Forward
This is quite a lengthy post (again). I have written it to develop my own understanding as much as to share these ideas fellow forum members. It is arranged in titled sections for quick reference. Hopefully, it meets its task in describing the future digital photography as I believe is envisioned by Fujifilm and Panasonic.
Introduction
Ever notice that when the future is mentioned the biggest interest is what lenses are coming next. And, that when the X-Pro 2 did not appear, it has been "naturally" assumed this must mean a FF or at least an FF X-100 equivalent. While it is pleasant recreation to visualize the next new lens announcement from Fujifilm, at the same time, the idea of a FF camera from Fujifilm has spawned endless debate. There is even some acrimony in a few individuals for the X-Trans sensor being APS-C size in the first place. Frankly, it has occurred to me that these folks may well be "barking up the wrong tree". (Non-native English users may highlight and wiki as needed).
The Fujifilm-Panasonic organic sensor project is continually glossed over in various threads as some tangential new gimmick of little consequence. I suspect this a strategic error of a high order.
Furthermore, my feeling is Fujifilm's "maybe in two to three years" answer about whether they will have a FF camera is just another example of polite Japanese manners. That just like an earlier answer about FF camera possibilities: It is not to be taken literally. In both cases their answers merely let people hear what they want to hear.
What is it they do mean? Probably some coy thought along the lines of, 'To be to honest about it, we have something different up our sleeve'. If this is true, that something different is the organic sensor project: Their polite answer merely indicates the sensor project is at least 24 months to completion. After all, saying "no FF camera for at least two to three years" is not really saying much, now is it?
Synopsis
There is going to be a new technology for sensor construction. It's major difference to present day sensors is the replacement of a relatively large and noisy photo sensitive diode within each pixel with a micro-thin, photosensitive organic membrane that covers the whole surface of the sensor. Fujifilm has perfected this organic photo sensing material. In partnership with Panasonic, who have perfected miniaturization of the remaining electronic bits of sensor technology, they have agreed to finalize production technology that combines their expertise into a new type camera sensor.
The broader implications of an organic sensor concept are drawn from the investment community: They anticipate a wide ranging organic sensor revolution, not only for digital photography, but across all sections of photo sensing applications from the automotive industry to medical equipment. The automotive industry is expected to be the largest consumer of these devices, not the camera industry nor even the personal device industry. Some say medical product development will hold second place. This sounds like a really big deal.
The New Sensor's Secret
There are two critical factors dependent on sensor size: Resolution and noise. They technically equate to pixel density and light gathering ability. Today, we do not want pixels too small because this affects light gathering ability. With present sensor technology, the relatively small senior area and high pixel density of MFT (Micro Four Thirds) has visible limitations with respect to resolution and sensor noise. With at APS-C size, the noise issue can be well controlled at 16mp, but resolution is less than a FF sensor, due to the lower resolution of the smaller sensor area APS-C camera.
However, if we want to be precise, we should say the limiting factor for resolution and sensor noise is not sensor size, rather it is the smallest dimensions with which we can make a high performance pixel or photosensitive cell. We are so locked into to the current sensor technology that this observation amounts to "outside the box" thinking. But, this is where we need to go.
Existing sensor technology quickly faces a situation where too high a pixel density in too small of an area lowers the per pixel light gathering area such that the sensitivity of the pixel, its s/n (signal to noise) ratio is diminished. This causes visible noise. Both dynamic range and high-ISO performance are thus limited by noise when we down-size sensor area. This has long been the reason FF cameras are favored for professional use.
The big secret of the organic sensor is s/n (signal to noise) ratio of each pixel is improved to 88db. Along with this decrease in noise, the nominal pixel size has been reduced by a factor of 2. The technical advantages of today's FF and APS-C sensors, the of best possible dynamic range and low level noise performance, will met by a much smaller sensor area than is presently possible. At the same time, resolution will be increased for a given sensor area.
These are the two implications brought to light from the very little information we have about the Fujifilm-Panasoinc organic sensor project.
Scope Of The Fujifim-Panasonic Project
Some time ago Fujifilm developed an ultra-thin, 0.5 micron layer of photosensitive organic film. Yes, literally photosensitive film. We are going back to film, after all. It is an unbelievably thin, photosensitive, organic based polymer. Organic, as in not an etched silicon electronic circuit that must be made with painstaking accuracy. Instead, it is a film that is relatively inexpensive to produce. A polymer film, as in similar to, but very much thinner than plastic wrap used in the kitchen. It is a flexible photo sensitive membrane that emits an electron charge when exposed to light. In our application, it is panchromatic and is coupled from below to traditional, but miniature, silicon etched, charge measuring, electronic transport circuits.
