Startup InVisage has announced a smartphone sensor with global shutter for distortion-free video and greater dynamic range than its CMOS rivals. The Quantum13 is a 13MP sensor based on the company's Quantum Film technology that moves away from silicon as the light-sensitive medium. The sensor, now available in small quantities for testing by smartphone makers, can also be built into a smaller package for slimmer smartphones. We spoke to CEO Jess Lee about the sensor and what it brings to the market.

Quantum Film uses nano-scale crystals that are precisely sized so that they generate electrons in response to specific frequencies (colors) of light. In this implementation, though, this specificity isn't used for color differentiation. Instead, to make it easy to incorporate into a conventional image processing and manufacturing pipeline, the Quantum13 uses a film that is sensitive to the whole visible light range, mounted behind a standard Bayer color filter.

InVisage's Quantum Film uses a film of quantum dots to convert light into electrical charge. This simplifies the design of the sensor which, in turn, should allow a series of benefits including greater dynamic range and a simultaneous 'global shutter' in video mode.

Global Shutter

The Quantum Film approach still has benefits, though. One advantage is related to video, where a mechanical shutter can't operate fast enough to be used, so you're stuck with electronic shutters and a well-known artifact known as the 'rolling shutter' effect.

Conventional CMOS designs collect light whenever the sensor is switched on and do so until they are read and reset at the start and end of every exposure. This connection between exposure and read out means the row-by-row read-out of the sensor also demands row-by-row starting of the exposure, introducing the 'rolling shutter' effect. By contrast, the Quantum Film sensor only collects light when the transparent electrode in front of the sensor is switched on, a process that applies bias across the film to funnel generated electrons into the rest of the electronics. That entire electrode can be turned on at once, meaning that the entire sensor can be switched on and off all at once in a manner that is entirely separated from the sensor read-out, which can be conducted after exposure has been stopped. Ultimately, this means the sensor can shoot video footage where all pixels are captured simultaneously and there's no rolling shutter effect.

Dynamic Range

There are numerous other advantages of shifting the photosensitive layer, and its role of generating electrons (or holes) from productive encounters with light, to the Quantum Film layer. For one, the Quantum Film layer is thinner (~0.4 μm) than the typical photosensitive silicon layer in traditional sensors (~3 μm). The decreased thickness of the photosensitive Quantum Film layer also means lower native 'crosstalk', or leaking of signals from one neighboring pixel to another. This means pixels don't need to be as optically and electronically separated from neighbours as with traditional designs. The net effect of all this is that less space is occupied by components required by traditional designs, meaning each pixel on the Quantum Film sensor is structurally simpler underneath that film layer. The result? The silicon circuitry underneath the film can be designed with more room to store electrons. Which, in turn, means each pixel can store more charge before becoming overwhelmed.

Lee says its sensors have pixels with full well capacity 3X larger than 1.1 μm pixels in competing 13MP sensors as a result of this. This should give a 1.5 EV improvement in dynamic range and, assuming you can give the camera additional exposure to make use of the higher saturation point, higher signal:noise ratio for most tones. Meaning cleaner crisper images.

The technology promises other benefits, including being more light-sensitive than existing silicon-based designs.

In addition, Lee says Quantum Film's response becomes non-linear as it approaches saturation. The voltage bias across the film is likely affected by the magnitude of the signal (electrons) generated, meaning fewer photon encounters are ultimately productive at the higher electron volumes near saturation. This ultimately 'compresses' highlights, much like the roll-off of silver nitrate film. In its standard mode the sensor doesn't attempt to use the data from this non-linear region but the company has also developed a 'Quantum Cinema' mode that attempts to process and incorporate this (at least 1 EV worth) information about extreme highlights.

Taken together, the increased full well capacity and non-linear highlight response means up to around 2.5 to 3 EV of additional highlight range and signal. This should, theoretically, allow the Quantum Film based sensor to challenge performance of larger sensor devices.

