Our friend David from Photoclubalpha got a confirmation from Sony that the Sony A7s has a complete silent shutter mode! Early beta testers of the A7s got a firmware upgrade that adds the feature. The firmware fixes and upgrades are:
– α7S now offers a ‘Silent Shooting’ mode
– ISO range for Movie Shooting extended to ISO100 – 102400, expandable to ISO100 – 409600
– Dynamic Range extended to 15.3 stops as sensor RAW output
David writes:
This is not the same as the Electronic First Curtain shutter found on the A7, A99, A77, A6000 and so on but conspicuously absent on the A7R. Nor is it the same as the near-silent leaf shutter terminated exposure mode of the RX100 models, RX10, or RX1 models. It’s completely free of all mechanical action and totally silent.
Sounds terrific!
A7s and RX100M3 Preorders:
Sony A7s preorder in USA/CA at Amazon, Adorama, BHphoto, Sony USA, Sony Canada.
Sony A7s EU pages: Sony DE, UK, ES, FR, IT, NL, BE, AT, CH, SE, NO, FI, PL. And Wex UK.
RX100M3 preorders in USA at Amazon, Adorama, BHphoto, Sony USA.
RX100M3 preorders in Eu at Sony DE, UK, ES, FR, IT, NL, BE, AT, CH, SE, NO, FI, PT. And Wex UK,
RX100M3 in Asia at Amazon Japan and Sony Australia.
———
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!
0 Responses
The A7S gets a firmware upgrade and it’s literally just been released and we’ve been shouting from the rooftops for RX1 and A7/R upgrades for yonks and we get nothing! To hell with Sony!! If only their stuff wasn’t so good… damn you Sony.
1st blood! ^^
this news is so terrific!
I think i beat you too it mate 🙂
you r 2 minutes late lol…
I had a comment 2m before yours but it got moderated out for some reason 🙁
You’re ahead of him in my browser 🙂
It seems it just needed approval but now i’m top! :p
The perfect timelapse device.
What is that, SR1 or official news?
How can they suddenly increase raw dynamic range?
I guess it is not a suddenly or magic, Sony realized why not push it to max for the user 🙂 so they informed sell store earlier and now deliver the stuff ‘fw’.
15.3 stops Dynamic Range?Is this trueeeeee?
15.3 Stops!!!!!????? that is insane for video, that touch the F65 and Alexa arena; I own the FS700 and with S-Log2 is around 14 Stops, and that is a lot of stops in video
It is in raw. Won’t be in video.
why not? the camera have a full 4K HDMI out 422 with Full Pixel Read-Out with S-Log2 Gamma. Maybe for internal recording won’t be present, but for external recording with the Atomos Shogun is posible
if we talk about benefit of stops normally is for editing, video is only 8bit, raw is 14bit, so the “recovery” in video is not as much as photo..But it will still giving pretty good result with slog2 and adjusting correct exposure in camera
13-13.5 stops max with s-log2.
If it was 15+ Sony would have made s-log4 for this camera…
The usefulness of the A7s’ silent shutter mode will depend on how fast the sensor can be read. If it’s like on Panasonic cameras, I doubt it’s going to be considered as something groundbreaking by any means. A very nice feature to have for specific applications, but not something that’s going to be useful for all applications where a silent shutter is a benefit.
Also, I’m eager to see that 15.3 stops DR. I especially wonder how it’s going to interact with 14bits files.
The GX7 is groundbreaking, perhaps not for you, but it was for me. Silent shutter completely changed my photography and do things at events I could never dream.
The GX7 looks very cool! I’m surprised that A6000 doesn’t have the GX7 silent move.
It must be a lower resolution image with silent shutter, basically a video capture.
If it could do a full resolution electronic shutter, then why even include a mechanical shutter?
It’s a full resolution electronic shutter, but this is not a global shutter sensor do it won’t work so well for every situation. If you understand that this is basically a video capture, you should be able to understand why it needs a mechanical shutter.
I’d be very confident that it will be full resolution. The compromise with electronic shutters on (almost) all CMOS sensors is that it takes longer to read the entire sensor than the traverse time of a focal plane shutter. What this means is that for anything much fast than about 1/40th of a second, the “active” part of the sensor will be a strip which gradually moves up the sensor. What this means is that with moving subjects you get things like leaning verticals.
Of course, you get this with physical focal plane shutters, but that doesn’t normally start until about 1/200th of a second (not, coincidentally about the region of maximum flash synch).
The reason that Sony can provide this mode for the A7S is that it’s “only” 12MP and the sensor is designed for fast read-out. It probably takes 2-3 x longer to read an entire A7 or A7R sensor, and then the issues of moving subjects become too intrusive.
