Tuesday, June 30, 2015

ESPROS Publishes its QVGA ToF Demo, Detailed Datasheet

ESPROS epc660 BSI ToF sensor is probably the world's only ToF device having its detailed datasheet (106 pages) openly available at the company's FTP site.

ESPROS also publishes a Youtube video with a demo of its sensor's capabilities:

Update: Other than the ToF sensor datasheet, ESPROS publishes the detailed board and system design guide, and eye safety analysis.

Sony to Raise ~$3.6 Billion to Invest in Stacked Sensor R&D and Production

Reuters: Sony announces its plans to raise $3.6 billion via new shares and bonds to expand stacked image sensor R&D and production. It's Sony's first new share issue in 26 years, worth close to a tenth of its current market value. The company expects to raise 321 billion yen ($2.62 billion) from a public stock offering and a further 119 billion yen from a convertible bond issue to fund a boost in sensor output capacity at its advanced plants in Japan.

For some reason, Sony PR document blocks copy/paste function, so I'm showing a snapshot of it:

Next iPhone Dual Camera Rumors

DPReview quotes Business Weekly Taiwan newspaper speculations that the next Apple iPhone might have a dual rear camera, similar to HTC M9+ and Huawei Honor 6+ phones. The original article talks about an interview with Largan CEO Linen Ping and the chances that Largan business would surge in future. No credible source was cited for these rumors.

Monday, June 29, 2015

Plasmonic Color Filters Review

Glasgow University, UK thesis "Structured Photonic Materials for Multi-Spectral Imaging Applications" by Iain James Hugh McCrindle presents a nice overview of the progress in plasmonic and metamaterials-based color filter designs. There are still not really competitive with pigment-based CFAs, but might improve over time:

Plasmonic CFA response (on planar surface)

IFTLE on Stacking Technology Progress

Insights From Leading Edge blog overviews image sensor stacking progress and recent publications. "Given the continued, aggressive stacked CIS development underway from independent device manufacturers (IDM) and foundries it’s predictable that stacked chip adoption will occur very rapidly over the next few years."

Sony ISX014 Stacked dice (Chipworks)

Saturday, June 27, 2015

Free Book on Photon Seeing

It came to my attention that The National Academies Press book "Seeing Photons: Progress and Limits of Visible and Infrared Sensor Arrays" (2010) is available for free download. The authors of the book are stated as Committee on Developments in Detector Technologies, Standing Committee on Technology Insight—Gauge, Evaluate, and Review, and Division on Engineering and Physical Sciences. The book is heavily skewed toward military and space applications:

Talking about the progress of single photon detecting technology, the book has predicted in 2010:

There is also in interesting graph comparing DRAM capacity with imager resolution growth:

and a comparison of LWIR technologies:

Friday, June 26, 2015

Sony Fast 1-inch Sensor Videos

Imaging Resource posts a nice compilation of Sony 1-inch sensor 960fps slow motion videos. While other slow motion videos were published in the official RX100 and RX10 Sony playlist, these video clips have long advertisement tails:

Thursday, June 25, 2015

DALSA TurboDrive Compresses Image Data on the Fly

Marketwired: Teledyne Dalsa presents TurboDrive, a proprietary and patent-pending data encoding technology that allows some DALSA GigE Vision cameras to achieve breakthrough speeds, increasing throughput by as much as 150% while retaining 100% image data.

"We're pleased to deliver an innovative speed advantage to customers who need to push beyond the current GigE bandwidth limitations with no loss of data," commented Mark Butler, Product Marketing Manager for Teledyne DALSA, "It's available now in our low-cost Linea line scan cameras, and will continue in future area cameras set to launch in the fall."

The company's technology primer explains how the compression works:

"Leveraging neighborhood effect Image entropy is the first principle used in TurboDrive. But to reduce even further the number of bits required to encode pixel information (with no loss of information), TurboDrive considers the neighborhood effect. The neighborhood of a pixel is the collection of pixels that surround it. Although the exact distance of a neighbor can vary, in this analysis, we will limit our example to the adjacent pixels (i.e. those that directly touch the reference pixel).

For most pixels, there is little pixel to pixel variation and a lot of redundancy. Therefore, it is possible to efficiently use the information of the adjacent pixels to more efficiently encode the reference pixel. One way to see this is looking at a high-pass 2D filter implemented using a convolution. A simple high-pass filter has the sum of all of its coefficients equal to 0. The filter we use in our model has a 3x3 mask and it provides the largest weight to the center pixel.

The result of this filter provides the difference between the reference pixel at the center, and four of its closest neighbor. It can be seen that, for a uniform image, the 9 pixels have the same value and the result out of this filtering operation is 0. Essentially, the less pixel to pixel variation, the smaller the value output by this high-pass filter. One can intuitively understand it takes less bits to encode a small value than to encode a large value. Obviously, it is possible to play with the weights of the 9 filter coefficients of this model to adapt to the image content."

Sharp Launches 14 New Progressive CCDs

Sharp keeps investing in CCD development, presenting 14 new products and few more marked "under development:"

Wednesday, June 24, 2015

Pixpolar MIG Pixel in Space Applications

ESA Technology Exchange: Pixpolar pixel has been considered for ESA space programs, although the original article only mentions "A Finnish company [that] has developed a new image sensor technology based on patented MIG (Modified Internal Gate) pixel architecture."

"The Company has participated in the development of a silicon drift detector (SDD) for X-ray applications within the TRP activity. The main goal of the project was to introduce extremely low noise SDD macro-pixels for X-ray detection applications. The Company is able to produce very low noise photon detectors for X-ray, visible, UV, near infra-red, and particle detection applications. MIG sensors are ideally suited for Space application as they enable simultaneously asteroid tracking as well as direct detection of planets around stars through continuous readout. Increasing the frame rate does not increase the noise, there are no interface issues, and they are tolerant to radiation damage."

