How to deliver large amount of video data with limited bandwidth and high perceived quality
Distribution of high resolution video content to a large number of users requires transmission of extremely large amounts of data. This has only been made possible by the application of video codecs that reduce the size of content sufficiently to be stored and distributed on a mass scale.
What we're doing
The video coding project at BBC R&D focuses on H.265/HEVC, a new video compression standard that provides large bit-rate reductions (up to 50%) over its predecessor H.264/AVC.
H.265/HEVC development goes back to January 2010 when ISO/IEC Moving Picture Expert Group (MPEG) and ITU-T Video Coding Expert Group (VCEG) issued a joint call for proposals on improved video compression technology beyond the H.264/AVC standard. The call for proposal targeted high and ultra high definition video and had a successful outcome with twenty-seven responses submitted from both academia and industry. Since then, MPEG and VCEG joined efforts in a partnership called the Joint Collaborative Team on Video Coding (JCT-VC) to develop H.265/HEVC.
As part of the MPEG, the BBC R&D is actively participating in the JCT-VC activities by proposing efficient compression tools particularly devoted to the video broadcasting scenario. BBC R&D is also working with the rest of the BBC and related organisations, such as EBU, to evaluate and define future usage of H.265/HEVC.
More project info
Why it matters
After HD and 3D, consumers are now demanding even higher definition and improved content fidelity. Enabling services for these and many other applications can only be achieved with ground breaking methods to handle the underlying huge amounts of data. This makes coding of high resolution sequences essential to content handing within the BBC.
How it works
Video applications commonly used in broadcasting require high compression ratios. In addition to exploitation of redundancies among video pixels, high compression ratios are achieved by application of lossy coding which discards some additional video details. A drastic reduction of the information inevitably affects the quality of coded videos and may introduce artefacts visible as blurred details and false blocking edges in some image areas. Therefore video compression ultimately addresses the trade-off between reducing the video bit-rate and maintaining the coded quality as close as possible to the uncompressed content.
Typically video codecs exploit two types of redundancies in video content: spatial and temporal. For example, spatial redundancy can be found in homogenous image areas where all pixels have the same value (or roughly speaking the same colour). If for an image region all the pixels share the same value, then we can achieve a good compression by sending to the decoder only the shared value. Temporal redundancy can be seen by looking at a video on a frame-by-frame basis. Between two adjacent frames there are many image areas which remain static. For these areas we can avoid transmitting any information but simply tell the decoder to use the pixel values from the same position in previous frames. Extensions of such observations motivated the development of numerous sophisticated algorithms that are the basis for modern video codecs.
Typically a modern video codec works by predicting small areas (macroblocks, coding units) of the picture, then only transmitting the difference signal between the predicted and actual images. Application of transforms, such as those similar to Discrete Cosine Transform (DCT), enables compact representation of the difference signal which is then quantised and entropy encoded to additionally compress the signal before a coded bit-stream is formed.
H.265/HEVC achieves significantly greater efficiency than H.264/AVC by using a much wider range of coding block sizes and shapes and by enabling many more coding modes based on advanced prediction algorithms.