Dear Sony,
keep doing things like this for future cameras AS WELL AS past cameras (at least the current gen cameras).
An increasingly satisfied customer.
Why not improve a7r in that regard?
I echo the same question, why not on the A7R?
Si
Maybe the technology isn’t ready back then or not compatible with A7r sensor?
The reason you won’t see this on the A7R is probably that it takes perhaps 3 x longer to read the entire sensor. That means the “travelling” strip of activated photosites would probably take perhaps 1/10th second to move up the full height of the sensor. That would mean you would get lots of issues with moving subjects with “leaning verticals” and the like. The A7S can probably do the same operation in about 1/40th second, whilst a decent physical focal plane shutter will not get this issue until about 1/200th second (which is fast enough, that leaning verticals are only perceptible on very fast moving subjects with physical focal plane shutters).
This issue is essentially the same problem you get with “rolling shutter” artifacts on CMOS video cameras (apart from those few with global shutters).
The A7R does not have a fast enough sensor readout. The A7 doesn’t either but it is at least fast enough to enable EFCS which is a big help.
haha a7r a 800E sensor lol… too old for improvement. cant even put in the 120fps
If the A7S has 15.3 stops DR, how many does the A7R have?
DxO has measured the a7r, you can get the dynamic range measurements for free on their site.
It seems not that long ago that Black Magic was touting 13 stops of DR. Now we’re at 15? Amazing!
approximately 14
A7 14.2
A7R 14.1
Toy camera.
Where is the APS-C sensor size RX10? 🙂
15.3 Stop is important for video, not for photo. (When it’s Video, usable dynamic range is shrinking.) – I guess, there might be a new firmware for S-LOG3.
A7R is already great for its photo purposes.
P.S: BTW, Sony doesn’t give a damn for the past system, that’s why the price of the cameras are cheaper than Red, Canon cinema cameras without cine alta series like F5/F55, F65.
Speak for yourself!
I make extensive use of the DR in my photo’s. The more the merrier!
I agree, sony’s way of upgrade photography solutions are great. It’s innovative. Honestly, If I were a photographer, I would use Sony’s camera. Price and quality of the product is very nicely arranged.
Oh and I love the bash on the way out.
You Sony trolls are priceless =:Þ
HEY I OWN FS700 AND F3
Look, If I had RED, they will allow me to upgrade camera with reasonable prices, but sony doesn’t even fix some problem like highlight problems including lying about 4k for fs700, which the option that sony gave us was $10,000 stupid priced 4k raw option.
Now, I am moving on red epic, because of these upgrade system for AFTER SERVICE!
I hardly think that the video will have 15 stops of DR, depending on the encoding process, video loses quite a lot of it, it’s really, really hard to find a stills camera that can maintain the DR in video. Maybe in 4K, if Sony does right, they will be able to retain a good part of it, but the internal 1080p video may not even output 12 stops of DR.
I also doubt we’ll see the full 15 stops in video—at least in this camera. You’ll need 10-bits or more so we’ll probably see a “pro” version in an F or VG body with S-LOG3 (or maybe 4?!) next year. Still if the video rivals the IQ and dynamic range of the F5 (which Den says it does) then for the size, weight, and price it is a game changer.
Well, the silent mode is basically what Panasonic has been offering for quite a while, isn’t it? But it’s interesting that Sony can offer it via firmware because it makes me question why not for others models.
The feature was most likely planned for the camera but the software end was not perfected when the demo-cameras shipped. They have now worked out the bugs and are supplying the upgrade. If the A7r shutter would have supported electronic first curtain they would have used it from the beginning. It is possible future sensors will be fast enough for silent shutter, but only the 12 mp is for now.
A7s geared towards video. Curious to see side by side stills against A7r, both in low and high ISO.
Agreed and I couldn’t figure out how the A7S would fit into a more stills centric workflow.
You now have a wonderful lowlight camera that’ll shoot silently for social events.
Plus it supposedly (according to people like Den Lennie) shoots near perfect straight 1080HD, unlike the GH4 which requires a 4K workaround.
Sony with this camera has achieved very high iso primarily by making big pixels..only 12mp. In addition very high iso settings are still noisy. This is the type of trade off necessary with todays sensors.
I just did some reading about graphene sensors.
A sensor has been desighned and patented using only graphene. It is 1000x as sensitve to light as current sensors..thats 10 f stps.
Using these highly sensitve sensors will not require noisier images to achieve higher iso, and higher resolution sensors will still have 10 f stops more sensetivity than current sensors.
