Vision science studies suggest that the eye is able to discern more than 11 bits of dynamic range for each of the three primary colors – red, green and blue – that typically comprise a given scene. The optical nerve connecting each eye to the brain, on the other hand, is only able to pass roughly five bits' (40 levels) worth of each primary color's data. Yet the brain still is capable of discerning more than 10 billion discrete levels of total color depth, equivalent to that of the 11-bit-per-color (2000+ level-per-color) source.

In my December column, I observed that smartphones and tablets are starting to be used in places where purpose-built embedded systems once reigned, such as point-of sale terminals. At home, for example, I have a small Android tablet that I use as an Internet audio player. And my local sandwich shop uses iPads as self-service ordering and payment terminals.

BDTI is well known for its software-related capabilities: performance- and power consumption-related benchmarking, for example, along with algorithm evaluation, and development and optimization work. But the company is no stranger to hardware, either.

"If it has speech recognition, why do we have to use our fingers?" According to Bernie Brafman, Vice President of Business Development at Sensory, that simple question has been at the forefront of many of the company's customers' minds throughout Sensory's 19-year existence. That same question has therefore guided the privately held company's technology and product roadmap. But actualizing this aspiration involves, at first glance, difficult tradeoffs.

Soon after BDTI got its start in the early 1990s, we became known for our benchmarks. We benchmarked whatever types of processors people were using for embedded digital signal processing: first DSPs, then CPUs, and eventually MCUs, FPGAs, and GPUs, too.

Algorithms are the essence of digital signal processing; they are the mathematical "recipes" that transform signals in useful ways. Companies developing new algorithms, or considering purchasing or licensing algorithms, often need to assess whether an algorithm will fit within their processing budget—and thereby within their cost and power consumption targets.

Back in September 2011, an InsideDSP article described a just-published analysis conducted by BDTI and sponsored by Altera, evaluating the viability of implementing complex hardware-accelerated single-precision floating-point functions on FPGA fabric. As I wrote then:

In a recent interview in EE Times, BDTI co-founder and president Jeff Bier commented:

Multi-core CPUs are very powerful and programmable, but not very energy-efficient.  So if you have a battery-powered device that is going to be doing a lot of vision processing, you may be motivated to run your vision algorithms on a more specialized processor.

If you’re a regular reader of this column, you know that I’m enthusiastic about the potential of “embedded vision” – the widespread use of computer vision in embedded systems, mobile devices, PCs, and the cloud.  Processors and sensors with sufficient performance for sophisticated computer vision are now available at price, size, and power consumption levels appropriate for many markets, including cost-sensitive consumer products and energy-sipping portable devices. This is ushering in an era of machines that “see and understand”.

Qualcomm recently opened up the QDSP6 (aka "Hexagon") DSP core in its Snapdragon SoCs to programming access by its customers and software developer partners. Multimedia applications, for example, can benefit from leveraging QDSP6 processing resources, boosting overall performance, minimizing overall power consumption, and freeing up the CPU to tackle other tasks.