IISW 2015 Review: Stacked Sensors

Albert Theuwissen continues his IISW review, this time talking about stacked sensors from Omnivision, Sony, Olympus, TSMC, and NHK.

Tuesday, June 23, 2015

IISW 2015 Pictures

16-megapixel IISW 2015 group picture and few others are posted on the Workshop site. If participants have interesting photos to post they are welcome to email them to jsolhusvik@imagesensors.org to add them to the web page.

Omnivision Released First C-PHY Image Sensors

PRNewswire: Tektronix names Omnivision among early adopters of the new MIPI C-PHY standard, published in September 2014. OmniVision's OV23850 and OV21840 20+ megapixel image sensors for smartphones have implemented the C-PHY 1.0 interface. Using Tektronix C-PHY Essentials analysis software, OmniVision and Tektronix were able to successfully validate and characterize the C-PHY interface used on the two new sensors.

"With an up to 17.1 Gb/s data rate over a nine-wire interface, C-PHY provides the performance needed for high megapixel count image sensors along with low power consumption," said Brian Reich, GM, Performance Oscilloscopes, Tektronix. "Our solution for C-PHY transmitter testing allows engineers to detect issues early on in the product development phase and confidently assures transmitter devices' performance and interoperability."

"OmniVision is excited to have Tektronix's oscilloscope-based C-PHY Essentials software for validating the next generation MIPI interface. With its three-wire signaling, C-PHY represents a significant design and test challenge," said Paul Gallagher, senior director of marketing and business development at OmniVision. "With the help of Tektronix C-PHY Essentials, we were able to easily set up cameras utilizing the OV23850 and OV21840 sensors and confirm the C-PHY interface's compliance with the MIPI specification."

For some reason, the original announcement of OV23850 and OV21840 in Nov. 2014 does not mention C-PHY interface feature.

e2v IISW 2015 Papers On-Line

e2v publishes its IISW 2015 papers on its web site:

"Electron Multiplying Device Made on a 180 nm Standard CMOS Imaging Technology" by Pierre Fereyre, Frédéric Mayer, Mathieu Fournier, Clément Buton, Timothée Brugière, and Rémi Barbier presents electron multiplying CMOS pixel:

e2v also publishes another IISW paper: "CMOS Charge Transfer TDI with Front Side Enhanced Quantum Efficiency" by F. Mayer, S. Pesenti, F. Barbier, H. Bugnet, J. Endicott, F. Devriere, T. Ligozat

Monday, June 22, 2015

IISW 2015 Review: Image Sensors for Photography

Albert Theuwissen continues his review of IISW 2015, now posting the part about image sensors for digital photography. This part talks about Samsung BSI APS-C sensor, Canon and Sony sensors with PDAF in every pixel, and DALSA large area CCD.

TowerJazz Manufactures Automotive Gated Image Sensors for BrightWay

GlobeNewsWire: TowerJazz announces that BrightWay Vision has chosen its TS18IS process to manufacture image sensors for its patented automotive cameras, specifically forward looking cameras in vehicles, to allow visibility in all weather conditions. BrightWay Vision has developed BrightEye, an advanced camera for day and night-time forward facing driver assistance functions based on gated imaging technology.

TowerJazz and BrightWay Vision jointly developed a unique gated pixel, something that until now has never been produced on silicon as it requires extremely fast charge transfer on the pixel. This sensor, based fully on a CMOS process, provides superior performance over competing solutions based on thermal vision technologies, and is much more cost effective. A "gated" sensor allows collecting light that is reflected back from objects at a specific distance from the vehicle, thus eliminating the noise level that is usually associated with ambient light. This sensor also allows the system to accurately measure the distance of certain objects from the vehicle and alert the driver accordingly.

BrightWay's Youtube video compares different vision technologies. The clip consists of three video channels captured simultaneously by three different cameras with the same field of view, positioned in the same direction:

  • Pulsed Light & Gated Image Sensor (BrightWay Vision active-gated imaging technology).
    Outcome: The image is clear and the person in the kayak is noticeable. The backscatter residual signal originating from the reflected signal from nearby droplets is negligible.
  • Continuous Near Infrared Light & HDR Image Sensor.
    Outcome: Backscatter originating from the reflected signal from nearby droplets is masking the image.
  • Passive LWIR.
    Outcome: The emitted signal of objects (e.g. person on the kayak) are masked due to the droplets’ emitted signal.

"We are absolutely thrilled with the performance of our sensor using TowerJazz's leading edge CIS technology," said Ofer David, CEO of BrightWay Vision. "The tight cooperation on the development between our R&D team and the excellent pixel development team of TowerJazz, allowed us to produce a very high performing sensor that is now being evaluated by leading car manufacturers."

"Our distinctive technology and deep knowledge and experience in global shutter pixels allowed us to develop a unique pixel that cannot be found elsewhere and can provide a gated image in a high frame rate with extremely good pixel properties," said Avi Strum, VP and GM of the CIS Business Unit at TowerJazz. "We have very high confidence in BrightWay's capabilities and ability to drive this breakthrough solution into the automotive market."

Sunday, June 21, 2015

Saturday, June 20, 2015

UHDTV News: NHK 8K Compact Camera, PERSEUS Codec

DigiInfo TV publishes a video of a compact 8K camera and environment presented at NHK Open House days at the end of May:

One of the biggest NHK challenges is implementing the high speed video processing and delivering the broadcast traffic. Videonet article discusses a rapid adoption of PERSEUS codec developed by V-Nova, which possibly solves these problems. The London, UK-based startup V-Nova has been established in 2011, but came out of stealth mode in April 2015 "claiming that PERSEUS completely shifts the bit rate curve and makes it possible to deliver UHD in HD bit rates, HD in SD bit rates and SD in sub-audio bit rates."

V-Nova founders discuss the disruptive nature of their proprietary codec technology in this pre-launch Youtube video:

Eric Fossum Papers at IISW 2015

Eric Fossum posts his papers presented at IISW 2015:

L. Anzagira and E.R. Fossum, Color Filter Array Patterns Designed to Mitigate Crosstalk Effects in Small Pixel Image Sensors, in Proceedings of the 2015 International Image Sensor Workshop, Vaals, The Netherlands, June 8-12, 2015.