The Results

At the point of writing, we haven't had a chance to see any samples from the sensor, but the company has shot a short film with it. The film, shot with a pre-production sensor and processed through a non-standard image processing pipeline, has an intentional color cast to mimic the aesthetics of a Wes Anderson movie, we're told. Between this and the effects of YouTube compression, it's hard to draw any solid conlusions about video quality. However, if the example below is representative, it does appear that highlights themselves are recovered and compressed (or perhaps compressed at the film/sensor stage, uncompressed and then recompressed in processing for low dynamic range displays). The technology may hold potential not just to extend the performance of professional full-frame cameras, but also in getting smaller compacts to perform at levels similar to larger sensor cameras due to the film.

We did not discuss cost with Lee, but he stressed that the read-out circuitry is also essentially the same as that in a standard CMOS chip, to allow much of the sensor to be built on existing fabrication lines. Lee did hint that perfecting the process for applying the film to the silicon circuitry explained the delay in launching the chips, which it first said would be ready in 2010. We'll be curious to hear about, and report upon, devices that may be integrating this technology in the near future, as we're sure that more practical demonstrations of the technology are something many viewers are keen on seeing.

You can read more about the fascinating technology behind InVisage's products here.


Press Release:

InVisage Launches Industry’s Highest Performance Smartphone Camera Sensor Quantum13

Using advanced quantum physics, the Quantum13 brings cinematography-grade dynamic range and smooth motion capture to smartphones, driving silicon CMOS image sensors into obsolescence

BEIJING--(BUSINESS WIRE)--InVisage Technologies Inc., the pioneering developer of QuantumFilm™ camera sensors, today introduced its inaugural product, the Quantum13 camera sensor. Leveraging the power of advanced nanotechnology found in the company’s QuantumFilm platform, the Quantum13 captures images with full tonal ranges in high and low light, and is the first smartphone camera sensor with an electronic global shutter. Today in Beijing, InVisage is showcasing Quantum13-enabled smartphones running on both Qualcomm and Mediatek platforms. Shipments to smartphone vendors will start this quarter.

The Quantum13 is the world’s first electronic image sensor that does not use silicon but rather a quantum dot film, QuantumFilm, to capture light. QuantumFilm has a natural light response curve matching the human eye. As a result, the Quantum13 has a powerful single-shot high dynamic range (HDR) mode called QuantumCinema™. This mode provides up to three additional stops of dynamic range compared to conventional CMOS image sensors. Both still and video modes will see this level of performance without any added HDR software processing, which can often cause a missed moment and a distorted image.

Quantum13 also offers the world’s first electronic global shutter for smartphones, a camera technology that enables full frame capture instead of the cumbersome rolling shutter method that CMOS sensors use to scan from the top of the image to the bottom. With this electronic global shutter, Quantum13-enabled smartphones can capture crisp still photos of fast-moving subjects at full frame shutter speeds normally reserved for digital SLRs, and deliver smooth 2K and 4K video without any rolling shutter distortion.

“The launch of Quantum13 marks a new era for the smartphone camera industry,” said Jess Lee, CEO of InVisage. “For the first time, smartphones will capture images on an entirely new medium. Not silicon. Not film. QuantumFilm. We are thrilled to showcase the capabilities of Quantum13 to the richest and most vibrant ecosystem for smartphones. And we are delighted to share that several smartphone vendors have already adopted Quantum13 for upcoming release.”

Quantum13 is a 13-megapixel, 1.1-micron pixel camera sensor that fits in an 8.5-millimeter by 8.5-millimeter module. With light absorption eight times faster than silicon, QuantumFilm creates an ultra-thin light capture medium that accommodates much higher incident angles of light, resulting in an unprecedented 4-millimeter camera module height. This thinner camera module allows for even thinner smartphone camera designs.

“InVisage is targeting the mainstream 13-megapixel smartphone camera market,” added Tetsuo Omori, senior analyst at Techno Systems Research. “According to our research, the worldwide volume of the 13-megapixel camera sensor market is projected to increase from 408 million units in 2015 to 995 million units in 2020.”

A short film shot with a Quantum13 camera was recently released and may be viewed here, along with a behind-the-scenes video that includes side-by-side comparison footage with a comparable silicon sensor here.

Pricing and Availability

Quantum13 is sampling now and initial shipments to leading smartphone vendors are expected this quarter. Pricing information may be obtained directly from InVisage at www.invisage.com/contact.