In its broadest sense, this architecture is an extension of the BSI (back side illuminated) technology introduced by Sony in 2009, and subsequently used in Fujifilm EXR pocket cameras. BSI was not used on larger sensors because this technology, by itself, was only an advantage in miniature sensors where pixel size was greatly reduced. But, with organic sensors, pixel size will see a 2x reduction and the BSI concept now becomes, by definition, the required construction format: There there are also longer any electronics in the uppermost, organic photo sensing or layer. All of the electronic part lies underneath.
Panasonic specializes in and continues to miniaturize all kinds of electronic circuits. Sensors consist of a electronic circuit capable of collecting the electric charge from, in this case, organic film, by way of gathering electrons in a capacitor "bucket". The electronics then burst the signal (our photo's pixel information in unfiltered form) through the rest of the A-D (analog-digital) conversion circuit. This all happens in micro-slices of time.
This A-D conversion process in digital photography records light waves of an image and is analogous to recording sound waves of music electronically. The two processes, A-D conversion of digital imaging and of electronic sound recording, have strong parallels we will exploit in explaining organic photo sensors.
The principle difference is sound A-D conversion is time variant and light A-D conversion is time invariant. This difference is a sometimes difficult switch of the imagination, but once accomplished, anything one knows about music reproduction and filtering has a perfect analog in digital image capture. Just remember the only real difference is a sound wave signal varies and is measured along a time line, while with light wave measurement, the signal varies and is measured over the area of the sensor at one point in time. Hence the term 'time invariant'.
The miniaturization of current CMOS photo sensors has always had a built-in roadblock. The photosensitive mechanism of conventional digital photosensors is an electronic component that can only be reduced in size so far. This is why FF and APS-C have superior noise qualities to smaller sensor cameras. Go any smaller in pixel size and noise becomes visible in shadow areas in the resulting image. The organic photosensitive layer developed by Fujifilm was the long awaited green light for further miniaturization of pixel size and its attendant electronic circuit.
As I understand the Fujifilm-Panasonic R&D agreement, they are not just trying to figure out if organic sensors will work. Both Fujifilm's organic layer and Panasonic's miniaturization of the relevant electronics already work. They are in the process of designing the production machinery to make organic sensors while they finish the details of the sensor's actual topology. There are several issues to be worked through, but they are spending money hand over fist to get this concept to into working products. "Sooner than later" is the operative word for concrete results.
It may be seen from the introductory comments about quotes from Fujifilm engineers that "sooner than later" could still mean a couple dozen months to the future. When Fujifilm engineers are questioned in public about a FF sensor they are saying, "two to there years" in response. I reason they also know something is coming within that time frame that will affect the evident superiority of FF sized sensors: Organic photosensor technology.
Impact Of The Project
When reducing sensor size, trying to keep acceptable resolution in small sensors always makes relatively more noise in the shadows and poorer high ISO performance. We have seen incremental improvements make APS-C the performance standard for the majority of amateur, any many professional photographers. However, noise in the smaller MFT (Micro 4/3) and particularly 2/3" sensors is still a problem. MFT has limited resolution and generally poor high ISO performance in comparison with APS-C. The 2/3" sensor additionally has both even less resolution and more noise in its images at any given ISO rating.
A single, brief media report made this July references two short papers given at a conference. One by Fujifilm group and one by Panasonic group. From this conference, the press release describes to the public the basic organic sensor topology for the first time. The reaction from the digital photography press has been tepid at best.
What is immediately apparent with the organic sensor concept is, while the tradeoff between pixel density and noise remains, the sensor size at which this trade-off becomes problematic is dramatically reduced. In short, based on the first bits of information about the organic sensor, we can envision a fundamental shift towards smaller, less expensive sensors of very high performance.
What we are discussing here is a change in technology that will alter the relationship between light gathering ability and sensor area. At the same time it radically alters our notion of pixel density. Smaller area sensors will become more very image makers with the new organic technology because pixel size related noise characteristics are changed. The impact could be quite dramatic.
Organic Sensor Details
The first fundamental change is the effective pixel size between organic sensors and silicon sensors has been shrunk by an order of at least two times. The organic sensor is being designed with cells (or pixels) of between 3.0µm (micron) and 0.5µm. The average FF sensor pixel spacing is ~ 8.0µm and the average APS-C sensor is ~6.5µm., while the 2/3" sensor (Fujifilm X20) is ~2.73µm.
