Today kicks off the 3rd annual Automotive Visual & Display Technologies (AVDT) event in Ann Arbor, Michigan (March 26-28), where leaders in automotive engineering, design, and manufacturing will discuss the next generation of car HMIs (human machine interfaces) and display integration. My colleague Matt Scholz presents a conference session this afternoon titled “New Measurement Methods to Solve Emerging Display Test Challenges.” (If you happen to be reading this blog from the show, check out Matt’s talk at 3:20 PM!) As I explore the AVDT agenda—with topics covering head-up displays to human-centric design and autonomous vehicles—it’s clear that we’re on the verge of a new automotive era.
Trends in automotive manufacturing are really changing the way I understand what a car is. Digital displays are taking the place of analog components in the form of virtual speedometers and center console controls—with all information provided in a digital medium, perhaps a car is a kind of computer. We’re now interfacing with our cars through touchscreens and wireless connectivity, making cars an extension of our handheld smart devices—so, maybe a car is actually a smart device, too. Autonomous vehicles are turning drivers into passengers and creating blank slates for completely new interior designs, like wide-screen curved displays, integrated computers, and augmented projections in windows. With these changes, a car is transforming into an infotainment room on wheels.
What’s cool about these trends is how displays are appearing in more places; and—aside from the steering wheel and pedals (still relevant, for now)—displays have become the central HMI for vehicle operation. Every day it seems another OEM is unveiling a design with huge display-based dashboards, intelligent rearview mirror displays, head-up displays (HUDs)… And some truly amazing concepts like this “transparent pillar” display from Jaguar Land Rover:
It may seem crazy to think, but did you know that in-vehicle displays aren’t even an option anymore? Beginning this year—effective May 1, 2018—the National Highway Traffic Safety Administration (NHTSA) will require all new “light vehicles” (including cars, SUVs, trucks, and vans) to feature “rear-view visibility systems”—in other words, backup cameras. Displays will follow the backup camera requirement and make their way into every vehicle interior, relaying a video feed to the driver. The requirement puts the pressure on automakers, not only to adopt display technology, but to understand and ensure display performance in order to comply with these safety measures.
What is the expectation for automakers who need to compete in this display-centric industry? Because displays are critical for vehicle operation, visibility of digital information is key. Display quality must be tested to ensure imperfections do not impede user control. Defects like non-uniformity, poor brightness and color performance, dead pixels and lines in the display, and other blemishes can impact operation. Plus, we need to be able to see information on the display in all ambient light and weather conditions, which change wherever the vehicle goes. The display should perform for the life of my vehicle; much longer than the average smart device display, such as a cell phone that I may replace every year or so. So, the manufacturer will need to test for any defects that might worsen with use, regular vibration, and extremely hot or cold temperatures over time.
The displays are important not only for operational reasons. The quality of a display can have an impact on my perception—as a customer—of the vehicle overall, including quality, value, and reliability. This influence only grows as displays occupy more real estate inside the vehicle. If I think a car has a low-quality display component, I may question the quality of the car itself. One bad display can be the Achilles heel for an automaker’s brand.
Yes, quality is key. With so many different flavors of displays in the automotive industry, and the role of in-vehicle displays becoming more diverse, the challenges automakers face to ensure quality continue to grow. The traditional methods of display test and measurement—which we’ve used to measure flat panel display (FPD) quality since the days when displays were still quite novel—no longer adequately assess the unique characteristics of all new manifestations of display technology in the market. For each display technology innovation (e.g., OLED), a new measurement method… a new camera, a new lens, a new software test…must be developed. Working with innovative automakers, Radiant has put our in-office test labs to good use, learning to adapt our photometric imaging systems to address new challenges in display testing.
Break the Mold, Not the Display
During his presentation at AVDT today, Matt will introduce several solutions Radiant has defined to acquire the data necessary to measure quality in new display tech. He’ll discuss the importance of using imaging systems to characterize defects across displays of any shape or size. Called free-form displays, a display that is manufactured to occupy a non-rectangular space in a vehicle could manifest as a circular speedometer display, or any undefined shape designed to complement the interior aesthetic. Because they come in bespoke shapes, free-form displays require new methods of manufacture and integration that may apply new mechanical stresses, which in turn introduce unforeseen defects, or defects that appear in new areas around a curved bezel or along an angled edge.
