Technological and internet evolution
In recent years, the real “internet revolution” has taken place in quantity and type of traffic exchanged, in particular due to new multimedia contents. The traffic explosion generated by mobile devices (especially by apps, social media above all) and on demand video content have further increased this trend. This also under the general pressure on data traffic caused by the advent of the pandemic and the strong increase in applications such as video conferencing, gaming and others.
In fact, today most of the internet traffic concerns the use of video content. Just think that last year the share of world internet traffic “consumed” by downstream video reached two-thirds of the total. Of these, a significant portion (about 50%) is handled by a few widespread services. Netflix alone (with 15%), YouTube (12%), Disney+ and Prime Video account for a third of the traffic.
A contribution to this evolution has undoubtedly been the diffusion of higher bandwidth internet connections, with the widespread availability of fiber connections and fast mobile network connections. 4G and 5G have in fact also enabled smartphones and tablets for any multimedia use, indulging the growing habit of interacting digitally or using content remotely.
But that’s not all, the supremacy of video has also been driven by continuous research and growth in terms of quality which, in terms of impact on the generated traffic volumes, has no equal in other applications. Now, in fact, in the digital video world we talk about “ultra-high-resolution” contents, starting from 4K, to 8K, HDR, and 360-degree.
Video encoding new protocols
Being able to support these application scenarios is not just a matter of infrastructure: new protocols are also needed. It’s no coincidence that the research for highest video quality support (without exceeding the bandwidth) is continuous and fervent.
The new H.266/VCC (Versatile Video Coding) standard promises to halve the bandwidth occupied compared to the H.265/HEVC standard, while maintaining the same quality, doing even better than AV1, which however is free. This will make it possible to support the most advanced scenarios such as 8K, 360-degree video, but these are fresh protocols, whose use today is still held back both by royalty costs and by the absence of products ready to directly support the protocol in hardware acceleration.
For this reason, and above all for the different performance and quality requirements compared to the 4K or 8K scenarios, protocols such as H265 and H264 still dominate in areas such as VDI (Virtual Desktop Infrastructure), where both are used, or video conferencing (applications such as Teams or Zoom use H.264 or technologically similar protocols).
Before that, the HEVC/H.265 protocol had surpassed the previous one by allowing a 50% reduction in the space required to store or transmit videos compared to H.264, using a more efficient coding, achieved through more advanced and sophisticated algorithms, maintaining a high level of image quality.
However, the H.264 standard still remains the most popular. In fact, the encoding used by HEVC/H.265 is much more expensive both in encoding and decoding than H.264, requiring much more computation time and more power/battery consumption, with performance problems on medium-low range hardware.
The same process that we observe today in the approach towards the H.266/VCC protocol (up to 7 times more complex to encode, and 2 times more complex to decode) took place between the H.264 and H.265 standards and also previously, with the H.264 codec, starting from 2003, then taking years to spread, until all the various hardware manufacturers have inserted specific decoders for this format.
H.264 vs. H.265
But let’s see now the differences between these two widespread protocols.
Both video compression codecs work by comparing the differences between consecutive frames, identifying those consistent common areas. These image areas are then replaced with information describing the original pixels. The big difference between the two is the different ability to expand these areas (called CTU coding tree units) dynamically, with dimensions ranging from 4×4 to 64×64, while H.264 is limited to a maximum of 16×16. However, the improvement of segmentation and spatial prediction requires more computational capacity for the encoding part, while it does not impact too much on the decoding part. HEVC/H.265 is also able to reference pixels blocks within the same frame (intra prediction) or in different frames (inter prediction).
To give an idea, the new H.266/VCC protocol goes up to 128×128 areas and has a much wider range of models applicable in the inter/intra prediction and partitioning stages. i video 360-gradi i video 360-gradi
Even desktop virtualization systems have always benefited from the improvement of streaming systems. From a technical point of view, the technologies used are very similar. For this reason, the main technology providers, such as Citrix and VMware, have already introduced the possibility of using HEVC/H.265 encoding for desktop streaming for years. The setup requires you to mount a video card capable of supporting this encoding on your server, although, when it comes to enterprise installations, graphics cards that can allow multiple access to their computing power can be expensive.