L. Anzagira and E.R. Fossum, Two Layer Image Sensor Pixel Concept for Enhanced Low Light Color Imaging, in Proceedings of the 2015 International Image Sensor Workshop, Vaals, The Netherlands, June 8-12, 2015.

S. Masoodian, A. Rao, J.J. Ma, K. Odame and E.R. Fossum, A 2.5pJ Readout Circuit for 1000fps Single-Bit Quanta Image Sensors, in Proceedings of the 2015 International Image Sensor Workshop, Vaals, The Netherlands, June 8-12, 2015.

E.R. Fossum, Multi-Bit Quanta Image Sensors, in Proceedings of the 2015 International Image Sensor Workshop, Vaals, The Netherlands, June 8-12, 2015.

Friday, June 19, 2015

2015 IISW Recap by Chipworks

Chipworks image sensor analyst Ray Fontaine publishes his thoughts on IISW 2015 and related topics, and also covering some parts of Ray's paper at the Workshop. Few quotes:

"As we head into the 3D integration era, I predict packaging inventions will rise as a percentage of the whole. It’s also natural that as the imaging community gains access to lower power and higher density processing all the dreams of years past will come to fruition in the form of innovation at the systems and software level."

"For the time being pixel scaling has essentially plateaued at the 1 µm generation so all the mobile small-pixel leaders have been looking for other knobs to turn. A lot of development has been done to embed color filters into the tungsten aperture grid metal. Sony and ON Semiconductor (Aptina) currently hold the world record (based on Chipworks’ analysis) for getting down to a 1.5 µm thick optical stack, as measured from the back-illuminated (BSI) silicon (Si) surface to the surface of the microlens. There doesn’t seem to be much more to do here, but time will tell."

"Sony has about a two year lead in the domain of stacking image sensor chips... Depending on which flavor of Sony stacked chip system used, the image signal processor (ISP) can be either a homegrown (Sony) 65 nm chip or a TSMC fab’d 40 nm chip... My summary presentation slide and message was: if you aren’t excited about chip stacking, you should become excited about chip stacking!"

Ray's IISW paper can be freely downloaded from Chipworks site.

IISW 2015 Review

Albert Theuwissen posts his brief review and impressions from IISW 2015 held in Vaals, Netherlands last week. "In conclusion : no major new technologies were introduced, neither any pixel size below 1 um, but everything is getting better in performance and more compact in size."

SuperPix Presents "the First Excellent Price 8MP 1.12um BSI Pixel Sensor"

Beijing, China-based Superpix launches SP8408, "the first excellent price 8-megapixel CMOS image sensor chip, the product in the world’s leading automatic 12 inch BSI process line production, production process without human intervention, to ensure that the product has excellent consistency and stability.

The main performance index of SP8408 is better than other first-class 8M products in the market, the key index sensitivity is 9.5%, and high dynamic range (HDR) camera mode and low illumination enhancement mode. SP8408 pixel size of 1.12um, can be supports 1/4 inch 8 pixel optical lens. SP8408 capacity adequacy, as long as 2.5 months in advance to give demand forecast, can guarantee the stable supply of demand.

The company's LinkedIn article lists SP8408 features:

"The 1/4 inch 8-megapixel RAW data image sensor SP8408 is one of SuperPix’s upgrade products of SP84XX series color image sensors. It is a high performance sensor based on advanced 1.12um BSI pixel architecture. The SP8408 is a high cost-performance image sensor product that can be embedded in portable equipment, and is especially suitable for mainstream smart phones and tablet computer applications.

The SP8408 incorporates 3280 x 2464 effective pixels, advanced low power analog circuits (ASP). The on chip ISP circuits performs sophisticated signal processing including picture flip,auto defect pixel correction,picture flip and lens shading correction,etc.SP8408 supports high frame speed up to 30fps at full resolution, 60fps at 1080P, 120fps at 720p format transferred over a 4-lane Interface. The new "Bright Mode" feature delivers high-frame-rate(slow-motion)video without causing a luminance drop.
" (sounds a bit like Toshiba "Bright Mode.")

Another Superpix LinkedIn article presents SP0A38 VGA 2.2um pixel SoC.

2015 IISW Awards

Here is the list of the awards presented at last weeks IISW 2015:

IISS Exceptional Lifetime Achievement Award
This Award is made to a member of the image sensor community who has made substantial sustained and exceptional contributions to the field of solid-state image sensors over the course of their career. (Established 2013). The first recipient ; Gene P. Weckler (2013)

2015 Recipient : Dr. Jaroslav (Jerry) Hynecek
“For the invention and development of significant advances in solid-state image sensors including the VP-CCD, impactron, and BCMD devices”

(Jerry Hynecek received his Dipl. Ing. In EE from the Czech Technical University in Prague in 1962 and Ph.D. from Case Western Reserve in Cleveland Ohio in 1974. He joined Texas Instruments in 1976, becoming a TI Fellow in 1990. In 1998 he founded ISETEX and has consulted for several companies. He has published 61 papers and holds 126 US Patents, and among other awards, won the 2003 Walter Kosonocky Award for the Impactron CCD. He has worked on image sensors for most of career, and his seminal work on pinned CCDs and the use of impact ionization is reflected in most of today’s image sensors. He has also served our community through nearly 10 years of editorial work at IEEE TED, as well as technical program committees, and at IISW.)

IISS Pioneering Achievement Award
The Board of Directors of the International Image Sensor Society (IISS) is pleased to announce the creation of a new award, the Pioneering Achievement Award.
This new award is to recognize a person who made a pioneering achievement in image sensor technology, and judged by at least 10 years of hindsight as a foundational contribution.