The good stuff doesnt stop there….grahene is on atom thick an can be used to make stacked senors supperthin…..single atom graphene layer..followed by an applied single atom insulating layer.
Theres more..graphine is superflexible and will not suffer from fatigue no matter how oftem it is bent, and it is transparent and supercunductive.
Variable curved sensors will be possible with this stuff..
This is also one of the strongest materiels known..100x as strong as steel.
Now here is the good part…grapene sensor manufacture is fully compatable with cmos sensor production.
The same manucturing techniques and equipment can be used to make current cmos or graphene or any combination/hybrid you can imagine…for example use graphine base layers for current stacked sensors to make them much thinner.
This is new stuff and will be the next generation of sensor tech..its here but needs to be developed.
Sony has still not released stached sensors ten years after sigma already has.
We can already see this unrealesed generation of sensor beboming obsolete to graphene even before they have been released.
Admin is like the Robert Snowden of the camera world and jas recently revealed to us that sony is deliberately withholding the new sensors because the believe they can still sell us obsolete bayer sensor.
Thy are correct until someone else releses the new tech..or we stop buying bayer.
Those who buy bayer now this close to stacked sensor release could very well regret doing so when the stacked sensors come out.
I wonder if the curved FF sensor patent have anything to do with this tec?
I do not think any of the current or ready for release technologies will include graphene technolgy.
This is still brand new stuff and is likely to be ten years before it reaches our cameras..but of course thats just a guess.
The main factor to our advantage that could speed things up is the compatability with current production..that means that although there is a lot of r&d left to to..getting it into production afterwards should be fairly quick.
Just a heads up, the Sigma stacked sensors were a big fat commercial failure.
The Grapene sensor sounds good though. Hope we get to see it in action soon.
Thats because sigma is not a viable camera system.
A mediocre camera that only takes sigma mount sigma lenses.
No hope at all that sigma will be able to do much development, and a very primitive version of the technolgy.
It never could have been succesfull.
Let’s hope it doesn’t end up in smartphones first. As if we need more reasons for people to choose against purchasing a dedicated ILC. I think graphene batteries are also very promising. The one thing Sony cameras could benefit from- longer battery life.
The potential for extremely high resolution and high iso without noise exists. Also new technologies can reduce diffraction.
Smaller cameras will definitelly meet and exceed the quality of current large cameras.
Unless large cameras also advance they will not compete with cell phones.
Any big change is going to come in only one of two places: military/spook applications, or smartphones. Because that’s where the money is.
Also, I found out that what makes graphene so sensitive to light is it’s defects- which I find odd. As long as it shows defects around the edges it’s able to maintain it’s light sensitive capabilities. I wonder what that means for IQ?
Graphene is a flat crystaline carbon structure.
Carbon normally has 4 bonds however in this struture there are only 3 bonds.
That clearly inparts some unusual properties to the materiel since electrons are shared in a non conventional type of bond.
As in any regular crystseline struture those atoms at the edges cannot fit the regular crystal structure.
I would expect various impurities including pigments can be bonded into the carbon.
I am not sure if this is possible but possibly impurities can be bonded within the crystal structure via the unused 4th carbon bond..however I am not sure if this would disrupt the crystsl structure or funtion similar to impurities in silicon crystal transistors.
The greatest potential seems to be for ultra thin materiels.
Graphene is a one atom thick supercunductor.
By arternating with one atom thick isulators extremely thin layers can not only make superthin stacked sensors but also supercapasitors.
Ever heard of photon shot noise? You can increase ‘sensitivity’ gigallion times, but you won’t get rid of it.
Nased on the uncertainty principle of quanton physics individual partices such as photons are random bu as the numbers of particles increase the randomness decreases.
If we counted individual photons that would be a problem but we are currently not there.
Remember we include 4 dimensions in our photon count.
3 dimensions of space and one of time.
Thats a lot of photons.
not that much of them… there’s a concept of “unity gain”, at which one photon corresponds to one unit of analog-to-digital converter. For most DSLRs it’s about 800-1600 ISO. so if you’re making a 25600 ISO shot, this means only few hundreds of photons per pixel. Guess why it’s so noisy?
My quess is with current but not developed technology the practical limit might be..
200mp x 3 layers…600mp…clean iso 6400-24000.
Please clarify what you think the significance as we approach those limits will be..I am also assuming we can resolve the diffusion issues.
A better approach is the one that Eric Fossum and his team are currently exploring (as background: Fossum’s name is on many of the early CMOS patents). Basically you capture data ABOUT the photon. Where it arrived, when it arrived, and from which direction it came. From that you can build a much more interesting sensor, though you need a lot of CPU horsepower to deal with the data set. You don’t have to worry about saturation in bright light, either, because all you’re doing is recording data about EVERY photon you received.
