Today Fuji announced the X-M1 (fulls pecs and price at Amazon, news and reviews on MirrorlessRumors). Fuji is becoming a serious competitor in the mirrorless world and now offers a wide range of cameras (the X-M1 is their first entry-mid level camera).
Also on the sensor front they promised quite some innovations: Two weeks ago Panasonic and Fuji (Source: Fujifilm) unveiled the new organic sensor tech that in theory allows doubles the dynamic range of the current best sensor Sony 36 megapixel sensor used by the Nikon D800 (Source: Adorama). The Sony sensor has 15.3 EV while the new Fuji-Panasonic sensor achieves a theoretical result of 29.2 EV.
But as reported prevails on SAR also Sony is working on the same tech. And Egami just spot a brand new Sony patent describing the organic sensor design. And the new Fuji-Panasonic sensor is expect to hit the market in 2015 only. I believe that by then also Sony’s organic sensor will be ready for mass production.
39 Responses
This sounds awesome. I can’t wait….
Seems like we a real revolution in dynamic range / ISO performance coming…
Admin,
I read the paper where this organic sensor is described.
I can’t truly recall the numbers, but they were not in EV but in Db.
In a discussion forum, someone called foul because the doubling of Db number does not mean doubling of sensitivity. Therefore, the material as presented in ‘double’ ‘twice as much’ may just be marketing babble.
That being said, I don’t know if anyone in SAR here could enlighten everyone and give some concrete translation between the Db to EV.
Let’s also remember that in light terms, one double the sensitivity (by including twice the amount of light gathered) by increasing ISO one stop (i.e from 400 to 800)
Sorry to not clearly contribute anything but I am still at odds with claims in a language I can’t clearly translate into real life stops.
Sensitivity and dynamic range are not the same thing. Dynamic range is, more or less, the ratio of the smallest signal above the noise floor of the sensor to the largest signal the sensor can handle without clipping.
One way to improve dynamic range is by reducing the noise floor. If that’s what this sensor does, then it would also allow use of more gain in low light situations, providing better high-ISO performance.
Another benefit of what Fuji/Panasonic are claiming here is better off-angle light acceptance of the pixels (thing WA rangefinder lens close to the sensor). Like the Exmor-R sensors, this new design puts the wiring behind the photosite instead of in front of it as it is in most large sensors.
Looks like very good news for short registration ILCs, and yet another reason that I may find myself switching to Fuji if Sony doesn’t come up with a more reassuring NEX roadmap.
I sold all my Nex gears and switched to Fuji already 🙂 Fuji lenses road map is much more assuring, and most of their lenses are incredibly great! No regret moving.
On the other hand… Sony doing research into this is great.
What about that graphite sensor developed in Singapore?
Thom Hogan already debunked this techno-babble:
http://www.gearophile.com/newsviews/no-it-aint-so.html
Good one.
As I wrote. The “theoretical” measure is of course different than the “practical” measure. We simply don’t know yet how an organic sensor will really perform in real world testing.
The problem is that 88db is mere 14.6EV and not that 29.2EV you put up there.
Even the very basic principles of photonic registration mean that for 29.2EV of DR you need an ISO value of about 1/160 at modern quantum efficiency, and then maybe only 1/3EV better if the sensor is totally ideal, i.e. an ISO 1/125.
Such low ISO means very long exposures, and may be easily achieved even today by stacking multiple exposures.
I think the real jewel with organic sensors is the angle of incident light.
One of the major problems facing all camera makers is miniaturization of capturing systems. That dynamic range is improved is just another bonus.
Also there’s another dimension, often forgotten: The fall in efficiency through the years. On which organics sensor may be more prone to.
Fuji/Panasonic’s sensor has a 88db signal to noise ratio? My God, that is sick! Todays best sensors have something arround 65db?
I hope Sony can compete with this new design without violating the Fuji/Panny patant. Otherwise its over for Sony and other sensor makers.
This is some scary news for the industry. Im very concerned about this causing a potential patant monopoly.
Wow….this is huge.
6db = 1 EV
So 88db = 14.6 EV.
That 29.2 is clear mistake.
From what I understand EV can be expressed with two different equations
EV = Db / (20*log(2)) which is common for photography
Or
EV = Db / (10*log(2))
The top one equals 14,6 EV and the bottom 29,2 EV
You need to keep in mind, the published dB is for the sensor itself, without color filters, to get the colorful output, you need cut 2/3 of the light on sensor, and interpolation it (digitally boost them 3 times) for mosaic sensors. That means, 6dB for 1EV is good to go
Seems like some could say it is market spin / babble from Sony’s competitors and I can’t wait for the next generation of Canon (rumored that next year cameras with their revised Sensors could be a mass improvement on their current generation) sensors to hit the market next year along with what many feel is the market leader Sony and their excellent Exmour Sensors.
Hopefully we see Sony improve their Alpha lens lineup (more pro and semi pro F/4 line would be good), introduce FF E-mount cameras, A-mount cameras from entry level – Semi pro and to pro (D4/1d & 1DC equiv) and E-Mount Camcorders (amateur, semi pro and pro) and by then will be grabbing the market from others in most sectors especially Video. I think Sony could be the maker who will introduce 4K across their range sooner than later compared to the some what conservative makers of Canon and Nikon – who now tend to bring their products late to the show, e.g., Mirrorless.
Its not double the dynamic range, but 1.2 times of current sensor technology “This boosts sensor sensitivity by 1.2 times compared to conventional sensors to deliver clear images even in low light.” according to the source
You have to really understand what’s behind the numbers when we are talking about the dynamic range. Dynamic range is nothing more than pixel saturation level divided by read noise. For example in Nikon D800 pixel can hold ~45000 electrons with read noise of ~3 electrons on base ISO. That makes ~14 EV Dynamic range.
Now Panasonic/Fuji claims that they have managed to increase the sensor sensitivity by 1.2 times. OK, as the current m4/3-sensors with 16 Mpixels have ~saturation level of 25000 electrons, this would transfer to 30000 electron saturation level within pixel. To achieve 29.2 EV dynamic range, read noise would have to be 0.00006 electrons. Sorry to all, that’s not possible. That kind of sensor would have absolutely no noise coming from electronics, and with current technology that is simply impossible. With read noise ~2-3 electrons, you would get ~13-14EV Dynamic range. That’s something more believable.
I feel like I’m picking on you but my understanding of what you are saying is the calculation for the db which is Db = 20 * log(Nsaturation/Nnoise).
Yes, that’s the correct calculation for what we’re trying to measure here. But as I pointed out, we have other published sensor values for smaller sensors that come close to the Fujifilm/Panasonic number. The Sony Exmor R numbers are, to my knowledge, not published, so we actually don’t know if the F/P sensor actually exceeds them.
Also note that the F/P sensor is targeted at smartphones, and much of the marketing spin in the statements are also targeted at that. If it can achieve what they say it can, it would indeed be a breakthrough for the very small sensors. But as with BSI, how much benefit you’d get in a larger sensor is likely to be lower, as those sensors already have larger fill factors.
Time doesn’t stand still and all the main players have patents for none Bayer sensors (Canon, Fuji, Sony, Nikon, etc) Digital is in its infancy and perhaps next year there will be a big leap forward in IQ especially in DR.
The described sony patent appears to be for a 2 layer sensor with green beeing an entire layer and c and y in another layer in a bayer like array.
If these pixels were to be the same size as the a77 and tbe sensor were c size it would be a 36mp sensor.
If we enlaged that to full frame of the same sensor size that would translate to 79mp for a full frame sensor or 117.5mp for a 36mm square sensor.
Based on the above stated perfomance specifications these new sensors would be one f stop faster than the a77 and have almost twice the dynamic range plus substantially lower noise.
Based on the currently available specs and a 2 layer sensor rather than a 3 layer sensor this is what sony is currently capable of if they choose to make it.
If sony duplicates the pixel size of the a99 we are talking about a 36mp full frame sensor or 54mp square sensor. Again one f stop faster than the a99 with almost double the dynamic range and less noise.
If sony makes a three layer sensor (not described in the above mentioned patent) add 25% to the pixel counts above but I have no data to speculate about the performance specifications.
Clearly sony can make a higher resolution camera of very impressive iq.
If sony ONLY makes a “30+mp” camera and does not also make a higher resolution camera that they are CLEARLY CAPABLE OF then I will be very dissapointed and sony will have missed an opportunity that I am sure thier competition will take advantage of.
One more thought.
Since the green pixels are on a seperate layer sony could interpolate the y and c pixels only thereby doubling the pixels with far higher quality than can be done later on a computer. This would be a firmware improvement and not involve hardware changes.
Not exactly. By having a layer of green, you basically have the luminance data for all positions, much like the Foveon sensor. Acuity could be very high. The issue in two layer designs has always been color spill, as you can’t get a perfect break between the layers. The color layer will likely be noisier and less precise in this design, which has implications for low light use. But I’ve long advocated this approach over three layer color if you can figure out how to do it. Such a sensor should be very good for monochromatic work.
The ‘color spill’ as you call it is the property allowing the system to distiguish hues. Without that all you get is just one fixed hue for each of the red/green/blue in the case of narrow spectrum sources, e.g. flowers.
Color spill can be totally eliminated by simply subtracting the correct percentage of the y and c signal from the green signal.
With the new generation of super high resolution lenses coming out sony needs to produce at least one super high resolution camera to match. Otherwise there is no reason to buy the new lenses. Better to wait until sony does decide to make a super high res camera or until thier competition does.
Why waste the mony buying these lenses now if sony does not plan to make a camera to put them on.
As for me, Pointy finger (bend at top)+Shutter Button (Top of Camera) = Taking pitchers. New sensor would help but doesn’t really matter to me.
I’ll believe it when I see it!
Fuji are masters of spin. Just look at their overstated ISO and the X-trans APS-C sensor aren’t performing significantly better than the Sony 16Mp sensor, even though they once were claiming near FF performance when it comes to low light.
They have made and making really great products which they should let speak for them selves instead of building this BS internet hype.
Not true:
https://www.youtube.com/watch?v=jmwEQB0KM8U
I don’t think thats a fair comparison at all. Now if they both had the same lens say the new cz then that would be fair. I had that kit lens its not all that great.
I’ll take a controlled test over that any day.
So much marketing for a promised future tech that seems to never make it to the marketplace. I’d rather hear about advances in the upcoming NEX 7.
Almost every time Sony announces that the new camera will have THE sensor. Fantastic sensitivity. Of 52000 ISO. No, 104000 ISO. Oh, eah! That would be fantastic!!!
And it was eventually true fantastic:
In reality, after the camera test, the sensitivity increase is not so impressive.
DXOMark shows this:
If take year 2006 camera A100 – sensor sensitivity of 427 ISO.
If take year 2008 camera A300 – sensor sensitivity of 538 ISO (+1/4 stop in 2 years)
If take year 2010 camera A560 – sensor sensitivity of 817 ISO (+1/2 stop in 2 years)
If take year 2012 camera A57 – sensor sensitivity of 785 ISO (-1/10 stop in 2 years)
(ok, decrease is because of translucent tech. so will take nex then)
If take year 2012 nex 5R – sensor sensitivity of 910 ISO (+1/5 stop in 2 years)
overall in 6 years we see only about 1 !!! stop of progress. And recently it is only 1/5 of stop increase in 2 years. So almost grow ISO at all.
But we again happy to read the prediction of + 10 stops increase in 2014!!!
That’s increadible.
I mean exactly this: IN CREDIBLE
Only one stop for midtones and highlights, several stops in the shadows. See the full SNR curves in the full reviews, not just the 18% grey ones.
And all is relative, a stop in 6 years vs nothing in 9 years (Canon 20D vs 700D).
Patent, patent and more sensor patents, still waiting the last year curved sensor tech patent, where it goes???
Here is an example
OVTI just announce a new tiny conventional sensor for cellphones,
8M 1/3.2″, with DR 70.23 dB. Figure that
www . ovt . com/download_document.php?type=sensor&sensorid=134
29 stops of DR is not a double of 15.3. It’s about 16000 times expansion of DR.