2015 Recipient : Peter J.W. Noble
“For seminal contributions to early years of MOS image sensors”

(Peter Noble first worked in sensor arrays starting in May 1966, and rapidly developed the basics of MOS based active pixels and self-scanned arrays, now known as image sensors. The first was a simple 7 pixel linear array, then worked through a range of sizes. He set up his own company in the late 60’s and showed the first working moving image system using a 4,916 pixel square array in the early 70’s. For this his company was awarded the Queen’s Award for Technology in 1974. Throughout he looked at and developed other operating systems experimentally, some not published at the time. The company also made and supplied a variety of sensor arrays and systems. His overall ambition was, and still is, being able to give to blind human beings the ability to see well enough to appreciate the world around them. Peter Noble is a consultant to major companies on the strategy combining technology trends, economics and global situations. He is author of two books, and many papers, articles and patents.)

Walter Kosonocky Award
The Walter Kosonocky Award is presented bi-annually for THE BEST PAPER presented in any venue during the prior two years representing significant advancement in solid-state image sensors. The award commemorates the many important contributions made by the late Dr. Walter Kosonocky to the field of solid-state image sensors. Founded in 1997 by his colleagues in industry, government and academia, the award is also funded by proceeds from the International Image Sensor Workshop.

2015 Recipient : “A 1/4-inch 8Mpixel Back-Illuminated Stacked CMOS Image Sensor” ISSCC Dig. Tech. Papers, pp. 484 - 485, Feb. 2013.
Shunichi Sukegawa(1), Taku Umebayashi(1), Tsutomu Nakajima(1), Hiroshi Kawanobe(1), Ken Koseki(2), Isao Hirota(1), Tsutomu Haruta(1), Masanori Kasai(1), Koji Fukumoto(1), Toshifumi Wakano(1), Keishi Inoue(3), Hiroshi Takahashi(1), Takashi Nagano(1), Yoshikazu Nitta(1), Teruo Hirayama(1), Noriyuki Fukushima(1)
1 : Sony, Atsugi, Japan,
2 : Sony LSI Design, Atsugi, Japan,
3 : Sony Semiconductor, Kumamoto, Japan

Best Poster Award
The best poster presentation is selected by the participants of the IISW.

2015 recipient : “Analysis and Reduction of floating Diffusion Capacitance Components of CMOS Image Sensor for Photon-Countable Sensitivity
Fumiaki Kusuhara, Shunichi Wakashima, Satoshi Nasuno, Rihito Kuroda and Shigetoshi Sugawa, Toholu University, Japan

Thursday, June 18, 2015

Occipital Reveals its Structure Sensor Internal Design

A high-profile Kickstarter company Occipital making a structured light 3D sensor integrated onto iPads or iPhone 6 back, presents its internal design in an Youtube video:

Chipworks Publishes ST SPAD ToF Proximity Sensor Analysis

Chipworks has published ST VL6180 ToF proximity sensor reverse engineering report:

Intel Acquires Smartglass Maker Recon Instruments

Intel has completed the acquisition of Vancouver, BC, Canada-based smartglass vendor Recon Instruments for an undisclosed amount. Intel has invested in Recon in 2013, and has a time to learn about the company business. The recent Recon Jet glasses launched in April 2015 has an integrated camera, somewhat similar to Google Glass:

Wednesday, June 17, 2015

Mobileye and ADAS Market

SeekingAlpha publishes comments of Baron Capital's Ron Baron in CNBC appearance: 13 years of data-collection efforts related to driver-assistance technologies has given the company a "essentially a monopoly" position for its ADAS chips and software. "When there's a tilt in the road, these guys will know about it and able to have the car steer a certain way because they've collected millions of miles of traveling [data]..." Mobileye has relationships with 23 out of the 25 top automakers.

Edwin Roks Becomes Teledyne DALSA COO

Business Wire: Teledyne announces the promotion of Edwin Roks to the position of Chief Operating Officer of Teledyne DALSA. Reporting to Rex Geveden, EVP of Teledyne and President of Teledyne DALSA, Roks will direct the operational activities of Teledyne DALSA and its digital imaging, medical, X-ray and professional imaging, and semiconductor divisions. His responsibilities will include financial performance, strategic planning, business development and research and technology integration. Dr. Roks also remains a VP of Teledyne, and will continue to advise Robert Mehrabian, Teledyne’s Chairman, President and CEO.

Tuesday, June 16, 2015

0.3e- Noise Histogram

Eric Fossum kindly allowed me to post a histogram of his and Jiaju Ma's image sensor that apparently achieves 0.32 e- rms noise on a single CDS read. It was achieved by minimizing the FD node capacitance as discussed in the March 2015 JEDS article:

"A Pump-Gate Jot Device With High Conversion Gain for a Quanta Image Sensor" by Jiaju Ma; Fossum, E.R.

The low readout noise manifests itself in a comb-like structure on the histogram, where the distance between the peaks is equivalent to 1e-. This preliminary result was first presented at IISW 2015 last week:

More Details on Sony Fast Full-Frame and 1-inch Sensors

Imaging Resource publishes an interview with Kimio Maki, Senior General Manager, Digital Imaging Business Group, Sony, discussing various trade-offs in image sensors for A7R II full-frame camera and RX100 IV and RX10 II 1-inch compacts. Few interesting points, in no particular order:

  • Why Sony needs a backside illumination in its new 42MP FF sensor:
    (a) to match a pixel sensitivity of its earlier 36MP front side illuminated sensor
    (b) to be able to use metal routing more freely, without obscuring the light path
    (c) better metal routing, together with Cu instead of Al in the previous sensor allows to implement 4K video mode in that big sensor
    (d) the new sensor is 3.5x faster than the previous generation one
  • 42MP resolution has been chosen to ensure an easy image scaleability in various 4K video modes. If not that, Sony marketing would prefer 45MP or 50MP resolution.
  • Other than the speed, Sony has improved the noise performance in the new FF 42MP sensor over the older 36MP one.
  • The 14b raw file format suggests that 14b ADCs are used on the new FF sensor
  • The new 1-inch stacked sensor is 5x faster than the previous one
  • The fast readout speed creates close to global shatter experience, although the shutter is actually rolling. Sony calls is "Anti-Distortion Rolling Shutter."
  • To get to that high speed, the 1-inch stacked sensor employs no less than 16 ADCs per column (!)
  • Since the 1-inch sensor has 5,472 columns, the number of column-level ADCs is  87,552
  • The data from ADCs is transferred to the bottom DRAM die and, then, outside of the chip
  • The DRAM die also has a "low level" processor integrated on it

Sony 1-inch stacked sensor assembly. The chips are inverted
and connected to the bottom of the sensor package using some
sort of solder-bump technology.