I was not aware that this tech exists..sounds like my earlier comment about a photon detector.
Tell us more about this technology.
http://ericfossum.com/Presentations/2013%20OSA%20Quanta%20Image%20Sensor%20%28QIS%29%20Early%20Research%20Progress%20talk.pdf
Thom
I read the link but only unterstand small pieces of what is being said.
I enjoy a conversation that pushes my limits but I could use some help understanding what this is saying and how it relates to new senor technology.
You mentioned in your earlier post about detecting a photon as well as its direction and wavelength.
That brings to mind 2 questions..
Is that the technology this report is talking about?
And if it detects direction will it eliminate the need for lenses?
No, doesn’t eliminate the need for lenses. Yes, detects a photon and when it arrived.
Does it measure wavelength?
Not that I know of in its current form. But it’s still very much developmental.
A conventional bw sensor converts photon energy ro electrical energy and measures the voltage.
This actually counts the photons and creates a dogital bw bittmapped image from the data.
In its current form there would be no color data.
That’s the way it almost always starts with new technologies: first you make sure you have the ability to convert light to data, then you figure out how to make that color. Putting filter layers on top of something is a different problem than making the conversion. In sensor terms, it’s usually called the “sensel.” Sensels are just the primary engine of the conversion and storage. Once you have working sensels you start working on the other things that make it into a useful product.
In Sony large sensors, for instance, the current generation consists of the same sensel technology, only scaled to purpose. The same ADC structure is added. The big difference is in how data is pulled from the sensor. With the 36mp sensor in live view modes, you have to subsample with a rolling shutter, for example. With the 12mp sensor you don’t have to subsample, I believe, though it still apparently has a bit of a rolling shutter.
This is all the reason why we believe that 54mp full frame sensors are coming, by the way: it’s really just the medium sensel size currently used in the 24mp crop sensors applied to the full frame. The problem with the 54mp sensor is the same as the 36mp one: the camera’s digital circuitry needs a lot of bandwidth to deal with the fire hose of data it’s going to get. Moreover, live view becomes a bit of a problem, as does video (rolling shutter again). But these are iterative engineering problems, and will be solved.
Thank you for all this great info about sensor tech. I really enjoy talking about this kind of stuff.
So currently it seems this tech bears watching but is very very far from practical use so far.
Graphene is good stuff..probably most usefull in conjuntion with current technology as opposed to a replacement for current technology.
54mp stacked sensors are on the way.
Personally I still think I will wait for this 54mp ff to purchase my next camera and hope we see it sooner rather than later.
Unfortunately economics gets in the way. Yes, 54mp sensors are coming, mainly because you can just repurpose the current IP on a larger piece of the wafer and only have to engineer a few new things.
Yes, stacked sensors are coming. On smartphones. Unfortunately, this is an economics problem as well as a physical one. The problem is simple: you have to recover costs these days (across all large sensor technologies) across maybe 16m units. So until you’ve recovered costs, you’re not going to do it. Where can you recover costs? Smartphones and security and auto sensors, which are selling two+ orders of magnitude already and still rapidly growing.
Maybe BSI will make it to APS, maybe it won’t. The problem is that with the current size of the APS photosites, BSI doesn’t really give you enough bang to pay for the bucks it would take to do that.
As I wrote, military, spook agencies, and smartphones are driving sensors. That’s where any new technology will be deployed, and it won’t be deployed to large sensor photographic cameras quickly because the market is too small.
Sometimes–and I think three layer technologies are one of them–you HAVE to start with bigger sensors because you can’t iron out the low level noise quick enough to make the right impact in small sensor devices. But graphene, jots, and stacks are not in that category.
Part of this comment is a little ambiguous.
You first predict 54mp sensors and that stacked sensors will occur first on small sensors like smartphones.
Your last statement however predicts stacked sensors will begin with bigger sensors.
I think the Sony is planning to introduce a 18mp x 3 layer 54mp organic/silicon hybrid ff stacked sensor in the near future..possibly 2015.
I also believe sony will not introduce a higher res exmor R sensor than any that are currently available….24mp c and 36mp ff.
I also think they will release a ff BSI cured sensor with 35mm fixed lens..4 element apo desighn.
Stacked sensors are not the same as BSI, so no I don’t think I’m ambiguous. In terms of economics, traditional sensors are the cheapest to create, BSI next, stacked the most expensive.