But, wait, why bother? The 2/3" sensor, the smallest silicon cell of reasonable photographic quality is roughly the size of the largest announced organic cell. On the face of this information, such as small pixel size means the organic sensor project is suitable to replace only the smallest sensors in current use. This first glance is miss-leading. A 2x pixel size reduction merely sets the stage for the second, and the truly remarkable feature of the organic sensor.
This second feature is a most startling breakthrough. The new organic sensor's s/n (signal to noise ratio) is given at 88db. Currently the best s/n is the 43db of the 36mp Nikon D800, which happens to use about the same size sensor photo cells as the 16mb Fujifilm X-Trans. Making the X-Trans is not far behind in s/n ratio, it just has less resolution for its 16mp.
Well, half the noise, you say? No, 88db is not twice as low noise as 43db. db values are an exponential series. The noise reduction is on the order of 40 times less noise! Four full stops less noise.
Comparative Analysis
Pause and reflect……..Today's electronic 16-bit music signals have a 98db s/n potential. The first music reproduction went from 28db of distorted and noisy sound on 1910 wax cylinders, to 30 db of early, noisy shellac records of the 1930's. Then, it went to a quiet 58db of vinyl in the 1950's through to today's 98db digital reproduction. That last change allowed music to have, not so much less perceived noise than a good vinyl record, rather it allowed music to be presented with its original dynamic range intact. This is precisely what 88db digital photography means: Superior dynamic range and high ISO performance. Note this is achieved with a 2x smaller pixel.
At an estimated 14 stops of dynamic range, we will no longer be subject to choosing between blown highlights or a lack of shadow detail. High ISO performance, which has the effect of reducing dynamic range and increasing shadow detail noise, will be dramatically improved. If all is equal with the new type sensor, the ISO 1600 image quality of the Fujifilm camera will be possible at ISO25600. If this scale up of s/n ratio holds, , this thing will practically shoot by candle light, hand held. I will believe it when I see it!
If a vinyl record is played for someone under thirty, they are shocked at how truly good vinyl can sound. Just as present day camera's s/n ratio is just good enough for an excellent image, so it was with music on vinyl records. But, clearly the difference between an organic sensor and a conventional CMOS sensor at 2/3" or MFT size is much more striking, on the order of the improvement from 1930's shellac 75rpm records to digital: Supremely better. For larger sensors, such as APS-C, the noise reduction has a sweeter analogy: We are going to the equivalent of quality vinyl to the quality of digitally re-produced music, but this time for our eyes instead of our ears. An HDR like image quality.
Much of what can be said presently is somewhat speculative. We have precious little information to go on. This is a new sensor technology being co-developed Fujifilm and Panasonic. Not research and development by Canikon, rather by a pair of camera makers with a huge vested interest in APS-C and MFT sensors. These are the two sizes the new sensor is designed for. The implication is APS-C, currently thought of as the 35mm film camera of the digital era, is about to become the FF (full frame) film camera of the digital era.
We do not have a detailed picture of the new type sensor's capabilities, especially what novel and new firmware it may make possible. We are also unaware what technical difficulties have to be yet overcome. With that said, we can make some very rough comparisons and indicate just how important this new sensor technology is for Fijifilm's APS-C and Panasonic's MFT cameras.
First, recall that a 16mp APS-C sensor with a 6.42µm pixel spacing has about the same pixel spacing as a 36mp FF Nikon D-800 camera. What happens if we apply the largest announced, 3µm size organic cell to APS-C? An APS-C camera can be sized up to a theoretical 75mp with 14+ stops dynamic range.
This is not saying that the next APS-C camera from Fujifilm will have a 75mp sensor. It is just saying it could in theory be done from the information that has been provided about the organic sensor project. The organic photosensitive membrane no long needs to use space for conventional shrouding to prevent cell to cell optical cross-talk interference and it is continuous across the entire sensor surface. Some have even suggested such a sensor architecture need not be "pixelated" at all. As tantalizing as that may be, we must unfortunately table discussion of working megapixel numbers and filter array construction details until more information is released about the organic sensor's practical capabilities.
What we do know is, whether technology proves 24mp, 36mp or 72mp to be optimum, there will be plenty megapixels in the organic sensor equipped, APS-C camera. We can note that MFT is also in similar green pastures, up to 42mp possible in theory, although this too is difficult to imagine.
A larger sensor camera, such as FF will be again that many more pixels, but it will not have any better dynamic range nor will such ultra-high resolution be explicitly practical as against costs and processing overhead. The file sizes for data would be a bottleneck for general photography and, as many have observed, there is a widespread lack of suitable display media for such a fine detail resolution. With that said, such an ultra-high resolution FF camera be off into 4x5 film territory. Nice resolution to have, but not essential to everyday photography.
We do not even know if organic sensors are fast enough to be made at FF and larger sizes. Over 100 million pixels is going to take parallel processing and other tricks to get a sensor of such pixel density and size to even work at all. They would certainly demand different and more expensive electronics. So, FF organic sensors are, at least on first analysis, an open question.
The real surprise will be the 2/3" "pocket" camera. It is conceivable to have 16-20mp and ~14 stops dynamic range, a huge improvement: The new "APS-C" size of cameras!
Light Waves And The Organic Sensor
A nano scale device such as a 3µm (micron) sensor is only a few times the photon wavelength. I'm not sure how or what sort of quantum effects come into play specific to photo sensors, but intuitively 3µm is still pretty large versus the diffraction limit. This is roughly 1/2 wavelength, so 400nm for red or 0.4micron.
As to working with photosensors this small, when quantum effects do come into play there are ways to further improve resolution by physically shaping the photodetectors (e.g etching of the electronic well type structures). Because of the near to sub-micron size sensor cells announced, it must be precisely this interplay of quantum effects that Fujifilm and Panasonic engineers are crafting. This truly is cutting edge.
The joint project must fit the underlying electronics to the organic polymer layer in what we should think of as a hybrid structure: The continuous organic photosensitive receiver as the top layer and conventional, but highly miniaturized electronic circuitry on the lower layer. The top organic layer gathers photon energy and emits electrons, the bottom electronic component captures this electron flow and performs the A/D conversion. All on an almost unimaginably small scale of both time and distance.
In the end, the organic sensor is simply quantifying the photon energy striking the photosensitive film by outputting electrons to what is a varyable energy capacitor "well". This "well" discharges electronically a signal wave form several times as the exposure is made, emitting a complex digital signal.
While the equipment of the pixel can now become extremely small in size, it remains an orderly measurement of relatively large photons moving at the speed of light. The process is sampling photon wave energy and the converting this energy value into much smaller, time discrete electrons particles, 1/1200th the size of photons. A digital signal, now moving rapidly through electronic circuitry, is processed by a highly efficient digital filter. This separates out the high frequency chrominance (color) portion and the low frequency luminance (intensity) of the signal burst. Firmware can output to RAW or straight to a JPEG or TIFF image.
Implications for optics
Although very small in scale, the organic sensor technology is really just so much applied physics. However, the increased pixel density of nano scale of organic sensor technology makes diffraction limits of camera lenses the next big issue. To sustain ultra high pixel densities suggested by organic sensor technology, new demands will be made of lens optical glass and coating formulas. This is a completely separate area of research now made ever more important by the proposed organic photosensor technology.
Costs of using exotic glass formulations and coatings with rare earth elements drive up the price of the highest quality lens making. The truly exotic end of glass making is part of the reason some of the latest Leica lenses are so expensive. While we may question the relatively high price of Fujinon lenses, from this perspective we can appreciate some of the cause. Current and future lenses have to be optically excellent just to support the resolution capability of the organic sensor.
It is very interesting to note that Fujifilm as just introduced firmware to address lens diffraction problems into the X-E2's X-Trans II sensor and processor. This is a technology that would help ultra-high pixel density APS-C cameras become a practical device.
With regard to diffraction in objective lenses, a second feature of really small pixel cell sizes is that organic film micro lens architecture accepts a very much wider angle of incoming light. With wide angle light acceptance and their small size, this may add some additional flexibility to future lens design. Perhaps novel new lens features and firmware will appear to aid the looming problem of diffraction control.
Conclusion
I have already pointed out it is entirely possible that Fujifilm is saying "no FF for at least two years" for a slightly different reason than anyone would imagine. Now we can see the whole picture: They do not expect to need a FF size sensor to compete in the main stream photography market of the next decade.
Standing back from it all, I get the feeling the organic sensor is the sub-text of their current message. What must they be thinking when pressured by press and public for a FF Fujifilm camera time line….There is a 'Catch-22' quality about all of this Q&A they are put through.
At the same time, Fujifilm and Panasonic do not want to jinx their project by making any definitive announcements or hastily committing to a specific time table. The result is certainly still multiple months away. I gather from the things being said, as well as things not being said, the project is not yet tantalizingly close to fruition. It may even be still at the stage where it risks of becoming bogged down by unforeseen problems in achieving the necessary quality. Besides just producing it, they must also devise cost efficient quantity production methods. Only time will tell.
If completely successful, this coming technology breakthrough means the Fujifilm camera body certain customers want to buy, if only it were FF, will most probably be delivered in APS-C format. What will happen to general photography in the next couple of years could be quite remarkable.