Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage News New H.266 Video Coding Standard Claims to Be 50% More Efficient Than H.265

New H.266 Video Coding Standard Claims to Be 50% More Efficient Than H.265

This item in japanese

Versatile Video Coding, also known as H.266, MPEG-I Part 3, and Future Video Coding (FVC), is the successor to H.265 and promises to reduce data requirements by 50% while keeping the same level of visual quality as its predecessor.

The previous standard H.265/HEVC requires ca. 10 gigabytes of data to transmit a 90-min UHD video. With this new technology, only 5 gigabytes of data are required to achieve the same quality.

According to Fraunhofer HHI, which has been working on the new standard for several years along with partners such as Apple, Ericsson, Intel, Huawei, Microsoft, Qualcomm and Sony, H.266 should be especially beneficial to 4K and 8K video streaming. The new standard is deemed ideal for any kind of video, though, from SD to 8K and including high dynamic range video and omnidirectional 360° video, says Fraunhofer HHI.

Some concerns about the new standard come from the chosen licensing model. The previous standard, High Efficiency Video Coding, also know as H.265, was released in 2013 and found wide adoption only in the last two years, mostly due to patent-related concerns. This brought some unclarity about royalties and made content providers stick with previous standard H.264, in spite of the fact H.265 offered doubled performance. In an attempt to reduce this risk for H.266, licensing for H.266 is handled through a new group called Media Coding Industry Forum, but it is not clear whether this effort will succeed in easing adoption for potential implementor as far as patent issues are concerned. Additionally, since MC-IF does not control the standardization process, this could further hamper H.266's adoption process.

Advances related to video compression seems to happen at a fast pace, with the new standard coming when the transition to the current one, H.265, is not complete yet. As several Hacker News readers pointed out, this comes down in part to the evolution of playback hardware:

Some of the advances in H.265 depended on players having enough RAM to hold a bunch of previous decoded frames, so they could be referred to by later compressed data.

Another factor opening the way to new encoding schemes is the change in video resolutions. In short, each encoding scheme is optimized for a given set of resolutions, those that are common when the codec is released. This usually means its efficiency is reduced when used with larger resolutions. In particular, one design choice, the size of macroblocks, which is the coarsest block that is compressed, greatly affects the efficiency of the codec at a given resolution. So, H.264, having a smaller macroblock size than H.265, is less efficient for 4K video.

Such factors sum up with algorithmic improvements to determine the overall efficiency improvement.

For all interested parties, Fraunhofer HHI plans to publish its own encoding/decoding software implementing H.266 this Autumn. Hardware support from chip-makers has not been announced yet and will take seemingly longer.

Rate this Article