People are reading far too much into all these sensor technologies for large sensors. Large sensors are already quite efficient pretty much any way you slice it, and there are potentially less expensive ways to get to the same level of modest gains. For example, use a smaller process fab. Canon can’t do that because it would cost billions for a new fab. Nikon can do it because they just use other people’s fabs ;~). Sony has been doing it because as they move the small sensors to even smaller process fabs, they repurpose the ones that they’re moving down from.
As for three layer, I’ve not seen ANY hint of technology that would allow us to go to that and get current state of the art noise performance except for Nikon’s patent, which uses mirrors. At present, I don’t know of anyone that can make mirrors that small and embedded in the silicon. But given that Nikon (and Canon and ASML) make the equipment that would be needed to be converted to do that, you’d have to think that they’d have more of a chance at that than Sony. Unless Nikon decides to cooperate again with Sony. So I think you’re being overly optimistic about 2015.
Sony has two different groups with conflicting needs. The sensor group needs to sell sensors. Increasingly they’re doing less of that to camera companies and more of that to phone, auto, and security companies. The camera group is already creating cameras with better sensors than Canon; they need to concentrate on getting the cameras and lenses more competitive.
They might introduce a curved sensor, but that’s a lot of engineering work and cost for what would be a low volume product. The problem with curved sensors is that the lens really needs to be designed to the sensor, and not all types of lenses would need the same curve. So it works for–wait for it–again smartphones with fixed lenses, but for interchangeable lens cameras, which is where the main action is for large sensor cameras, it just isn’t economically justified.
If as you say we are approaching a peak sensor efficeincy than in a way that insures the survival of large sensors as they have a h ok gher guality potential than small ones…at least until photon detectors are perfected.
Yes, and remember that economics play much a part of end game than most people tend to think. Quick, which would you rather have: a cheaper full frame camera or one with another stop of DR? By using older tech on the larger sensors you get cost savings, which you could turn around to the user. Or you could spend more money in a declining market to get a sensor that does a stop better.
The reality is that we’ll get SOME of both those things, I think. At least until a new disruptive technology comes in that would allow for a hugely visible economic or quality improvement. You’re betting that graphene, jot, stack, et.al. ARE that disruptive technology, but the problem is that they just don’t add the benefit to a full frame sensor they do to a tiny cell phone sensor. So the bucks will be spent there, not with big cameras.
FWIW, we’ve always gotten the left overs in the big sensors. I can track all the way back 25 years where the smaller sensors got the new tech first and it only spilled up to larger sensors later, mostly when it had been paid for elsewhere. Indeed, Nikon had to persuade Sony to make the original D1 sensor. Sony didn’t want to do it. To much cost involved when they were still trying to iron out costs in smaller sensors. It was only Nikon’s all-in approach, coupled with some tech sharing, that got the volume high enough that it interested Sony. At the time, Sony was the leading seller of digital cameras, something that isn’t true today (they’re currently number three). And all those cameras used smaller sensors.
Sony has another electronic division that is more desperately in need of sensor magic, by the way: the smartphone group they bought in the Ericsson deal. If I were running Sony and had sensor magic coming, THAT’S where it would be deployed, not cameras. Sony’s nearing another “last chance” in the smartphone realm (previous last chance that was jettisoned was personal computers).
Follow the money.
Thom
I just had an interesting thought about the photon detector.
Assuming it can detect shorter tham light wavelengst it can both photograph and analize hidden odjects with very little modification.
For our example lets put an unkown object in a sealed wooden box.
We build a focusing electrmagnetic source.
We place an identicle filter wheel in front of both the emitter amd our photon detector and focus the emmitter to illuminate the object inside but not the front of the box.
By taking multiple exposures with mulitiple filters…which can be pulsed to reduce radistion exposure..
We now have a multiple non visible color picture of whats inside the box.
By coparing how various materiels look with the same wavelengths we can determine not only what the object looks like but also its composition.
This may have limited photographic use at this time but great potential in medical, sky, and analysis fields.
If this photon detector can be mad to determin 3 parameters..
Detect each photon
Determine itd wavelength
Detrmine its location…
This could replace sensors.
Resolution would be unlimited because the processor knows the exact location of each photon it can generate as few or as many pixels as desired.
Wavelenght info would go way beyond color information. Spectral analysis for example can analize the chemichal composition of every item in the pictire from a distance.
Non visible wavelengths can reveal non visible images such as internal structures and subjects behind walls.
Photon detection would be the endgame for image detection as there is no currently known image beyond the photon..unless we discover something new?
I’ve read that 1000x claim for graphene sensors, but given that existing camera CMOS sensors already get quantum efficiencies (QE) of over 50%, it’s simply physically impossible to improve sensitivity of 1000 fold. If you do the physics, you already find that current CMOS sensors can detect detail when the number of photons received at a photosite are counted in the several tens. Improving sensitivity 1000x would involve sensing fractions of a photon, which is clearly nonsense.
