro blocks and allowing up to 64 x 64 pixel blocks.
While HEVC/H.265 increases the compression ratio, at the same time it is also more effective at predicting the details of moving objects, subdividing the compression blocks to quarter size, and managing colours more efficiently. This is necessary because, while the 4k sensors have an increased number of pixels, they may also have an increased number of frames per second (50, 60, 120 or even 240 fps).
Currently, 25 fps (frames per second) is considered as sufﬁcient for “live motion” in CCTV. However, larger displays using 4k resolution (and soon to be released 8k displays) have a more noticeable ﬂicker effect when viewed up close. The ﬂicker effect is the result of the old 24 pictures per second phenomenon from the early days of ﬁlm, known as human eye persistence. The eye persistence effect is more noticeable with our peripheral vision and it is stronger with bigger and brighter displays. In order to minimise the ﬂicker effect on our eyes, displays need to produce more frames per second.
One way to increase the displayed frames per second is to simply duplicate the frames within the TV itself, without necessarily having to capture footage using 50 frames at the camera, you just double the frames of a 25 fps signal. That said, the proper way to reduce the flicker is to increase the frames per second on the camera itself. Eventually, all cameras in the near future will not only have 4k display or more, but they will produce more fps as well. To encode such a high megapixel stream with such a large frame rate, more capable video compression will be required. This certainly has been considered in H.265.
Switching from HD to 4k
When switching to 4k video, with H.265 video compression, your system will use the same bandwidth as with HD video with H.264 video compression. There is no need to update the network or your storage. Of course the cameras and the displays will need to be upgraded from HD to 4k.
Many broadcast studios, and many production houses, are already using 4k on their movie sets. Many CCTV camera manufacturers have already showcased and are now shipping 4k cameras. Some of them are using H.264 video compression for the 4k format, but some have already embedded H.265 encoders inside their 4k cameras.
4k Display screens are already available and they are not much more expensive than their HD predecessors were a few years ago.
How about 8k?
An even bigger and more impressive format known as 8k is being experimented with now.8k offers quadruple again the resolution of 4k equaling 7680×4320 pixels. This is 32mega pixels of live streaming video, also known as Ultra-HD-2.
When viewing 4k, and 8k video, a viewer sits closer to the display relative to the viewable details and this immerses the visual sensors completely. It is said that the viewing experience is almost three dimensional, without having to use 3D goggles. This was reported by many viewers watching the London Olympics in 2012 with the experimental 8k video.
More processing power and better displays
There is a price to pay for such an advancement in video compression. This price is even higher demand for processing power then what was needed for H.264. HEVC technical papers state that H.265 encoders require 3~5 times more processing power than the H.264 encoders. This would typically be done in the silicon (encoder chips) of the new 4k cameras. The H.265 decoders should be a little bit less demanding then the encoders, but would still require 1.5~3 times more processing decoding power than H.264. This is important for anyone in CCTV to understand, as most of the decoding in CCTV is done in the software of the operating system, as is the case today with most HD client stations. This will clearly require even more CPU and GPU computer power, and more efﬁcient viewing software.
Certainly, display quality needs to be adequate as well. When using 4k cameras, one should really use 4k displays as well. While it is still possible to display a 4k feed on a HD display, unless you have digital zooming ability in your client software, having a have 4k cameras on an HD display is almost pointless.
When making the switch to 4k, make sure you have suitable workstations that have sufﬁcient processing power to decode and display 4k video streams smoothly. You need to be even more cautious if you wish to view multiple 4k streams simultaneously.
Vlado Damjanovski is an author, inventor, lecturer, and closed circuit television (CCTV) expert who is well-known within the Australian and international CCTV industry. Vlado has a degree in Electronics Engineering from the University Kiril & Metodij in Skopje (Macedonia), specialising in broadcast television and CCTV.
In 1995, Vlado published his first technical reference book – simply called CCTV, one of the first and complete reference manuals on the subject of CCTV. Now in its 4th edition, and translated into four languages, Vlado’s book is recognised the world over as one of the leading texts on CCTV.
Vlado is the current chairman of the CCTV Standards Sub-Committee of Australia and New Zealand. In his capacity as chief contributor, Vlado has helped create the Australian and New Zealand CCTV Standards (AS4806.1, AS4806.2 and AS4806.3).
He can be contacted through his website www.vidilabs.com