Invisage Opens World's First Sub-5nm Fabrication Site

InVisage has inaugurated its first high-volume, fully automated QuantumFilm sensor manufacturing facility in Hsinchu Science Park, Taiwan. “After a review of various U.S. and Asian locations, we chose to establish our high-volume manufacturing in Taiwan because of the vitality of the semiconductor ecosystem here, and in particular because of our partnership with TSMC,” said InVisage’s CEO, Jess Lee. “Thanks to our high-calibre engineering and manufacturing staff here, we were able to source and install custom, state-of-the-art fabrication equipment and are now well positioned to ramp up QuantumFilm production.

The manufacturing facility, called QFAB3, is said to feature an unprecedented sub-5nm process geometry within an ISO Class 10 cleanroom. It is designed to support a wide range of products, from mobile phones to high-end cameras as well as drones and other IoT devices that require high performance cameras.

Because we use quantum dots, we don’t need finely patterned lithography to achieve our high performance and sub 5-nm scale. This QuantumFilm fab has a multiplier effect in terms of product performance, value, and output. So our factory is compact and accomplishes much more than the typical fab,” said James Chou, VP of Manufacturing Operations. “All the steps could fit within one shipping container. This enables many expansion opportunities we’re already scouting for.”

InVisage’s product wafers are first manufactured by nearby TSMC and then transferred to QFAB3 for QuantumFilm deposition. The deposition itself is performed by a single, custom tool that combines spin-coating and CVD deposition technologies into one machine for the first time. This combination allows the deposition process to be both modular and fully automated. This is said to be the first of many facilities to come as InVisage increases its capacity aiming to QuantumFilm technology to become the new standard for cameras.

Invisage Aims to Conquer the Imaging World

EET-China posts an article on InVisage progress and plans, also quoted by wtuitiao.com site. Jess Lee, Invisage CEO, gave an interview on the opening a new 6,000 sq. feet fab in Taiwan that will deal with adding Quantum Film on top of TSMC-produced wafers. Few quotes, in Google translation:

"Spend so much money gone? After the factory opening ceremony, "Electronic Engineering Times," given the opportunity and InVisage company CEO Jess Lee one interview, this is the first question I can not wait to put forward.

"Basic money spent these years of research above and the corresponding, because the TSMC outsource the manufacturing part of the silicon, so in fact our factory investment is not large. "InVisage all the issued and pending patents has reached 102 items.

And now Fab3 upcoming production, InVisage finally began to have an income. In addition, InVisage also won the two top camera's contract, and has received a total of 10 million dollars in the first phase of development costs. "However, if you need a more rapid expansion of production, we may also consider re-attract some investment." Lee expressed.

"In the semiconductor industry, which is the only one I've heard from inception to launch the first product requires 10 years of IC companies, and that decade also continued to be favored by institutional investors, with 100 million dollars was built out first a product."

Jess Lee has shown a small bottle of Quantum Film substance which is said to be enough to produce 10,000 image sensors:

Once Invisage's Quantum Film sensor reaches the market, it's projected to capture a big chunk of it:

"VisEra do this layer of the color filter, in fact, before I was a little worried, after all, is different from the silicon quantum film, will not be splitting ah, for specifically ask Mr. SC Schin VisEra, he said, "There is absolutely not, on the contrary, because the surface of the quantum film will be more smooth for us, but also easier."

"Pixel size is not a limiting factor, can be nanoscale, "but in order to cater to the mainstream, our first product will choose between the 1.1-1.4um until mature market culture Then, again using smaller pixel size. Lee said."

"For InVisage of Sensor, Lee has revealed that the main chip and the two major mobile phone manufacturers to develop, integrate optimized for InVisageSensor ISP in its main chip."

Thanks to HTL for sending me the link!

TSMC Looks to Expand its Image Sensor Presense

Digitimes reports that TSMC is looking to expand its presence in the CIS field, according to the newspaper's sources. TSMC gets orders from Aptina and Galaxycore, and partners with Invisage. "TSMC will also be able to secure orders from OmniVision, if the fabless firm wants to stay competitive in the mid- and high-end market segments."

"In other news, InVisage recently announced the inauguration of a manufacturing facility nearby TSMC's manufacturing site at the Hsinchu Science Park, northern Taiwan. "After a review of various U.S. and Asian locations, we chose to establish our high-volume manufacturing in Taiwan because of the vitality of the semiconductor ecosystem here, and in particular because of our partnership with TSMC," said InVisage CEO Jess Lee in a statement.

The facility, dubbed QFAB3, will be engaged in the manufacture of CIS based on InVisage's proprietary QuantumFilm technology. "InVisage's product wafers are first manufactured by nearby TSMC and then transferred to QFAB3 for QuantumFilm deposition," said InVisage.

Caeleste Publishes its IISW 2015 Papers On-Line

Caeleste puts its IISW 2015 papers on-line:

1. "CMOS image sensor reaching 0.34 eRMS read noise by inversion-accumulation cycling"
Qiang Yao, Bart Dierickx, Benoit Dupont, Gerlinde Ruttens

Record for lowest read noise in CIS. CMOS pixels is pushed a little bit further again, as reported in our 3rd IISW paper. The value of 0.34 (sic) electrons RMS is obtained at the cost of accumulation-inversion cycling, severe oversampling and cooling down to -40°C.