There are even some specialist CMOS sensors with QE > 80%
http://www.laserfocusworld.com/articles/2012/03/e2v-cmos-imaging-sensors.html
Read what’s been released (which is weak on detail), then you find that what the graphene sensor does offer is a much wider range of wavelength to which it’s sensitive. That’s fine if you want to image down into the far infra-red, but not much use for visible light. I think that’s where the 1000x sensitivity comes from – it’s more about the sensitivity to photons of a much wider range of wavelengths (photon energy decreases as wavelength increases, which is why (because of the way that silicon semiconductors work) their sensitivity runs out into the deeper infra-red, as there’s simply not enough energy in the photon to kick an electron into a new energy level.
Of course, the QE efficiency is measured after colour filtration, so approaching 2/3rds of the photons are filtered out by the Bayer (or other pattern) filter. Only if somebody could design a sensor (perhaps using prismatic separation) which detected colour without filtration could this be improved.
There is also still some gain to be made by reducing read noise (which dominates when you get down to the tens of photons), but short of truly heroic stuff – like supercooling sensors – some will remain due to thermionic noise. Note, reducing read noise does not increase sensitivity as such, but just means that lower light levels can be sensed before being swamped by noise.
Even if sensors could be designed to hit 100% QE and colour could be sensed without filtration, it’s just physically impossible to be 1000x more sensitive than current CMOS sensors. Maybe 10x might be achievable, but by then you are into single digit photon counts and then you reach inescapable limits due to the quantum mechanical nature of light.
I fear Barry is not up to understand QE xD
>Only if somebody could design a sensor (perhaps using prismatic separation)
Sony can do that in place of SLT mirror 😉
Surely Barry can appreciate if >50% of photons are already being detected, you can’t improve that by a factor of 1000.
Incidentally, there are cameras (mostly for video) which have used multiple sensors and prismatic separation, but it’s not really an attractive approach for still photography as it’s very expensive and bulky. What I have in mind is the separation occuring on the sensor, but I suspect that the dimensions are far to small to allow this to work at photosite levels. (The Foveon works in a completely different way and is rather wasteful with poor QE).
I seem to recall there was an ancient minolta digital camera (long before the Dynax) which had multiple sensors and prismatic separation, but the images were very low resolution. However, I might just have imagined that.
He doesn’t seem to know what is photon.
>as it’s very expensive and bulky.
things like A77 or A99 are bulky anyway
> but the images were very low resolution.
because these cameras used small sensors with focal reducers through aerial image (the design goal was to preserve OVF)
No, much, much more bulky. Imagine the space taken up by three FF sensors and all the optics required. Professional 3 CCD video cameras with even moderately large sensors are simply huge.
There are much smaller video cameras with 3 x CCD, but the sensors are tiny and the still image quality will be dreadful.
Which optics? Only beamsplitters.
Professional 3CCD videocameras are huge because they have fixed bright zoom lens (and as a bonus, allow user’s customers recognize instantly that he’s a professional)
3 CMOS APS-C camera body wouldn’t be much larger than A77.
They can do even 2-CMOS camera beaming green to one sensor and red&blue to another (with CFA), this would take exactly as much space as regular SLT camera and improving SNR almost two times.
Mayby sometime in the distant future we will replace sensors with photon detector/counters that count each photon and record its wavelength.
But that technology does not yet exist
nb. on this graphene sensor hype, I’ve just done a bit more research. It’s been misreported. Read this article.
http://www.engadget.com/2013/05/31/graphene-camera-sensor/
Then read the comments, especially the stuff by John Stern, but there are other bits which show it’s been wholly misrepresented.
Quite simply its nonsense.The battle is not with the technology of sensors, but the nature of light itself.
The article you refer to supports my comment..in fact I read that article before making the comment.
Barry, you miss the point. The article is just wrong. Incorrect. Misleading. Just out and out garbage. It’s the work of a journalist who doesn’t know what he’s talking about. That’s why I said read the comment section.
In fact the mystery is solved if you follow some of the links in the comments below the article. The 1000 x improvement was apparently the improvement over previous graphene optical sensors, not existing silicon based ones. It was a journalistic slip.
Your talk about graphene as brining a revolution to camera sensors is simply the repetition of an faulty article.
It’s quite simply impossible to achieve a 1000x (or anything like it) increase in sensitivity, simply because current sensors are already quite close to the theoretical limits. I would suggest reading up on the quantum mechanical nature of light to see why it simply can’t be done.
