Product differentiation – one of the industry’s most dreaded topics – seems to be making a strong comeback as more and more vendors try to find ways of setting themselves apart to avoid the bottomless pit of price competition.
Vendors need the so-called USPs (unique selling points) within a product that consumers will find attractive and meaningful. These sometimes subtle, yet very important, differences will give companies a market power and help them overcome the Bertrand Paradox for pricing similar products.
Let’s see how this applies to using our market-leading PowerVR video hardware encode and decode IP cores. With these cores customers can build powerful SoCs able to support a wide range of video standards while also bringing overall system power consumption down and significantly offloading the CPU from the intensive task of video transcoding and other video functions.
When looking at our solution, as it is with every standard, a number of valid questions always arise. By adopting it, won’t I have just another ‘me too’ offering that might get lost in the noise of dozens of other solutions? More so, how do I know that my final product will have the unique features and required performance that will beat whatever the competition will be showing off in the near future?
The first step in answering these questions is simple: stop looking at a standard like a check-list, where all of its features come together in a recipe-type manner. Instead, look at it as a foundation on which your product is based, where its flexibility enables you to make changes that suit your needs and those of your customers.
PowerVR video IP and the importance of standards: a case study on H.264
We’ll use one of the widest used video standards at the moment, H.264/MPEG-4 AVC, as an example. This is a block-oriented motion-compensation based codec designed for Blu-ray Discs and internet streaming and also incorporated into several major terrestrial HDTV broadcasting standards (ATSC, DVB-T etc.) Developed by an ITU-T committee, it has an extended number of features which make for more efficient video compression while providing versatility for different applications in network environments. The standard also defines a number of profiles which include subsets of capabilities that are tailored to suit specific areas; the most common are the Baseline Profile (BP), the Main Profile (MP) and the High Profile (HP).
In implementing H.264, we have developed a broad range of video IP to meet and exceed the demands of the market by avoiding the destined-to-fail ‘spec sheet’ approach, where items are simply ticked off a list to meet compatibility. Instead, we’ve focused on key factors like efficient hardware architecture, full implementation of toolsets, a broad software support and an optimal design that incorporates hardware, firmware and software elements. We’ve also integrated support for additional pre- and post-processing and have made sure that we are using the latest benchmarking tools to compare and rate our solutions.
When testing against industry standards, we never settle for the minimum. Pictured below is an example of how using just one evaluation algorithm for quality measure can sometimes be a dangerous path to tread on:
MSE vs. SSIM in measuring image quality
Two indices are used to measure the similarity between the original image and the processed ones. Even though they have the same MSE (mean squared error) value, it is the structural similarity index (SSIM) that offers a better approximation of what image from the second row is closer to the source because its mathematical formula models human eye perception.
Imagination has been providing industry standard video IP solutions because we understand the focus should always be put on the ecosystem, which will naturally drive the differentiation factor. Using our IP will enable you to address the common challenges of a high performance solution with minimum risk and fast time to market, as we have been the only video IP provider to have shipped in massive volumes and across all areas.
An overview of existing PowerVR video solutions
Looking at the decoder video IP cores, the PowerVR Series3 video IP family has an area-optimised, Blu-Ray capable encoder that supports all major operating systems, as well as a decoder that targets Internet standards and has long latency for memory together with post processing options. We’ve also developed multi-pipe cores to suit high-performance requirements and announced the next generation of hardware decoding IP, PowerVR Series4 VPUs, which will support 4K display surfaces and features 10 bit colour precision and full 4:4:4 resolution. On the encoding side, our area optimised encoder covers H.264 BP multi standard encoding while the high profile is covered by two other members of the Series3 line-up, one of which offers a complete toolset and introduces support for 4K encoding.
The new PowerVR Series4 video IP core family delivers an encoder that is able to match the fidelity and depth of raw sensor footage, but also enhances the compression algorithm’s performance by maintaining colour precision from source to display.
Flexibility also should be taken into account. You could have everything covered for now, but what happens when new features are added to the standard a few months later? The smart way to do it is to spend that time choosing the right solution from our ever-expanding portfolio and thinking about integrating it into a SoC that can deliver the functionality you desire. We have a solution for every application, developed because we have a wide customer base that justifies the effort.
With all the potential choices from our growing list of available video IP, you also make system development and integration less difficult because we have the background required to provide the support for these innovative solutions.
Why companies need to abandon in-house designed video solutions
Companies must let go of in-house designed solutions: they will always be time-consuming, with a lot of engineering effort being put into something that, almost every time, ends up being standard-compatible at best. Teams might think they are creating a completely custom solution but instead are just wasting valuable resources that would be better deployed elsewhere, struggling to catch up with the competition. It is also cost inefficient, as maintenance and portability can quickly become a nightmare, especially when moving between SoC architectures. Instead, the engineering effort can be directed to more innovative things like integrating video transcoding with the GPU, improving memory system traffic and optimizing the SoC architecture to lower the overall latency.
Imagination has been working with a large number of customers and has had the experience of integrating its products with many industry leaders, an advantage proven by the millions of units shipped worldwide. Its video IP families are market proven across all application areas from handsets and tablets to gaming – we are the only IP vendor to have shipped in excess of 500 million solutions for video encode and decode. We offer tight integration with all our IP, including audio, demodulation and graphics technologies and can support all operating systems, including Windows, Linux and Android. By looking at all the above, one can see the clear road to not only industry leading solutions but also fast time to market and multiple standard support.
If you want to find out more on PowerVR video and display IP solutions, keep reading our blog as we have many exciting articles coming up on our key technologies in the near future. Also, follow us on Twitter (@ImaginationTech) for the latest news and updates!