2. "Four Concepts for Synchronous, PSN limited, true CDS, HDR imaging"
A.K.Kalgi, B.Dierickx, B.Dupont, P.Coppejans, P.Gao, B.Spinnewyn, B.Luyssaert, A. Defernez, J. Zhu, J.Basteleus, Q. Yao, W. Verbruggen, D. Uwaerts, B. Uwaerts, G. Ruttens, G. Cai

Comparing 4 HDR methods. Here we use a pretty strict definition of “High Dynamic Range”: It must at the same time comply with: 1) knee-less fully linear 2) allowing RWI synchronous shutter and 3) complete CDS. 3 methods are actually Caeleste patents, the 4th is the historical “overflow barrier” method.

3. "Imaging sparse events at high speed"
Gaozhan Cai, Bart Dierickx, Bert Luyssaert, Nick Witvrouwen, Gerlinde Ruttens

Ultra-high speed readout of a sparse image. The first imager to realize ultra-high speed sparse readout of a megapixel small-pixel (8µm) array, by copying a “SLICE” of the pixel array data into a processor array. The below 1Mpixel prototype is able to output about 8000 frames per second.

Monday, June 15, 2015

Omnivision Announces 400 x 400 Pixel 1/18-inch Micro Sensor

PRNewswire: OmniVision announces a new ultra-compact CameraChip sensor for medical and industrial applications. The 1/18-inch OV6946 captures 400 x 400 pixel resolution images and video in an extremely compact 0.9 x 0.9 mm package, making it an ideal camera solution for minimally-invasive, reusable endoscopy.

Utilizing a 1.75um OmniBSI+ pixel, the OV6946 captures video at 30fps. The OV6946 enables endoscope modules with a width of 1.65 mm and height of 5 mm. Additionally, the sensor's reduced pinout and basic ISP functionality offer extremely easy integration. The OV6946 is currently available for sampling and is expected to enter volume production in the third quarter of 2015.

The correct pixel size is 1.75um

PRNewswire: OmniVision partners with preferred solution providers Precision Optics Corporation, Inc. (POC) and Fujikura Ltd. to develop the 160K (400 x 400 pixel) camera module, a quick go-to-market camera solution for OEMs. POC is a leader in designing and manufacturing micro-optics and associated medical imaging devices, and Fujikura is a leading miniature CMOS imaging module manufacturer.

Based on OmniVision's new OV6946 sensor, POC's Microprecision lens technology, and Fujikura's flexible micro-coaxial cable technology and small packaging capability, the 160K CIS module provides high quality images and video with a distal end diameter of only 1.60 mm. Evaluation units of the 160K CIS module are currently available, and volume production is expected to begin in 2016.

POC shows a nice demo video with the previous generation OV6930 sensor:

Sunday, June 14, 2015

TowerJazz PDK Adds Support for Cadence Layout EAD

TowerJazz announces its support for the Cadence Virtuoso Layout interactive Electrically Aware Design (EAD) for all of its 180nm processes, including TS18IS (image sensor) process.

Harvest Imaging Forum Agenda Published

Albert Theuwissen publishes a tentative agenda of this year's forum "3D Imaging with Time-of-Flight: Solid-State Devices, Circuits and Architectures" delivered by David Stoppa:

  • Applications for 3D image sensors
  • Range sensing technologies: taxonomy, pros and cons
  • Active range sensors (Triangulation, Interferometry and Time-of-Flight): operation principle and solid-state implementation
  • Time-of-Flight measurement principle and implementations
Time-of-Flight System
  • Building blocks
  • Optical power budget and system design considerations
  • Technical requirements for 3D Time-of-Flight image sensors
  • History and evolution of 3D image sensors versus 2D
Detector Technologies: Photo-demodulators (PDM), Electronic Shutter (ES), SPADs and APDs
  • Evolution of PDM-devices from CCD to CIS
  • Description and analysis of State-of-the-art PDM implementations
  • Description and analysis of SoA ES solutions
  • From photodiode to SPAD through APDs
  • SPAD operation principle and behavioural model
  • SPAD front-end and processing electronics
  • Description and analysis of SoA SPAD-based solutions
  • Pros and cons of each detector technology and best application scenarios
3D Image sensors readout architectures and main processing blocks
  • Sensor architectures, column amplifiers, ADCs, on-chip processing blocks
  • Extra electronics needed by 3D Time-of-Flight with respect to 2D (TDCs, TACs, drivers)
Main noise sources in 3D imagers
  • Noise analysis and case study for PDM, ES and SPAD sensors
3D Image sensors characterisation and Figures of Merit
  • How to test a 3D Time-of-Flight imager
  • Main Figures of Merit
Additional topics
  • Best 3D sensor technology for your application
  • System aspects and trade-offs
  • Future perspective and roadmap

Saturday, June 13, 2015

ON Semi-Aptina and Sony Think Alike

Sony patent application US20150156387 "Image sensor, image sensor operation method, and imaging apparatus" by Daisuke Miyakoshi proposes to split the continuous exposure into exposure on/off pieces to reduces flicker and jerkiness artifacts associated with short exposure times in bright light:

Aptina patent application US20150009375 "Imaging systems with dynamic shutter operation" by Gennadiy Agranov, Sergey Velichko, and John Ladd proposes, basically, a similar thing, albeit with no storage cap in the pixel:

While we are at Aptina, SeekingAlpha's ON Semi earnings call transcript for Q1 2015 praises its performance:

Keith D. Jackson - President, CEO and Director
In the automotive segment, our momentum in fast growing advanced driver assist systems market remains intact, and we continue to be market leader with approximately 70% share for image sensors for ADAS-related applications.

Q: John Pitzer - Credit Suisse
I was wondering if you could spend a little bit of time just talking about the competitive landscape for Aptina. I mean clearly the image sensor marketing in auto ADAS is an area of pretty strong strength, but we are getting some signals of sort of industry capacity going up on the image sensor front. How do we think about that relative to the competitive dynamics? And can you talk a little bit about sort of the capacity needs that you have for your own imaging sensor business?