Not only are Radiant’s photometric imaging systems able to capture the full spatial area of free-form displays to locate defects within the context of the entire display, but our software offers new tools to align non-standard shapes. Our Register Active Display Area (RADA) tool is typically used to locate the corners of a rectangular display to auto-rotate and crop images for analysis. With free-form displays, we continue to employ corner location on a rectangular test pattern; then, we use a unique feature called RIDI (Register Inside Display Image) to zero-out the pixels in the negative space between the pattern and the display edge, leaving only the active display area for analysis.
RADA locates the corners of a rectangular test image (left), and then RIDI detects the edges of the actual display by zeroing-out the pixels in the negative space (the difference between the test image and the actual display area) for accurate analysis (right).
Fixing Fickle Pixels
As displays appear in more shapes and sizes, OLED technology is not far behind. Because OLED pixels emit their own light, there is no need for inflexible backlight units (BLUs) to power the display, as in LCD displays. For this reason, OLEDs have been the tech of choice for the latest flexible, curved, and free-form displays. Along with the benefits of OLED’s self-emitting pixels, however, there are corresponding challenges. Luminance (brightness) levels across OLEDs can vary quite a bit when each pixel produces its own light. Luminance on a per-pixel basis also varies quite widely over gray levels. Non-uniformity is a common issue in new displays that use OLED technology, and therefore a hot topic in quality testing.
At his talk today, Matt will describe a novel measurement method that Radiant has developed specifically for the purpose of OLED pixel-level measurement and correction. Called “OLED Demura,” this method acquires the luminance value of each pixel in the display and then calculates a correction coefficient to be supplied via input control to correct the appearance of pixel non-uniformity and mura (unevenness). The demura process employs Radiant’s patented method of measuring Spaced Pixel Test Patterns (US Patent 9135851), wherein the camera is presented with a series of dot-matrix test patterns on the OLED display. Each pattern has a small portion of the display pixels turned on, while the rest are off. A Radiant imaging photometer or colorimeter is used to measure the luminance of the on-pixels, and then the matrix is adjusted to turn the next series of pixels on for measurement. This process is repeated until all pixels in the display are measured, ensuring the accuracy of luminance calculations across OLED displays of any arbitrary resolution.
During Spaced Pixel Test Pattern measurement, several patterns are measured with different pixels turned on, until a luminance value has been acquired for all pixels across the display.
Values acquired during pixel measurement are combined into a synthetic image for analysis by the camera, which compares luminance values at each of the pixels’ X,Y locations to determine uniformity. Using this image, the software calculates the necessary correction coefficient to adjust the luminance values of each pixel until luminance across the display is uniform. A sample of these results are shown below in Radiant’s TrueTest™ Software (using false color mode), to compare the look of a non-uniform display and the same display after demura correction. Using the demura method, automotive manufacturers and suppliers now have a solution to reduce waste in OLED manufacturing, saving “defective” components from being thrown out, and enabling cost-effective design of new display shapes and curvatures.
Demura is used to correct non-uniformity of orange-colored emissions in an OLED display.
If you’re not at AVDT on your way to Matt’s presentation by now, I’ll spill the spoilers and detail the last two display test methods that Radiant will cover during our presentation. Both methods precisely define the measurement hardware, software, and setup required for testing unique aspects of automotive display types; specifically, HUD and anti-glare displays.
Heading Up Automated SAE HUD Testing
In regard to HUD, Radiant defines specifications of a measurement method that supports a new standard from the Society of Automotive Engineers, SAE J1757-2 “Optical System HUD for Automotive.” While SAE doesn’t dictate a measurement system for HUD testing, it does outline specific needs for capturing both photometric (uniformity, luminance, chromaticity, contrast) and dimensional (distance, distortion, ghosting) data. Radiant’s HUD measurement method employs a photometric imaging system, which combines the capabilities of a photopic measurement device like a spot meter with the power of a machine vision camera or imager capable of dimensional measurement. This system greatly reduces the cost and complexity of a HUD measurement solution.
HUD testing requires photometric and dimensional measurements across the full area of the HUD projection. In the top image, a piecemeal solution requires a spot meter capable of photometric measurement, a robotic arm to position the spot meter at various points across the display, and a machine vision camera capable of dimensional measurement. In the bottom image, a single imaging photometer can be used to capture the entire display in one image and apply photometric and dimensional measurements simultaneously.