Assuming you are correct and we cannot achieve substantial sensitivity increases it will still revolutionize sensor tech.
The fact that this is a single atom thick superconducive film and can then be coated with a single atom thick insulating materiel is revolutionary.
Since the next generation sensors will be stacked the distance between the layers is problematic.
Using this in conjunction with current sensor components can produce a virtually 2 dimensional stacked sensor…if this is backlit like the exmor r than based on the current exmor r footprint we can have a 600mp sensor…200x 3 layers…with good iso because of the shorter microlens focal length.
The real benefit is that this tech is fully compatable with current tech and does not need to be used exclusively.
Several people comment here that the improved sensitivity claims are huge exagerations…I believe now thayt they are correct however I still see huge potential for graphene.
Yes, and one of the guys who has been commenting works has spent 20 years working in sensor technologies (and he works for Aptina, who designed the sensor for the Nikon 1).
He points out (in other places) that there is only the potential for 3 x improvement (about 1.5 stops) in the efficiency of sensors, and that’s if you achieve 100% conversion including eliminating the filter. It doesn’t matter how it’s done – you are just down to the raw statistical nature of random photon arrival and no technology could ever improve that.
However, I’d give one proviso to that. At very low light levels, noise is dominated by read noise. It’s that which will swamp the signal at the very lowest light levels, and ultimately sets the noise floor and hence the lower limit in establishing the total DR. Reduce read noise, and you therefore increase DR, so whilst sensitivity as such is not improved, you get more useful signal as it’s not swamped by noise. Such noise can’t be eliminated (at least without cryogenic cooling), but it might be improved a bit.
In theory, therefore, there might be a bit more than 1.5 stops of extra useful sensitivity (by eliminating filtering and hitting 100% QE) by reducing read noise, but I wouldn’t count on it.
Of course, you can postulate a device which counts individual photons (along with their wavelength), which would get you to the theoretical quantum shot noise limits (perhaps 4 stops or so better than where we are), but it’s very difficult to imagine how such a device would be produced.
Personally, I think the future is going to be dominated by more refined versions of the technology we’ve already got as it’s cost effective and, for most purposes, is good enough.
So would you therefor agree then that the room for the most improvement is usable iso..clean iso…particularly in low light…
And in use as base layers to produce ultra thin back illuminated stacked sensors
And that we are not likely to be using pure graphene sensors but rather hybrids of current technology and graphene?
Or do you think some of the above assumptions are incorrect?
Also in terms of resolution how much resolution…eg ff sensor…do you think is ultimately possible and why?
Given what we know about the state of graphene in sensors, I think it’s premature to
As far as resolution is concerned, then the limiting factor will be diffraction in the lens, which is purely a function of aperture. Theoretically, an f1.2 lens on a 35mm FF sensor could allow for more than 2 gigapixels. However, practical lenses can’t resolve that much detail as it would equate to about 700lpmm. If you aim for 200lpmm, then that requires 2.5 micron photosites, which comes in at 140MP, which will essentially out-resolves pretty well every current commercial lens. Of course, AA filters would then be wholly unnecessary.
However, there are a few provisos. Firstly, anything slower than F4 will mean the image will be diffraction limited (although F4 tends to be a sweet spot for lens resolution). Secondly, unless the read noise can be reduced (as it tends to be fixed per photosite and thus accumulates during downsizing) then there will be a bit of a penalty to be paid on “high ISO” performance, even at reduced output sizes. Thirdly, there are issues with the amount of data to be read off the sensor which impacts in all sorts of areas – the frame rate, practicality of electronic first curtain (or full electronic shutters).
In practice, we might see 24 x 36mm sensor cameras reaching 80MP or so, but I’d be a little surprised to see them top 100MP as that’s well into the area of diminishing returns.
The article you refer to refers to a ruby cmos sensor for use in night vision and does not appear to be applicable to colr sensors because of the narrow wavelenghth it is sensitive to.
“It is 1000x as sensitve to light as current sensors..thats 10 f stps.”
If that means that in daylight you have to shoot ISO 102400 at 1/10,000 sec and f32 I don’t want it 😀
The value is for smaller higher res pixels.
Besides sensitivity can easily be reduced and even lower noise levels can be achieved.
Doesn’t work that way, and virtually every publication that picked this up didn’t notice that the 1000x increase of sensitivity was not in comparison to current silicon sensors but to previous graphene ones, which have been terrible. We went through a similar set of misleading statements back in the 90’s when people were publicizing CMOS results and people kept thinking the comparison was to CCD.
Sensors are currently about 50% efficient. To over simplify: two photons fall, one is collected.
