Benefits

QuantumFilm™ can help you deliver your best images with even the smallest of cameras. Keep reading to find out how.

With higher dynamic range, more accurate motion capture, dynamic pixel sizing, higher resolution, and greater near-infrared and visible light sensitivity, QuantumFilm takes camera technology to the next level. Remarkable performance no longer requires compromising on portability or power consumption.

Have you ever taken a photograph or video of something moving that came out distorted? This defect, which impacts traditional smartphone image sensors, is known as the rolling shutter effect. In these sensors, each row of silicon pixels is read in turn, and different parts of a picture are taken at different times. This skews moving objects in still images—for example, a cube-shaped van driving across a scene can end up looking like a trapezoid. In rolling shutter video, shake of the camera produces the “Jell-o” effect, or a wobbling of the picture at its edges.

When shooting moving objects, the QuantumFilm sensor’s unique electronic global shutter reproduces true-to-life motion by capturing all parts of an image at once. QuantumFilm’s sensitivity to visible and infrared light can be turned on and off in a matter of microseconds, maintaining a high frame rate of 30 frames per second while minimizing distortion when shooting quick-moving subjects. The global shutter mechanism used in silicon image sensors requires a much larger pixel size than those used for image sensors in smartphones and other compact mobile devices, but QuantumFilm makes global shutter possible at any pixel size.

The benefits of capturing motion with global shutter extend beyond the purely aesthetic. Eliminating rolling shutter distortion minimizes disorientation in virtual and augmented reality, and is also vital for safety when it comes to accurate collision avoidance and face and iris authentication.

QuantumFilm isn’t just sensitive to light in the visible spectrum; it also senses longer wavelengths, or infrared light invisible to the human eye. Many everyday devices, such as remote controls and ambient temperature regulators, rely on infrared light detection to function. Just as QuantumFilm senses visible light better than silicon-based image sensors can, it also has vastly augmented infrared sensing capability. InVisage develops products and QuantumFilm for both the visible (400-700 nm) and the near-infrared (700-1100 nm, also known as NIR) ranges of light.

With its increased sensitivity, high-definition imaging and global shutter capability, QuantumFilm for NIR—available in our Spark™ product line—allows you to communicate more accurately with your smart devices, and allows them to gather more accurate information about their environments.

A structured light pattern on a drone shot with InVisage's Spark4K. Spark4K's ability to detect variations to infrared light patterns with greater sharpness allows autonomous and augmented reality devices to perform with enhanced accuracy.

A structured light pattern on a drone shot with InVisage’s Spark4K. Spark4K’s ability to detect variations to infrared light patterns with greater sharpness allows autonomous and augmented reality devices to perform with enhanced accuracy.

An increasing number of IoT devices are using structured light patterns to map their environments in 3D. But how does this work, and how does QuantumFilm help?

Using a near-infrared (NIR) laser or LED to emit a specific pattern of light invisible to the human eye, devices incorporating QuantumFilm NIR sensors detect changes to that pattern as the environment reflects this light back to the camera. However, it can be difficult for conventional structured light cameras to perform accurately outdoors or in bright sunlight because they easily saturate and fail to detect the structured light patterns their devices emit. QuantumFilm’s higher sensitivity to NIR extending to the 940-nanometer wavelength takes advantage of the fact that water in the atmosphere absorbs most of the 940-nanometer infrared light in sunlight, minimizing solar interference with structured light systems.

Combined with the higher resolution delivered by the 1.1µm QuantumFilm pixel and undistorted motion capture of global shutter, QuantumFilm’s sensitivity provides IoT devices with more accurate, higher resolution 3D depth mapping. QuantumFilm-enabled devices are therefore better able to avoid obstacles, recognize users, track movement, and model objects in their environment. InVisage produces modules tailored for specific applications, such as the Spark Authentication Module SAM and our Spark4K Micro-LiDAR SML20 module. SAM allows for iris and face authentication systems to recognize users more effectively indoors and outdoors, enabling ultra-thin bezels and as much as 50 times less system power consumption. SML20 fuses depth mapping with high resolution 4K video for unprecedented obstacle avoidance at a 20-meter range, without the added weight of conventional LiDAR.

This diagram shows a device emitting NIR light on the left and projecting a pattern of structured light onto an object. This pattern is reflected differently from an object closer to the device than from the background. By detecting this variation, the device can recognize the shape, distance, and other information about the object.

This diagram shows a device emitting NIR light on the left and projecting a pattern of structured light onto an object. This pattern is reflected differently from an object closer to the device than from the background. By detecting this variation, the device can recognize the shape, distance, and other information about the object.

The height of a camera lens affects the angle at which light enters the camera, or its chief ray angle (CRA). With silicon-based sensors, there’s a limit to how short a camera lens can be in order to prevent photons from passing into the wrong pixels. This is because silicon, a weak absorber of light, lets considerable light incident on one pixel pass to its neighbors. And the thinner the sensor, the more light escapes the correct pixel. This effect, known as crosstalk, produces color error, blurring, and washed out images, especially at the edges of the image sensor.

InVisage’s QuantumFilm is such a strong light absorber that the right pixel consistently absorbs the light incident on it, without passing significant light to its neighbors; in other words, it minimizes crosstalk. As a further benefit, light which passes through the focusing optics can be picked up by the sensor at a steeper angle. With conventional CMOS sensors, compromises have had to be made between sensor thickness and color accuracy, resulting in unsightly camera “bumps” on otherwise sleek devices. QuantumFilm offers a unique capacity to detect colors accurately, freeing  manufacturers from that compromise.