In addition, Radiant offers a specific test suite tailored to requirements for HUD quality evaluation. These tests can be run in sequence to enable rapid automated testing of all HUD parameters, which means HUD testing can be performed more quickly, even in production settings. Tests for Radiant’s HUD measurement solution include Uniformity, Contrast, Line Defects, Pixel Defects, Checkerboard Contrast, MTF Slant Edge Contrast (based on ISO 12233), MTF Line Pairs, Image Sticking, Distortion Grid, and Eyebox Limit. An example of an automated test sequence in Radiant TrueTest Software can be seen below.
The Radiant HUD measurement system can be integrated to control display images, enabling automated testing while limiting setup time and reducing complexity to acquire all necessary data. Using this method, manufacturers can meet SAE standards for HUD quality with extreme efficiency.
All That Sparkles Is Not Gold
The last topic Matt will touch on at AVDT today introduces Radiant’s method for characterizing the effect of a visible property called “sparkle” on anti-glare displays. Because automotive displays must be visible in a wide range of arbitrary ambient lighting conditions, anti-glare and anti-reflective layers and coatings are important for reducing glare reflected from the display surface. An anti-glare (AG) layer has a micro-structured surface that reflects light in a diffuse pattern, so as not to obstruct the display in bright sunlight, for example. However, a visible effect called sparkle can occur when the surface structure of the AG layer competes with the pixel geometry of the display – creating a “sparkly” appearance.
A display with AG layer (left) is affected by sparkle, causing the display image to appear pixelated and less clear than the display without the AG layer (right).
While the effects of sparkle may be perceived as “poor quality” to the human eye, these effects are extremely difficult for machines to quantify and measure. The goal is to acquire precise tolerances to employ in quality control between display suppliers and automotive OEMs, so only the highest (perceived) quality displays make it into our vehicles. Extensive testing performed by Radiant both in labs and at customer sites has helped us to refine a method for repeatably measuring sparkle in AG displays. This is made possible using a high-resolution ProMetric® imaging colorimeter precisely aligned to the display, and a Random Mura algorithm in TrueTest Software, which calculates a standard deviation percentage of pixel-level luminance variation across the display. Rather than getting too technical here, if you’d like to know more I encourage you to join Matt’s webinar on this same topic, broadcasting live next week: “Defining a Sparkle Measurement Standard for Quality Control of Anti-Glare Displays.”
The key to Radiant’s sparkle measurement method is its ability to not only acquire repeatable data, but also to correlate this data to human visual perception of display quality. In a case study at a Radiant customer site, several display manufacturers were asked to submit their displays for review by a panel of human observers. Each supplier’s display was ranked by the observers on a relative scale from 1 to 10 (with 1 being the best and 10 being the worst) for overall display quality, given the presence of sparkle. Once human ranking was complete, Radiant measured the same displays using our ProMetric imaging colorimeter solution and Random Mura algorithm. A near 1:1 correlation was shown between the human observers’ rankings of the displays and the relative values obtained for pixel-level luminance variation by the Radiant system, with the higher percentage correlating to the worst quality display.
Because the results of these studies were so consistent, the automotive OEM hosting the study has since issued a directive to its suppliers that they must use the Radiant Sparkle Measurement Solution for their outgoing quality control (OQC) operations. This ensures that AG displays meet a minimum tolerance for sparkle, as defined by the OEM for their own custom quality control standards. This measurement method empowers any OEM to define similar internal standards for sparkle tolerance, making AG display quality a measurable and controllable value in their acceptance of supplier displays.
The more display test applications we encounter at Radiant, the more evident the display’s ever-increasing impact on the overall vehicle performance becomes. Whatever a car really is, or will be in the future—a computer, smart device, or room on wheels—I wager that the display will determine the final definition. Or, as displays continue to take new shapes and new roles, perhaps cars are never just one thing. Think about it: displays provide the interface to virtually anything. Of course, we can’t apply just one display test method to evaluate all of them! But that’s what’s great about Radiant—the flexibility that comes with a core technology based on imaging, proprietary software tools, and an engineering spirit, keeping us in the thick of display innovation, with the agility to meet ambiguous challenges as a solutions provider.
Are you changing the landscape of automotive displays? We’re ready for new challenges. If you’re at AVDT, you know who to connect with, or contact us here and let us tailor a display test method that’s made just for you.