QuickLogic's Visual Enhancement Engine (VEE) Brings iridix to Mobile Devices
The software implementation of iridix requires high CPU load and consumes a lot of power. QuickLogic and Apical have partnered to architect and develop solutions for mobile and portable multimedia products.
The iridix algorithm developed by Apical enhances image quality by employing dynamic range compression. This innovative technology is the result of more than a decade of research by Apical scientists based on how the human eye perceives image contrast under various lighting conditions. This program research led to the development of the iridix algorithm which first found success in the Digital Still Camera (DSC) market and wide screen LCD televisions. The DSC applications used software based solutions, and the television hardware solutions were based on large FPGAs – neither option suits the mobile market due to high power consumption.
The iridix core dynamically calculates and applies a different tone curve to each pixel depending on content and position in the image. This pixel-by-pixel dynamic range and color correction enables far better control of contrast than traditional techniques. However, it also requires high CPU computing resources and, therefore, higher power consumption. Also, the iridix software must be optimized for different CPUs and DSPs. For example, in an ARM11 CPU the iridix software library processes 30 fps QVGA while consuming about 75% of the resources. The processor load scales linearly with pixel count and pixel clock rate.
With such a high CPU utilization requirement, it is very challenging to support a smooth and responsive user experience in a multi-tasking operating system. Users may experience slow transition from one application to another or see jerkiness in the video when another application with higher priority interrupts and slows the process down. Other features such as sharpening, dithering and color correction in addition to the iridix core further exacerbate the situation by putting more load on the power-hungry CPU. Even for processors that have a dedicated accelerator (such as the TI OMAP), it is estimated that the iridix core would consume most of the dedicated resources. The development effort required to optimize the software code for a specific processor is not a trivial task.
Apical required more than a year to optimize the iridix core for a TI application processor (ARM-based) used for a DSC application, primary still images only. Much of the time was spent optimizing resource sharing with other applications, which usually involves a lot of benchmarking on the actual device within the software application and testing. Normally, it takes three to six months of intensive development process for applications such as a DSC, and much longer for complex smartphone applications.
The only feasible implementation for the mobile market is a hardware-based solution. QuickLogic and Apical have partnered to architect and develop the optimal blend of algorithms with QuickLogic’s patented ViaLink programmable fabric for mobile and portable multimedia products. The VEE technology is based on the proven iridix algorithms and is supplemented by additional image and video enhancement blocks such as dithering, hue rotation, color correction, and non-linear sharpness filtering, among others.The VEE solution substantially enhances image and video quality by optimizing the dynamic range, contrast, and color saturation pixel-by-pixel to provide a natural and usable viewing experience under low backlight or bright ambient light conditions. Seamlessly integrated into the display path, VEE enhances the user’s mobile multimedia visual experience while drastically reducing backlight power to extend battery life.
Advertisement