A: Keith D. Jackson - President, CEO and Director
Well, there are two other very strong players both of which have very significant handset exposure and the capacity on the handset side remains I think a key area for expansion. As you look at the automotive, it’s more than just the ability to produce pixels. You have to have the entire package for the automotive wins, particularly in safety critical applications. And we think we continue to have an edge there with our many years of experience. Relative to capacity, we have been working to increase capacities both internally and with our external partners, and are very comfortable we can see very strong growth and share gains supported for our image sensors this year.

Q: Ian Ing - MKM Partners
...my follow up for image sensors, could you talk about the implications of OmniVision getting acquired by a China entity, would that be a positive or a negative? Would that make the China business more competitive, perhaps favoring a local supplier?

A: Keith D. Jackson - President, CEO and Director
I think over time there can be some influence on the market from a favored China supplier. I think in the short term, there’s going to be lots of confusion and changing; some announced intension to change of manufacturing, et cetera. So we’re looking forward to some short-term opportunities. And then we’re just going to keep driving on the technology side so that our products basically will be favored and the local favoritism won’t matter.

IHS Sees Drones as a Big Image Sensor Market

IHS Electronics360: IHS believes that a rise of low-cost consumer and industrial-oriented drones opens a big market for chip makers in general, and, specifically, for image sensor vendors. Few quotes:

"Today, the major UAV vendors include DJI, a Chinese company expected to reach $1 billion in sales this year; Parrot in France; and 3D Robotics in the U.S. Larger companies are trying their wings as well. Google bought drone-maker Titan and plans to start testing drones later this year. Amazon Prime Air plans to use its drones for package delivery."

"In many consumer drones, the camera function is handled as part of the microprocessor and so a separate camera chip is not needed. Depending on which chip vendor is talking, they refer to drones as either “flying smartphones” or “flying cameras.

"DJI designed Ambarella chips into its Phantom line of drones 18 months ago. Although it is a new market for Ambarella, sales have “really, really taken off,” says Chris Day, Ambarella’s vice president of marketing and business development. “It’s not just a niche. We think by the end of this year, it’ll be significant.

"Ambarella chips produce high quality video in very low light conditions and incorporate “dewarping” technology that can correct distortions from wide-angle lenses, allowing for better discrimination of detail, he says. They also have frame rates as high as 60 frames per second, enabling smoother video of fast-moving events like sports."

A Cnet video explains the 4K/30fps camera operation of the latest DJI Phantom 3 drone:

Friday, June 12, 2015

Yole on Omnivision Acquisition

Yole Developpement publishes its analysis of Omnivision acquisition by Chinese investment funds. Few quotes:

"...now OmniVision is Chinese. This is certainly good news for the industry since it should renew competition facing a much more dominating Sony. If you are looking at the recent changes in the situation, there are 3 aspects of competition at play and OmniVision management is playing very well:

First, OmniVision management must fix the technology gap by massively investing in technology with foundry partners TSMC and now with their move to use XMC as main foundry partner.

Second, they must build market partnership and tie themselves closer to loyal customers. Smartphone cameras have become so important that trust and reliability has become essential.

Third, they must get access to capital in order to finance technology and product development in a quickly evolving market.

By becoming Chinese, OmniVision is providing the right answers to those 3 aspects.

Thursday, June 11, 2015

Yole Interviews Seek Thermal

Yole Developpement publishes an interview with Seek Thermal founders. Seek Thermal sells smartphone attached thermal cameras based on Raytheon microbolometric sensors. Yole forecasts a growth of this market with quite a few companies offering their products:

Mentor Graphics CIS Verification

Mentor Graphics publishes a whitepaper on using its FastSpice transient noise analysis for image sensor simulations. Some of the Mentor's claims are disputable, but probably nobody expects a deep understaning of CIS issues from an EDA company:

"Another type of temporal noise source is random telegraph signal (RTS) noise, which introduces “blinking pixels” from which the output exhibits three discrete levels. However, RTS noise does not appear to be a limiting factor on the performance of today’s image sensors."

Sony Unveils: Full Frame BSI Sensor and 1-inch Stacked Sensor

PR Newswire: Sony announces a mirrorless Alpha 7R II camera featuring a world's first full frame BSI image sensor. The newly developed 42.4 MP back-illuminated CMOS sensor is said to be "the most advanced, versatile and highest resolution full-frame image sensor that Sony has ever created."

The new sensor combines gapless on-chip lens and AR coating on the surface of the sensor’s glass seal to improve light collection efficiency, resulting in high sensitivity with low-noise performance and wide dynamic range. This allows the camera to shoot at an impressive ISO range of 100 to 25600 that is expandable to ISO 50 to 1024002.

Additionally, the sensor’s back-illuminated structure, with an expanded circuit scale and copper wiring design, enables faster transmission speed, approximately 3.5x faster than the original α7R.

The new full frame sensor features 399 focal-plane phase-detection AF points, said to be the world’s widest AF coverage on a full-frame sensor – that work together with 25 contrast AF points to achieve focus response that is about 40% faster than the original model.

The sensor supports 4K (QFHD 3840×2160, 30fps) video recording in either Super 35mm crop mode or full-frame mode. In Super 35mm mode, the camera uses information from approximately 1.8x as many pixels as 4K by using full pixel readout without pixel binning and oversamples the information to produce 4K movies with minimal moire and ‘jaggies’.

In full-frame mode, the α7R II utilizes the full width of the 35mm sensor for 4K/30fps recording. It is said to be the world’s first digital camera to offer this in-camera full-frame format 4K recording capability.