The animation below demonstrates how lens height and crosstalk go hand in hand in a silicon-based BSI sensor, and how QuantumFilm bypasses that problem. As the lens height decreases, red photons generate a signal that is collected in the green pixel region, causing color crosstalk that appears as color noise and color error. This crosstalk happens because silicon is a slow absorber of light and it takes multiple micrometers of silicon thickness to absorb sufficient light for imaging. For near-infrared image sensors, which conventionally have a significantly thicker photosensitive layer in order to maximize infrared sensitivity, QuantumFilm’s advantage with respect to CRA and monochrome crosstalk is even more pronounced. With a thinner layer of more infrared-absorbing QuantumFilm, infrared camera pixels can  be smaller than ever and deliver unsurpassed resolution.

In smartphone cameras with silicon sensors, users are forced to rely on digital zoom—and to pay the accompanying price of noisy images. But what if a new type of zoom, not limited by the size of the camera lens, could allow you to get closer to your subjects without sacrificing image quality?

Traditional silicon pixels have fixed dimensions: the silicon is divided up into fixed bins of light sensors and charge stores at the time of fabrication. In contrast, because QuantumFilm is a physically continuous layer, it can be reconfigured on the fly. As a result, InVisage is able to offer programmable pixel size. When a light–starved scene calls for making the most use of available light, QuantumFilm gathers information from a super–pixel and funnels it all to a single, low-noise junction for image rendering. When a rich and complex scene calls for the finest, sharpest possible picture, QuantumFilm can maximize its resolution for record pixel counts. Effectively, QuantumFilm allows the sensor to combine information in the physical (analog) domain to maximize the signal-to-noise ratio, regardless of light condition.

Conventional cameras with smaller pixels suffer from a problem called saturation, in which parts of image taken in bright light appear as a “blown out” wash of white and lose all detail. Saturation happens when all the photons being detected cause the charge storage portion of the sensor to overflow with too many electrons. InVisage’s separation of the two functions of light sensing (photodetection) and signal processing (electronics) into two dedicated planes leaves added room in the silicon plane for a higher full well capacity (the maximum number of electrons stored before saturation). In terms of performance, this greater capacity allows for images with enhanced dynamic range.

Dynamic range describes the finesse with which nuances of brightness and color can be distinguished, and is measured as the ratio of full well capacity to noise. Because of its limited real estate, silicon can store only a small number of electronic charges in each pixel. By freeing up silicon storage area, QuantumFilm enables storage of a much greater number of electrons. This translates to improved in-scene dynamic range for the photographer. In other words, because more information can be stored in InVisage’s image sensor, you’ll see less saturation in bright conditions and more details in dark ones.

The above diagram compares the full well capacity of three pixels: from a silicon CMOS smartphone camera sensor (1.1um), a QuantumFilm smartphone camera sensor (1.1um), and a silicon D-SLR camera sensor with much larger pixels (4.4um). The overflowing electrons in the silicon smartphone pixel on the left indicate it has exceeded its saturation point, which is much lower than that of either of the other pixels. QuantumFilm delivers higher dynamic range with a smaller pixel thanks to this higher full well capacity.

Silicon digital imaging has taken the place of photochemical film in most applications today, but the shift from film to digital has always involved a sacrifice of dynamic range—until now. QuantumFilm features a unique, expanded dynamic range capability called QuantumCinema that allows for more details to be captured simultaneously in both low and bright light.

Conventional digital image sensors rely on silicon to sense light linearly and therefore saturate when the number of electrons a pixel can store exceeds a fixed limit (known as a pixel’s full well capacity). Because full well is largely determined by the size of the pixel in a silicon image sensor, sensors with smaller pixels in smartphones and tablets suffer most from this lack of dynamic range. In contrast, the silver halide crystals used in photochemical film have a non-linear response to bright and low light that can preserve details in more extreme light conditions.

Not only does the QuantumFilm sensor provide a higher FWC than a silicon sensor at any pixel size, but its photosensitive layer also has a non-linear response to light, just like film. QuantumCinema takes advantage of this non-linear response to expand the dynamic range of the sensor even further. The images below illustrate two scenes shot with three different cameras: a conventional smartphone camera with high-resolution silicon image sensor, a camera using Kodak film, and a QuantumFilm smartphone camera sensor in QuantumCinema mode.

Silicon

Kodak Film

QuantumCinema

Silicon

train-nokia

Kodak Film

train-kodak

QuantumCinema

train-inv

Note that details outside the window in the first set of images and outside the door in the second set are lost in the case of the silicon image sensor, which saturates in the bright sunlight. In contrast, QuantumCinema captures a level of detail similar to the dynamic range of film.

QuantumCinema changes QuantumFilm’s response from the fully linear default to a combination of linear response in the dark and mid tones and film-like response in the bright to very bright tones. Hard clipping of information is eliminated in this mode, and the response precisely mimics film’s response to bright light. This extends the dynamic range of the sensor in bright light, allowing capture of details over multiple stops of exposure, more than would be possible with a linear sensor. In other words, QuantumCinema effectively combines the best aspects of silicon and film.