The last couple of months I got three mails from three different anonymous sources sharing the same info. The Hasselbad-Sony collaboration is going beyond the simple “luxuruzation” fo Sony products…read [shoplink 22672 ebay]Hasselblad Lunar (Click here)[/shoplink] and [shoplink 22691 ebay]Hasselblad Stellar (Click here)[/shoplink].
According to these sources Hasselblad and Sony developed a brand new sensor and they will show this sensor sometimes in early 2014. I can give you one important info about it:
Every single pixel can have a different shutter time! This means the sensor allows a dramatic increase of the dynamic range. What sources didn’t tell me is how exactly this works and if the sensor is going to be first used by Hasselblads new medium format camera or by a new generation of FF sensors. Anyhow, its great news to see that Hasselblad is working on some exciting new tech with Sony!
—
Message to sources: You can send me more info anonymously by using the contact form you see on the very right of your browser. You can also contact me directly at sonyalpharumors@gmail.com.
For readers: Subscribe the RSS feed (Click here) and Twitter (Click here) stream and the Facebook fan page (Click here) in order to not miss any news!
—
Reminder (SR = Sonyrumor):
SR1=probably fake rumor
SR2=rumor from unknown sources
SR3=50% chance it is correct
SR4=rumor from known sources
SR5=almost certainly correct!
142 Responses
This may also mean that you could use a flash with any shutterspeed. Or worse: you cannot use flash at all.
Strongly support your statement. We will never know what Sony going to do.. hahahaha
If different pixels can have different shutter speed, then there’s little need for flash. It would be impossible to sync it with pixels of varying shutter speeds, anyway.
If they make such a sensor, it’ll be a revolution, because if made in small format, 1/1.7″ for example, it could have DR of much larger sensor, effectively making larger sensors obsolete for most photo needs.
“It would be impossible to sync it with pixels of varying shutter speeds, anyway.”
No, it is possible. Even with different shutter speeds, they will be opened synchronously for a duration of the shorter SS. But I hope there will be an option to fire the flash outside that duration for creative purpose.
I’ve said this before, if each pixel has the ability to record the time it takes to reach saturation, you can calculate how bright the light that created was, in a sense giving each pixel it’s own shutter, and gaining near infinite dynamic range, how does 50 stops sound?
Human eye has DR of around 20 stops, I’m not sure… So, I guess most of us would be satisfied with such sensor. For night shooting, it would be interesting to have more, but for daylight – no. It would look too flat.
1. You would probably not use this feature with flash.
2. If the exposures all started or ended together (I.e. the electronic front curtain staggered and the mechanical shutter ends the exposure) you could likely use either front or rear curtain flash sync at ‘normal’ speeds
3. This is pretty awesome sounding… The processing challenge must be intense
1. why not? suppose odd rows have exposure time 1/100 + flash, even rows have 1/1000 + flash. so you get extra bracketing capabilities. might have creative use, or be in handy sometime
Maybe, if you are blending flash and ambient light… But generally with Hasselblad, we are talking studio and the light is all strobe.
They didn’t mention it but the fact is that every pixel has its own flash too! 😉
Whaaaaaaaaaaat.
That sensor idea should upset the people who argue about ISO-less cameras.
Not really, just another step forward.
Sounds like each pixel gets its own electronic “shutter”. Madness is so!
exactly!
*if
This one will be very interesting to follow, a large sensor (APS-C or more) with such technology is probably the dream of every landscape photographer.
sure as long as there is NO WIND and a lot of light.
That would vaporise Canikon…
Sony sells sensors to both, P&S ones to Canon and different sizes to Nikon.
Canon has switched to their own sensors for P&S cameras too for a while now.
Only Nikon would be the first one to release camera with this sensor – lol (D800 story all over again?)
“That would vaporise Canikon…”
Technologically, yes. But don’t underestimate Canikon consumers loyalty.
True as a donkey with a carrot in front of the snout…
But
Every empire will fall
Every monument crumbles
Yea, only Sony is in no position to threat any of these. Not with 1000 euro for 50 f/1.8 lens. lol
Sorry, but bodies aren’t everything, and until you and Sony will understand that – Sony won’t be any threat to Canon or Nikon.
Barry would be really pleased. I guess this new sensor would be really expensive, I’m not surprise if it’ll appear in Hassy medium format camera first.
Another way of approaching it would be for each pixel to have a variable ISO.
How about just hacking the RAW file and do this in software instead? Cheaper.
No to both, this is even much better as you can set the saturation of each pixel independently while ISO works with the general saturation (which is always a compromise) of the sensor and then amplifies from there. In the former scenario you can have noise benefits (less noise in shadows) and easier to avoid clipping in highlights due to oversaturation.
>In the former scenario you can have noise benefits (less noise in shadows)
for most modern non-Canon sensors, it is irrelevant. Only Canon still uses noisy ADCs.
With most Sony sensors, underexposing 4 stops at ISO 100 + pushing shadows is very much like shooting at ISO 1600.
And you cannot escape saturation with that, because saturation occurs in the pixel.
It’s not irrelevant at all. If you expose for the highlights with a current Exmor sensor and that results in pushed shadows at the equivalent of ISO 1600, with this technology you could expose shadows longer at say the equivalent of ISO 100. That’s still going to be a big difference in shadow noise.
But yes, Canon sensors would benefit even a lot more.
One can do a lot in RAW, but not magic. This sensor could give you fantastic HDR with one shot, without tuning raw data to the limits.
Sensational if it is true, and delighted with the ‘Hassy’ tie up with Sony also! Of course this means the A7/A7r would quickly be consigned to the rank of museum artefacts, and my beloved A900 and A77 would conversely become ancient geological curiosities!!!!
New sensor technology?
Dare I hope that I might see the new sensor I am waiting for before 2015?
It’s going to be slow (low shutter speeds) and expensive, to begin with. Hence, Hasselblad or Sony MF back. Amazing feature!
Pixel peeper heaven, setting exposure for every pixel..
Will it have manual exposure mode?
that will take some time to set up!! hahaha 😉
“Ready? Just hold it there a few hours! I just hav to set the manual exposure in my 36 million shutters first!”
LOL
So, it’s basically going to be a Sony 36MP sensor used in the A7R but 3x the price. We’re onto you Hasselbad. You’re not fooling anyone.
It sounds possible that this might be a stacked sensor and that it might start out as a Hasselblad exclussive until 2015
I would of course like to see it in a sony camera i can afford sooner but am not expectin that
Maybe it’s a FF version for Sony and a MF version for Hasselblad!
I suspect sony will introduce stacked sensor in 2015.
If so that would be consistant with giving hassy 2014 exclusivity.
Sony is more a sensor giant than a camera giant.
stacked sensor?? in MF camera? Larger sensors is where new technologies hit the latest.
E.g. only recently Sony released a 1″ BSI sensor in RX100Mk II to market. While in 1/2.3″ BSI sensors were available for years. In a few years probably will hit APS-C – eliminating the colorshift plague.
I just bought the A7R, a very nice cam that will reveal every inherent properties of my large collection of legacy glass. If Sony presents a new, similar camera with a stacked sensor with individual pixel shutter technology for 3x the price, I will be more than happy to buy that one as well since the A7R was a bargain. 🙂
Is this just Hasselblad’s way of saying they are converting to CMOS?
Medium format (6×6 at least) mirrorless open camera wanted !
Medium format (6×6 at least) Flex Body wanted !
Medium format (6×6 at least) Arc Body wanted with shift and tilts capabilities with Schneider large circle lenses for Architecture and urban picture.
I want them!
Then they must store the exp time for each pixel in the metadata.
Hahahaha. Good one.
It sounds possible that this could be strictly for Hasselblad use similar to the first 36mp sensor being only for nikon use until now.
It seems logical that Hassy would require a new super sensor to compete with the advances being made in 25MM sensors.
Sony is the world leader in sensors and would be most qualified to build them one.
I have long been suggesting Hassy build a 36x36mm square with a mount
Possibly an a mount Hasselblad mounnt hybrid or a special a mount hassy adaptor.
This is likely to be a Hasselblad only product and may still only trickle down to sony cameras in 2015.
I am anxiosly awaiting to hear more.
Holy smoking crap! Almost like permanent HDR… that would be totally amazing.
True. Besides, that is closer to the mechanism of our eyes.
Sounds perfect for the Fuji X200 :-))
Admin: did they share any info on when such cameras would roll out?
Sony we need lenses for the FE Mount..so can you please stop sleeping around with every camera manufacturer (olympus.hassy)
and crank out some new lenses like elf’s at Christmas :%)
thought the focus was on “smaller, lighter cameras
Sony is the world leader in sensors.
Sensor sales are far more important to sony than lenses or cameras.
To suggest they give that up to work exclussively on a lens for you is….. …… well lets just say not going to happen!
He can ask, can’t he? You expect Sony to make a 58mp sensor just for you, so why shouldn’t he ask for lenses?
Moreover, he probably won’t ask for the same thing every. single. day.
The engineers working on the lenses aren’t the engineers working on the sensors. It’s not an either or situation.
Is that like…. DRO??
Let’s dream…
Make a sensor like that movable so it can focus with MF lenses. Put PDAF on it, make it APS-C and announce a range of Speed Booster like adapters for various FF mounts. Make m4/3 adapter that lets camera work with m4/3 lenses in crop mode. And put everything in a NEX-6 shaped/sized body…
One camera to rull them all :D.
New contrast detect seems to be as good as pdaf in focus speed and not require dedicated pixels..this is probably more likely.
I would be surprised if Hassy goes mirrorless at this time however because thier costomers would not be willing to give up the mirror fot medium format work.
PDAF because of focus tracking and compatibility with DSLR lenses, not for general focusing.
Brilliant idea, wouldn’t this be perfect for landscape photo? One exposure HDR RAW-file kind of output sounds like a great idea. Wonder how this works for moving subjects?
Sounds like such a sensor would have to take two shots: one to know how bright each pixel was (to calibrate the individual shutters), then another one to take the actual shot. So it would always be limited to half the speed of a conventional camera. The camera has to have no movement between the two shots either, or there needs to be some pretty nifty software going on to match the two images…. plus the camera shake between individual pixels would be different, etc. So potentially lots of software tech involved in the exposure: its not just the sensor that is important for this to work…
That’s all a small price to pay though, as it sounds like a really revolutionary concept!
Thinking about it, all those limitations would not be issues in a studio environment, so perhaps that’s why HB are partnering in this?
It must be a very poor studio that can’t afford good light. Rather distant landscape, I think.
BTW, this type of exposure can be done in single run: saturated pixel simply shuts itself down earlier and its exposure time gets recorded (in a digital or better analog form). During readout all pixles should report both intensity and actual exposure time. This needs double bandwidth but same shutter speed. –Just wild guess.
hmmm. Shutting off a sensor before it saturates is no better than letting it saturate, as both value would be just as wrong? You’d have to detect perfect exposure per pixel before the exposure, and then expose for real, so two shots.
Trouble is that color reproduction changes with exposure in current sensors (see http://howgreenisyourgarden.files.wordpress.com/2011/01/colour.png – this is a test where I exposed a color chart by 0, +1 and +2EV, and then exposed back by 0, -1 and -2EV in post and merged the results, the image shows by how much the colors change), and that’s why cameras don’t automatically shoot right and correct back to the left by default…. which is kind of what this sensor would be doing by default, except its doing it directly on the exposure time rather than in post
There is an alternative. Scan all the pixels regularly, when they near saturation level record the level and exposure time. The RAW processing could then take the exposure level into account.
Frankly I think it’s unlikely this we’ll see anything like this soon as the power consumption would be very high, but it might just be feasible for small sensors with relatively low pixel counts as it would allow for much higher DR. I suspect that it would be of most use in video where the resolution if higher. For most still HDR photography there are simpler techniques.
This is Hasselblad
They are not talking about small or video or low resolution.
Ah yes of course!
So they would be shutting off saturated sites early (i.e while they are still in their accurate range) then interpolating what their actual exposure should be based on a ratio that would be something like (main shutter time)/(individual pixel shutoff time).
The extra headroom you get on the high end helps in the shadows because it then allows you to set a longer shutter than normal.
No, that’s metering sensor’s job. Shutter variation will be activated based on exposure map supplied by the metering value. Or they make those pixels as a metering sensor as well.
Why worry about two exposures?
How does the current mirrorless sensor determine exposure?
Surely it’s all LIVE and only when the capture occurs does one need to be concerned about highlights and shadows – which would have been determined via the live view anyway.
This is why one can adjust the live display of the histogram before taking the shot.
I think the shadows will determine exposure time and the pixels capturing highlights will be shut off way before those ones capturing the shadows.
Wooden pixels! FInally!
Ebony(black)…..Redwood…..Blue Spruce….Green. Leafvein…
It seems possible to have each pixel charge short of clipping, and vary the difference between the first pixel shutting off, and the pixels receiving less light.
Am I right to assume that such a sensor would only be usable for non-moving scenes ?
I don’t see how it could adapt to a moving subject. Or maybe by setting faster pixels to the moving subject. But in this case, the moving subject would have a less dynamic range than the rest of the scene ?
Well, it’s too soon to guess 😉
I think GreenLens is right it will be cut off right before clipping but probably with a maximum exposure time so it will still work with flashes and movement. Let’s say your max “global” shutter speed is 1/50. If any pixels hit a max point before then they’ll shut down and record the time (1/75, 1/100, …) as well as luminance data (color is determined in the debayering). If there’s any movement that would show up in 1/50 of s second it will be confined to the only the pixels that stayed open the maximum amount of time. It will probably result in some strange new artifacts in some limited cases but it will probably more often save exposures that would be thrown own for too much motion blur. The same principle could be applied to strobist photography as well.
A very intriguing application of a global electronic shutter, although I believe Aptina does something similar in some of their sensors. I think if this all happened fast enough it could be amazing for sports photographers. Between the clean ISO of Sony’s EXMOR and the no-need-to-fear-clipping of this new tech, a photog would just have to worry about setting the shutter to freeze the motion. The camera brain could do the rest.
Single-shot HDR?
Looks nice but… What happens if the object is moving? The movement will be different for every pixel depending on how dark the pixel is. Strange pictures indeed! The result will be a 50% grey picture in worst case! Not so nice…
Ideally, amount of movement should be limited to single pixel. Otherwise the image will be blurred. Isn’t it always true?
If this turns out to be the sensor Ive been asking me for they should name it after me and give me a complimenary copy.
The worrying thing is that you might actually be serious.
I WAS YOUR IDEA!
You said iI could ask.
Where would I be without my most loyal fans like Lotus Eater.
Grow up, the world doesn’t revolve around you. I’m as much a fan of you as I am of piles.
You folow all my comments and consistently respond( always negatively) that makes you a loyal fan..it also made it really easy to bait you into making this comment.
Not at all, I barely ever respond to your comments. In fact, I’m extremely inconsistent.
But yeah, well done for baiting me *slow hand clap*
I thought you were asking for a sensor with very small pixels, to suit extreme cropping in wildlife photos.
This idea would give a sensor with very high dynamic range, which is quite different, and more needed for landscapes, architecture and streets at night. It might well have a low pixel count.
You are probably correct.
Mayby the technology will spill over into something I can use…….and afford….
If you think it through it will give you a output with zero dynamic range, all pixels saturated is that. But maybe that’s what photography will become. A feature less blank white frame, praised by all. Until some child says ‘but the king has no clothes’
for moving object scene its how fast metering react, it would be some smearing pixel on dark moving area. unless they have individual iso implement.
they will still have PSAM mode with one additional mode, i guess its DR mode. Speed Priority can’t work with this variable shutter pixel.
so what happen to organic sensor? this is just based on cmos tech with revolutionary proccessing power they have to metering individual 50 million pixel on medium format.
it might be no movie on 1st generation, they will have global shutter but sync-ing 50 million shutter without heat are the real challenge.
We dont yet know for sure if this is organic or cmos.
yes, i think this is the advantages of using organic.
with no shutter curtain, this would be mirrorless, and they have to active all the time with EVF, individual pixel metering have to respond to WYSIWYG EVF WITHOUT LAG!!
this is overwhelmed for conventional Cmos.
I’m guessing the differing shutter speeds for each pixel is ideal for still life’s and a capability of each pixel to have a different ISO at one shutter speed is what they are talking about.
Damn, don’t have the money to go medium format. Hope it’s FF, that’d be worth the wait.
…btw this is equal to suddenly solving the FF electronic shutter problem and even going beyond that, amazing!
And now the most important question: Will the pixels be big enough that you can put a cool shiny “H” Logo on each one of them???
this is revolutionary design, this is isnt cheap, sony needed H logo to go over 100G
they will come out on H first.
we have to wait 1-2 year b4 its release on affordable SONY brand.
or
we only have variable multiple zone pixel shutter instead pixel on sony FF
And what about at motion blur, if every single pixel has a different shutter time?
But really sounds like an exciting technology for a great DR!
Sure the motion blur will be distorted but a relatively steady subject might be captured well.
Motion blur would worry me too, especially with longer exposures (big risk of variable sharpness across the frame). Presumably Sony has come up with a way of managing this though… I’m curious about target market – for still life/fine art it has a lot of potential but for studio it sounds less attractive (assuming you’re using flash)
But I’m sure it will have a photo mode with equal exposure time for every pixel too, to handle more extreme motion blur situation.
i guess the effect would be lighter areas=sharp and darker ones=blurry, its like the super duper version of “night portrait mode” ( wich utilizes large aperture and rear sync flash in dark settings).
This sounds like Sony may be about to introduce a global electronic shutter.
With no mechanical shutter, flashes can sync at more or less any speed and just imagine the possible frame rates….
This is exciting.. I’m hoping we’ll finally see a digital 6×6 sensor (that Hasselblad is known for) or sensors closest to the sizes of 120 film (NOT Pentax 645D, no offense, but I don’t see it as a true 645)… That will be a good replacement for the V-series users after the retiring of the 503CW…
I’m not surprised that Sony is working on a MF sensor for Hassleblad, but for it to be cost effective I would expect the technology to be used in future Sony APS-C & FF sensors.
There is talk of Canon announcing a Medium Format digital system next summer, which they have developed with possibley Phase One. So is Canon also about to produce a MF sensor and does it use new technology?
Interesting times ahead 🙂
i’m just curious how the light metering will work??
As before, I think. You’ll simply meter for shadows, that’s it.
Can Hassel win back our faces without laughing after Lunar-Moonar nonsense?
Interesting news, I believe that it could lead to CMOS medium format sensor since Hasselblad doesn’t really have as strong offering as PhaseOne and I’ve heard that quite a lot of Hasselblad users are switching over to PhaseOne.
Shame that after discontinuing V series in April all hopes for square medium format sensor vanished.
Something does not sound right with this, like controlling it. Maybe on some kind of fully automatic camera maybe. I would think this is further off than 2014 and further out of my budget range
This type of shutter could make iso irrelevent
Each pixel could collect light until it has enough then record how long it took to collect it.
Using this data printers could print more than 4 colors and extend the d- max to accomadate the extended range. ( d-MAX. is the printing term for maximum ink density…when the d- max is too high the paper cannot hold the ink so it offsets…when the d- max is too low the shadows look washe out)
As Shamb has noted it must be strictly linear operation to succeed with color balance but apparently it is not. In current sensors at least.
IT WAS YOUR IDEA!
You said i can ask…
Where would I be wothout my most loyal fans like Lotus Eater.
This I would go into debt to purchase. Finally a camera designed as a _data recorder_ and not as a picture maker. Of course, they will still overload it with picture-making hammers and chisels, but at least it is a small step away from the fossilized idea of what a camera should do.
Extra bonus if they make the sensor hexagonal (or round) so that it records near the maximum amount of data our expensive lenses collect. As it is, framing should be done in PP, with the data your camera recorded. Now it seems we have a promise of having enough data to be masters of tonality as well.
Implicit in this, is, at long last, a core paradigm shift: photography has nothing to do with “truth”. It is no more than another set of manipulated data.
Nice post Kirby, this is what I have been saying all along.
You philosopher, you…
It’s to be called Quasar and will transport the subject to another location/dimension like Star Trek transporter. I can see a good use for such a camera. I’m naming mine “Spock”.
There are two ways this could be done.
Firstly, and not so probable, a map is measured and each pixel is given an individual integration time (exposure time). This requires two exposures, first to read out the metering results, compute it and then write it back to the Read-Out electronics for each pixel. It could be done as a SOC (system on chip) but that would be a very very expensive and complex chip. This seems a stupid and overcomplicated way of doing it. Multi frame HDR is more effective and less sensitive to motion blur.
Secondly, and more probable, is that each pixel has some kind of metering that stops the integration (exposure) when the pixel reaches an “ideal” value. The exposure time value is then added as a header to the pixel value in the digital readout.
Could work and very doable. Especially with large pixels where there is room for the additional electronics needed in the Read Out circuitry.
Don’t hold your breathe. Something I have learned: manufacturers do not bring radical innovations all at once or, they demand a formidable price. It;s going to be like this: if such a sensor comes at once with a significant jump in dynamic range and low light performance, it will probably cost about $15000 (most likely in Hassleblad brand). The following year, it will sell for $14000 and in subsequent years, it keeps gradually drops in price until it becomes almost affordable for masses. Another approach: a very basic version is introduced which improves half a stop and is priced just %10 higher. The three years later, a better implemented version is introduced which improves another half stop. Eventually after ten years of gradual evolution, we would see a significant improvement in sensor performance.
Another observation: Sony introduces a basic version with minor improvement. In the following upgrade release, Sony switches back to the typical CMOS sensor design, claiming this kind of sensor design is not suitable for photographers or perhaps, there is technical problem. A few years passes and there is a debate war going on in DPReview forums when many Sony users protest and demand Sony not to use the new sensor technology because it is against pure photography. They argue the camera has been compromised because of the new sensor and they are glad Sony decided to stop using the new sensor in their cameras.
In the meanwhile Canon has already started implementing the new technology in their cameras with significant improvements. All Canon and some Nikon DSLR’s already utilize the new sensor when Sony is finally convinced to follow the suite and starts using the new sensor in all their cameras.
Analogy: Sony hesitated before including video in their camera and they are slow in improving video features. Another analogy: Sony is hesitant in including touch screen interface in their cameras and are just waiting for Canon and Nikon to lead the trend before they try to catch on.
Nikon introduced video to DSLR, then take it away with the DF. Ironic.
Perhaps it is ironic, but certainly shows that they do listen to their customers, unlike Sony.
Yes, 5 customers asked for removal of a feature that was already developed and implemented for no other reason than… removal.
They removed it because they didn’t want mainstream users to know how bad the D4 sensor is in video. When even new Leica bodies have video, it’d be a joke for Nikon to think their users are more hardcore about still photography.
Just imagine the embarrassment Nikon would have to undergo if people hopeful about a small body FF from Nikon with not too many pixels (suggesting usefulness for video) ditch the DF for something like a Canon 70D?
You are taking the matter of who was first too literally instead of what really happened. Nikon with D90 and Canon with 5D MKII introduced video to the DSLR world within two weeks. It doesn’t matter who managed to announce its camera before the other one. Most likely, they were aware of each other’s plan and were already in a race to be the first. It wasn’t a matter of followup (like what happened with Apple’s iPhone and the world followed it). My impression is that Nikon is overall a much more conservative company and often plays catch-up. They knew Canon is coming with video and all they cared was to be the first to announce a video company. They didn’t care and still don’t care about much about its usability. It is the competition that drives them up. Canon, on the contrary from the beginning had a systematic approach to implement video in its DSLR’s and has been moving in that direction continuously but steadily in well calculated moves. D90 only shot 720P with a loose codec while Canon 5D MKII could shoot 1080P in H264 codec at 30mb/s one of the most most common and highly regarded codecs in the video industry.
Nikon is more like Sony, while Canon is more like Apple. Canon knows they wants to have video recoding function in their DSLR’s in as most effective and usable way as possible and gradually they add features one by one to improve the video in their cameras while, Nikon and Sony are following the suit only because the market demands it. On the other hand, while Nikon is only a catch-up player, one should give a credit to Sony managers and designers because they do have their own ideas and alternative designs such as TSLR and compact FF mirrorless cameras some of which is quite worthy to consider. One problem is that (unlike Apple) they most often they don’t have a clear idea of what they want to achieve so they are just trying alternative designs c0ounting on their luck that one of their innovations eventually succeeds in capturing the hearts of consumers. I actually think A7/A7r has a good chance to succeed particularly because they re priced reasonably. The only serious barrier is lack of lens options.
Sounds interesting. But I wonder how to operate the control
See the seventh paragraph of my first original post on the Nov 25 “A-mount showdown” thread:
“My entire position rests on the assumption that information is captured in the first instance, and a particular weak point IMO is dynamic range…I don’t think there could ever be sufficient DR. In any case, people are working on DR, and on filtering and localized sensitivity adjustment on existing sensor types.”
Well, now that the cat is peeking out of the bag, it should be clear enough that Sony knows where the low-hanging fruit is, as far as the future of photography is concerned. Improved DR + PP is where it’s at.
Which is awesome news for me personally.
Ultimately, one distant future of “photography”, and of “videography”, would be the re-creation of the images in actual 3D, which would likely require different initial image capture, vastly enhanced processing capabilities, and new display/presentation technologies…that will be the real “quality revolution” in photography/videography.
What has happened since the early large format guys like Ansel Adams has been the democratization of photography and, later, videography. It is still not very democratic at the highest level, especially videography, which is left to pro-level corporate interests and is still entirely undemocratic…but reasonable-quality videography and photography have come to the masses.
As far as technical improvements are concerned, the ultimate proof is in the images themselves. Has anybody yet bested the quality of Adams’ work, technically or compositionally? IMHO, no. It has certainly been equalled, but not bested, and those were done, technologically speaking, a very long time ago.
Of course, there is much to be said for the democratization of photography–but to get really good capability into the hands of the everyman, there is still a ways to go, and that’s what all this consumer-level stuff is all about.
The everyman of today is much more likely to be familiar with, and competent with, PP, than with the technical requirements of initial image capture. Camera development therefore, and rightly in my opinion, focuses on idiot-proofing the initial image capture as much as possible, just to get the information in the first place, hence huge resolution, where you can just relatively aim and select what you want later, rather than laboriously contemplating composition and framing before the exposure is ever made.
Also features like panorama sweep, auto-stitching, automatic bracketing, auto focus, and so on.
Improving DR is the critical dimension in which current sensors are weak, as regards initial image capture (yes, color depth is also wanting, but generally not so much so as is DR). PP implementations have been tried, where images of varying exposure are combined into one, but in my admittedly limited experience those PP implementations are not as satisfactory as are other applications using PP, so the best solution to increasing DR would necessarily have a physical technical component as well as a processing component.
Sony really seems to be heading in the right direction, at least as far as my uses of photography are concerned.
Your posts are proving worthy of me logging into this site day after day. Very interesting….
I’ve always wondered why you couldn’t could just sample each pixel at several times during the exposure (given fast enough circuitry). You also could possible sense how fast the “bucket” is filling and interpolate what the value would have been for a blown out pixel.
Just thinking out loud here, and this may have already been said in the long thread above, but the way I can envision the implementation of something like this is as follows. Basically each pixel exposes until right at the saturation point or a few photons before saturation. The time value that the pixel collected photons up to saturation is then recorded and the RAW decoding then changes that to a tonality based on the time to saturate value. Thus pixels that exposed for a longer time would be rendered darker than pixels that exposed for a shorter time – a completely different type of RAW format. If you applied today’s RAW format to a file like this it would just be solid white but if the time value was what was being decoded, a very clean, very high dynamic range image could be captured since we could rid ourselves of bits per pixel as the time value would have nearly infinite possibilities. It would essentially be a step-less analog signal with no need for hardware based D/A conversion. Probably a long way off for commercial applications but it’s cool thinking about the possibilities and there may be severe flaws in my thinking but that’s just a first pass thought process on this idea 🙂
One could record time to saturation or rate of photons hitting the sensor instead of counting the number of total photons that hit within a given time. This would be great for bright pixels, not sure how well it would work for dark ones. You still have limitations on the clock speeds, or how fine a division in time you can measure as well as limits based on the file size and how many bits can be devoted to each pixel. To stop motion or blur the minimum shutter speed overall has to be below a certain maximum, which may limit DR information. At fast shutter speeds you are still going to have issues with not enough photons hitting a given pixel in dark areas to get a good signal to noise ratio. It may well be the sensor just has some special modes that can be used in certain situations only, while it acts conventionally in normal shooting.
Each pixel would record th hue and time to reach critical exposure…that info would require no larger file than current methods.
This kinda like Quanta Image Sensor(QIS)?
http://ericfossum.com/Presentations/2012%20March%20QIS%20London.pdf
It would be cool to utilize a touch screen to designate different sections of the sensor to different shutter speeds through live view. For ex. Live view a landscape shot and draw a line across the horizon to assign it to a different shutter speed than the foreground. If you could integrate a photoshop-esque magic wand selector you could work wonders selecting blown out areas and assigning different exposure to them. Though, this would be the same as exp. bracketing and blending layers, wouldn’t it?
Basically the same, with some advantages such as less time difference(1 shot vs 3 shots) and it is live.
WOW! Apologies in advances if my comments are covered by previous comments above, but WOW! I assume that this means that every single pixel would now have perfect exposure and that the horrible HDR-Grey look would disappear? Would this also mean that we would need to totally re-evaluate the way we look at pictures and also how we process and post-process pictures? For example any normal scene with a normal combination of areas in shadow or deep shade, areas with bright highlights and areas with ‘normal exposure’ now would ALL be perfectly exposed. Does this mean also that the whole picture scene as a result would look FLAT (though perfectly exposed) and that pictures would lose the contrast/tonality that makes them interesting? I guess we would need the ability to tweak ‘bands of exposure range’ to make the picture recoverable and ‘interesting’ again?
I may be talking bo*****s, and feel free to constructively criticise a non-pro/enthusiasts’ comments, but this to me is one of the most interesting, game-changing and arresting photography announcements I have heard in a long, long time.
Add a switchable 100% bright OVF/EVF view to the next FF A-Mount and ditch the video and I’ll be in Photographer’s Heaven! 🙂
Yours in Photography,
Malinki_Malinki.
I think it’s too early to say, but we still have to use some HDR algorithms to construct/record an image otherwise it will just become a plain gray picture. Remember, every pixel has the ability to be perfectly exposed(18% gray), so the camera has to make a calculation based on surrounding pixels and decides which regions should have the longer/shorter exposure. Sky vs ground exposure will be a classic example of this tech.
Sony is definitely the new Kodak or Intel of the photography world. I hope we see a mirrorless medium format camera from this partnership. Electronic global shutter would do away with expensive leaf shutter lenses, provide super high min shutter speed (1/16000+, goodbye neutral density filters), and be a breakthrough in strobe photography.
This reminds me of the Fuji S5 pro super ccd. Maybe a different take on the same issue but not revolutionary in my opinion. http://www.dpreview.com/news/2003/1/21/fujisuperccdsr
A completely different take and fully different end result. It’s like comparing a boat and a plane.
The Fuji sensor was a CCD and only had two differently sized pixels for 2 different “exposures” (saturation points) per (double) pixel.
The Sony approach will allow near stepless individual exposures per pixel on a CMOS sensor. Millions of electronic shutters, would be nothing short of a revolution compared to current single mechanical shutter operated CMOS sensors (minus a few small sensor exceptions that allow or have “single” electronic shutters).
Ok. Fair enough. But you like all the rest of us here have absolutely no idea about the end result of this sony hassy rumored sensor.
Secondly the difference in the Fuji was not just the size of the photo diodes, they had different sensitivities. So in essence the lower sensitivity site had somewhat the effect of a faster shutter speed if you will.
If it was just a size difference of the sites I’d agree fully with you. But it wasn’t. It used photo diodes of differing sensitivities on the same sensor then the software combined it all to produce the image.
Naturally, sensitivity of the pixel is strongly related to size and contributes to what I said about the saturation points.
my guess is: expensive as hell. this will be medium format only for a couple if years. hence they need hasselblad to make the bodies for this sensor.
Gotta love how people take things literally and run off speculating in tangents. The story of humanity I guess.
At this stage all they could have meant was that it is the _equivalent of_ each pixel having its own shutter. If it’s that, it’s not a big stretch to implement it… no need to have millions of tiny gnomes opening and closing tiny windows and the like lol 😀
probably a very powerful image processor that has a electronic shutter for every pixel. but you still need flash because for example, you are a reporter and you want to take a photo of some VIP that rush around, you will still need fast overall shutter.
but i do agree with some of the posts here. unless sony has the professional support service, the sport/reporter circle wont even look at it. take canon for example, if you sign up to their professional service, one a big event day, they will have van /booth there to make sure you have all the back up kit you ever needed.
sony came a long way since a900, it was a true challenger for nikon.canon machines at the time, but what they did do after that? release a toned-down A850 (which is still ok), but user had to wait for ages for a true update few years down the line. until sony realise the power of brand loyalty rather than starting a completely new system or new design every time… people will continue to get annoyed and stay away. continuity is what they need.