PRNewswire: The second Sony announcement is the world's first 1-inch stacked sensor featuring in compact DSC-RX100M4 and DSC-RX10M2 cameras. The stacked Exmor RS CMOS sensor is said to have "advanced signal processing and an attached DRAM memory chip. The high speed signal processing and DRAM memory chip work together to enable more than 5x faster readout of image data and are responsible for a variety of standout features that have previously been available in only a select few professional-level video cameras. These impressive capabilities include 40x super slow motion video capture at up to 960 fps, an ultra-fast Anti-Distortion Shutter with a maximum speed of 1/32000 second, high resolution 4K movie shooting and more."

"Additionally, the ultra-fast readout of the image sensor is responsible for the high speed Anti-Distortion Shutter (maximum speed of 1/32000 second), which allow the new cameras to capture sharp, crystal clear images with a wide open aperture at brightness levels up to EV197. It also minimizes the “rolling shutter” effect commonly experienced with fast moving subjects."

DPReview states that the 4K video mode frame rate is 30fps, the same as with the previous generation non-stacked 1-inch sensor.

Sony posts Youtube videos demoing the new sensor slow-motion video capabilities:

Update: Additionaly, Sony says it "is the world’s No. 1 manufacturer of image sensors for digital cameras and video recorders, with over 50% market share. Data based on Sony research from April 2014 to March 2015."

Imaging Resource posts pictures taken on today's Sony press conference in NYC:

Update #2: DPReview writes "Internally, we're told, Sony's engineers judge that their new 1-inch-type stacked CMOS sensor is roughly five years ahead of anything else on the market, and on paper at least, it's hard to argue with that assessment."

Tuesday, June 09, 2015

Sony Aims for 20% Image Sensor Sales Growth This Year, Following 40% Growth Last Year

Reuters: Tomoyuki Suzuki, head of Sony's device solutions business, said he expected sensor sales to grow about 100 billion yen ($804 million) to 550 billion in the year ending March 2016. That would be slower than the 40% rise of the previous year. But Suzuki said demand was so strong that Sony was struggling to keep up.

"We're seeing very good demand at the moment," he told Reuters in an interview. "We don't have slack."

Apple is Sony's top customer, followed by Samsung. Chinese handset makers such as Xiaomi also use Sony sensors for high-end models, analysts said.

"We want to be inside a variety of customers," Suzuki said. "When it comes to semiconductors, if you can't make use of production capacity you quickly end up with a loss. So if you want to avoid that volatility the important thing is to have good balance with several customers."

"Of course we will meet the requirements of our top customer, but we are expanding capacity with orders from Chinese smartphone makers in mind," Suzuki said.

The superior low-light sensitivity would give Sony an advantage as it expands into automobile-related products, Suzuki said.

Espros Presents its ToF Advances

Espros Photonic sent me its June 2015 Newsletter which, I guess, will eventually appear on its web site. The company CEO, Beat De Coi, tells about their ToF sensor product, and promises to tell more in the next Newsletter edition:

"The ESPROS TOF technology is gaining huge momentum. The delivery start of the 3D-TOF QVGA imager was excellent. Customers immediately recognized the performance of our OHC15L technology. A QE of more than 80% at 850nm wavelength and pixels with 100% fill factor provide unmatched sensitivity with very low illumination power. A reference design of a camera engine with a horizontal FOV of 94° achieves a 10m range in the full field. Ok, this is on a white target. But the exposure time is a few milliseconds only."

Espros also opens a Youtube channel with a video demo of its epc660 ToF imager:

First 8K Video on Youtube

No Film School: It appears that Youtube has quietly prepared an 8K video infrastricture. The first such video has been published, made by stitching of two 6K videos streams shot on RED Epic Dragon 6K camera:

PDAF vs Contrast AF vs Laser AF

Anandtech posts a nice chart benchmarking AF speed of different smartphones. The top three group on the chart features phase detect pixel AF, but farther down, the results are less conclusive:

Correction: Samsung Galaxy S5 has PDAF too - added to the picture. This makes the chart even less conclusive.

Saturday, June 06, 2015

SETi Reverse Merged with G-SMATT

I've just noticed that a year ago Korean SETi had been reverse merged with G-SMATT Global. The LED panel glass manufacturer G-SMATT is now trading under the former SETi ticker at KOSDAQ. G-SMATT is still selling the older SETi image sensors, but its main business is in glass manufacturing. This is pretty sad news for a company that once controlled a significant portion of Chinese image sensor market. Thanks to BS for the info confirmation!

In spite of SETi demise, Korea still has a relatively large number of smaller image sensor companies: Pixelplus, Clairpixel, ZeeAnn, Maru LSI, and Rayence.

Update: Rayence has been acquired by GE Healthcare in 2013. Thanks to RT for the info.

ON Semi-Aptina Proposes Variable Size CFA, Digitally Corrected SAR ADC

Aptina patent application US20150146054 "Image sensors with color filter elements of different sizes" by Brian Vaartstra, Mitchell J. Mooney, and Stanley Micinski describes the small pixel challenges "Typically, each color filter element in a conventional color filter array has the same size and shape (i.e., each of color filter elements R, G, and B are square and occupy the same area). As pixel size becomes smaller with each successive generation of imaging sensor devices, one problem that may arise is the diffraction limit of red light. Due to this diffraction limitation, the quantum efficiency may be significantly degraded for pixels that are smaller than 0.7 microns on one side (as an example). It may therefore be desirable to form color filter elements of varying sizes optimized for the wavelength(s) of interest."

So, Aptina proposes to increase red color filter at the expense of the blue one:

"Color filter element 102-X may be a yellow color filter element, a green color filter element, a cyan color filter element, a magenta color filter element, an infrared-pass filter element (e.g., a filter that passes infrared light), an IR-block filter element (e.g., a filter that blocks infrared light), a clear color filter element (e.g., a filter that passes all visible light), or any combination of color filter elements (e.g., in each unit cell, one X pixel containing a green color filter and the other containing an IR pass filter)."

Another Aptina patent application US20150146055 "Imaging pixels with improved analog-to-digital circuitry" by Parthasarathy Sampath proposes a digital correction to compensate the capacitor mismatches in column-parallel SAR ADC: