PowerVR video IP solutions for wireless displays

After coming back from MWC 2013 in Barcelona where we’ve launched the PowerVR G6100 – our most smallest PowerVR Series6 GPU to date -, we’re dedicating this entry to a ‘behind the scenes’ look at the other important announcement made during the show. We’ve stated in a recent press release that power efficiency is now the dominant factor in mobile design. This also applies to our PowerVR video cores, especially since our partners have now shipped 500m devices cumulatively incorporating our video SoC IP cores out of over 1 billion devices which have integrated Imagination’s IP, many of these containing multiple cores from Imagination.

The growth in devices using Imagination’s video technologies shows the momentum that continues to build in the mobile and embedded market for new full-HD video technologies. The PowerVR video hardware IP cores are able to form the backbone of many new broadcast standards like Miracast, the wireless display protocol from the Wi-Fi Alliance. According to IHS iSuppli, Miracast-enabled devices would top 1.5bn units in 2016 alone. To address this we’ve designed our PowerVR Series3 and Series4 video decoders and encoders to deliver multi-stream, multi-standard video transcoding, supporting high definition resolutions like 4K and beyond as well as increased image quality with 4:4:4 colour resolutions and 10-bit colour depth.

PowerVR global shipments WiFI Miracast Video

Video codec requirements usually vary from standard to standard, therefore we’ve made sure all of Imagination’s video cores support all major video and imaging specifications including H.264 (Base to High profile), MPEG4, MPEG2, VC-1/WMV, AVS, Sorenson and JPEG. Many of our decoder cores also offer hardware rotation and scaling, delivering significant reductions in bandwidth and power consumption. By combining them with our PowerVR graphics cores, developers can write more efficient software that takes advantage of our heterogeneous processing capabilities. Scaling can thus be done by the graphics core, making the combination of VPUs and GPUs extremely powerful for a wide range of multimedia applications.

Imagination’s PowerVR video IP meets all requirements for wireless displays

We’ve built a dynamic and complete roadmap to address all multimedia processing needs, offering the best hardware IP technologies for all wireless display applications. Configurable multi-pipe and multi-core implementations allow system designers to choose the right combination of performance, features and area while always keeping the power consumption at a minimum. Here’s a brief list of the mandatory features each system for transmitting video wirelessly should have.

PowerVR Video Decoder Requirements for WiDi Miracast

PowerVR Video Encoder Requirements for WiDi Miracast

Notice how we’ve optimized our encoder lineup to sustain high bitrates and framerates, improved image quality and futureproofed our cores for upcoming trends like 4K @ 60fps.

High quality encoding is essential. Most consumers can get confused by advanced technical terms like flexible bitrate control and colour subsampling but will be able to spot any loss in quality very quickly. By integrating Imagination’s video cores, any platform can offer improved video quality at lower bitrates. Features such as 8×8 blocks for reference frames and CABAC encoding enable video transmission at the highest possible quality, whilst maintaining the low data rates essential for network transmissions.

Move away from “spec sheet” video encoders and decoders

Furthermore, any encoder needs to have all its elements implemented efficiently so that they all work together in harmony. It is not enough to have a “spec sheet” encoder simply stating H.264 compliance, where performance and features are treated separately. We instead have taken the path of building an efficient architecture based on a combination of optimized hardware, firmware and software for power efficiency and performance. For example, Imagination’s advanced search algorithms allow a ‘no-compromise’ approach for large, bandwidth efficient and accurate searches, implementing most of the functionality in hardware and requiring minimum CPU involvement, keeping the central processor usage typically under one MIPS.

PowerVRVPU PowerVR video encoder decoder roadmap

Thanks to a long and successful collaboration with numerous customers, we’ve developed a flexible, scalable design deployed in hundreds of millions of devices that are used daily instead of just theoretical systems tested under laboratory conditions and carefully controlled code. We’ve also made drivers available for several operating systems as well as low level APIs for easy porting to any new or existing OS.

PowerVR video IP encoders and decoders fully support 4K resolutions and beyond

4K resolutions increase the requirement for efficient compression, as wireless displays demand compression when used over limited channel. The same principles apply for our PowerVR decode technologies, where we have been the only IP provider able to offer 4K @ 60fps working at the required high bitrates demanded by wireless displays. Even in extreme bitrate conditions of 230 Mbps for 4096 x 1744 (2.35:1), our cores still deliver superior image quality without sacrificing the frame rate. We’ve also integrated advanced error handling and concealment algorithms to optimize our solutions especially for displays which are sensitive to transmit issues. Not only that, but this rapidly growing market area will need cores that support multiple standard, multiple stream capabilities as device manufacturers will always want to have the latest and greatest bundle of multimedia options integrated in their products.

The most recent additions to the PowerVR video family, the D4500MP and E4500MP, maintain quality throughout the processing chain and support 4:4:4 and 4:2:2 colour resolutions which are essential for high quality wireless displays as most of them require conversion from RGB to YCbCr. A simple case of mixing video and graphics like in OSD (on-screen display) will lead to poor results without 4:4:4 encoding, therefore increasing colour depth and resolution for the Series4 video cores was not only a matter of offering our partners a path to differentiation but also of further increasing efficiency, as most internal video processing is typically 10bit.

PowerVR video VPU family overview

The benefits above are mostly related to picture quality but perhaps the biggest gain from using our cores is seen in the overall system power consumption. By offloading the more intensive multimedia processing to the PowerVR video decode/encode cores, the host CPU can refocus on other tasks or even enter its idle state, thus reducing power consumption, maximizing battery life and increasing the responsiveness of the computing platform.

Imagination naturally becomes the best choice for wireless displays. Our PowerVR video cores exceed the performance and power requirements for wireless display, providing leading transcoding solutions which ensure platforms using our IP deliver the best quality for high resolutions at superior bitrates.

Stay up-to-date with the latest announcements from the PowerVR video team, follow us on Twitter (@ImaginationPR) and keep coming back to our blog.

, , , , , , , , , , ,

  • Tangey

    In an article which sets the premise that power efficiency is now the dominant factor in mobile designs, I find it strange that “low power” is glaringly completely missing from the list of requirements in the video encoder table, whereas it is present in the video decoder table with IMG solutions cited as having the “lowest power in the industry”

    It begs the question. Is low power not a dominant factor in Mobile design video encoding, which would be amazing given the compulsory hi res rear camera fitted to mobiles today, and the emphasis placed on video recording opportunities, or was it left out of the table due to IMG not being class leading in this area ?

  • alexvoica

    Hi,

    That’s a very interesting point. All our technologies, including PowerVR graphics and video cores, are intrinsically low power. On mobile platforms, decoding is typically used more often than encoding, but this does not mean that all our PowerVR hardware encode IP has not been optimized for low power.

    When encoding video streams, maintaining high compression ratios without reduced visual quality is a key element for most users.

    However, providing the industry’s best encoder is a balance between high quality (we are talking Full or Ultra HD resolutions) at low bandwidths and optimized power consumption. All our encoder cores have an automatic hardware power management system that uniquely allows them to provide a flexibile way of controlling parameters like bitrate or header generation. What this esentially means for mobile platforms (smartphones, tablets) is that you can still capture 1080p video at the best possible quality while also expecting the lowest battery usage.

    For example, Intel has been integrating both our PowerVR VXD and VXE cores across all their mobile-oriented Atom platforms (Lexington, Medfield, Clover Trail, Clover Trail+). In their keynote at CES 2013, one of the main points of their presentation was how, by adding our low-power, hardware video cores to their Atom SoCs, they have been able to offer smartphone and tablet users an extended multimedia experience, by prolonging the battery life of all platforms that include PowerVR video cores. This is an esential point especially for emerging markets where people rely on their mobile devices as their main computing device.

    Regards,
    Alex.