What intrigues me more about the graphene approach is the potential for lower dark current noise and a higher signal value to provide the ADC, which might make it possible to use higher gains for low light work.
I’ll say it again: short of having some way to deal with electron well saturation, it’s going to be difficult to get meaningful increases in SNR over current technologies. Depending upon who you ask, they’ll say the theoretical maximum is anywhere from 1 to 2.5 stops improvement using current architectures. And those won’t be easy.
If you are correct and we cannot gain additonal sensitivity..which I am not agreeing or dissagreeing with…
The superthin nature of the materiel used for a stacked backlit sensor will still result in shorter microlenses with smaller pixels..
Therefore iso will still be improved.
No. A microlens is attempting to adjust for both fill factor and for angle of light not being telecentric. BSI already has a very high fill factor, as do large photosite sensors (e.g. the Sony A7s, or the Nikon D4s). You can increase fill factor by using smaller process fabs (which is Canon’s full frame problem: they use a 600nm fab when others are using 200nm or even less these days).
As noted, right now sensors are about 50% efficient in converting photons. The quantum dot and similar technologies should get us far closer to 100%.
In terms of my reading of the research papers on graphene, there are three potential wins for it: speed, lower power, and better signal at the ADC (lower noise). It also has a big win for imaging in the non-visual spectrum.
Unlike Moore’s Law which has been a nearly linear progression, I’d tend to say that we’ve been on a more log progression with imaging. Early on we made big gains over X time, but this declines with each X period. Every now and again we can a technology that moves the curve over a bit (e.g. BSI), but the overall progress is still a curve with a long tail, not a straight line.
I don’t regard graphene as a big bump in the curve. Other technologies brewing in the labs are.
Thom, I think it was you who demonstrated stacking images from a 60 fps burst of a Nikon 1 camera. Isn’t that already a ‘technology’ able to deal with electron well saturation? Sure, even only taking two frames at 60 fps implies an exposure time between 1/60 and 1/30 s which is to slow for a lot of applications but combined with VR it can be sufficient for most static subjects.
Lot’s of small sensor cameras have implemented that for DR extension, though only in jpeg-mode, probably because some digital alignment as well HDR processing is taking place.
For static subjects, many of us have been using a variety of techniques (stacking, HDR, etc.) to deal with the fixed ceiling of well saturation and the noise floor. But it doesn’t work for moving subjects, esp. in bright, contrasty light. If the subject itself was well identified and isolated in “good” light, you could perhaps use masking techniques to deal with the background.
A lot of research has gone into trying to come up with a carry bit type of well (fills to saturation, sets a bit, fills again), but the problem is that you have to bring the well down to zero, which takes time, so we’re back in the realm of motion being a problem.
My understanding is graphene is still too expensive to mass manufacture. But I see big advances being made in its manufacturing process with one study even using a kitchen blender. So maybe in the next couple of years there will be an affordable sensor made with graphene.
The articles I have read seem to indicate this will be less expensive than current sensorsto produce.
If that were true they would switch in an instant. Profitability motivates change.
You keep saying Sony is sitting on technology in a marketing conspiracy to milk us from obsolete technology. I hate to trouble you but is there a link(s) supporting this? I’d just like to judge for myself how reliable this rumor/information is.
You use a Robert Snowden analogy. I’ll use a Popular Mechanics one. Popular Mechanics is known to have exciting tech on its covers. Maybe it becomes real, maybe it doesn’t. So – and this is a made up example – if they were to come out with an article on self-driving cars in the next issue (like Google’s) I’m not going to put off buying a new car because I suddenly believe all the cars with steering wheels in my local dealers’ lots are immediately obsolete.
As an aside, here’s an interesting sample of Popular Mechanics misses: http://www.popularmechanics.com/technology/engineering/future-that-never-was-next-gen-tech-concepts#slide-1.
The stuff I’ve read on, say, organic sensors looks exciting but haven’t rumors (e.g., Fuji/Panasonic will introduce cameras with them) haven’t panned out as accurate. Foveon is very interesting but we can see Sigma’s implementation of it and see there’s still problems to be solved. Certainly, Sony has better resources to solve these issues but that’s not to say they are easy ones to fix.
IMHO, Sony could just as easily release, say, a Foveon sensor if they had the tech ready and keep milking us for incremental increases in quality as they are doing with Bayer. In fact, it’d probably be easier for them than to keep trying to squeeze better quality out of mature Bayer technology.
Hi Mighty Sony, if you are reading this can you fix the shutter vibration problem on A7r before you bring out another new model. It makes ppl who bought your previous model like an idiot by buying your products.
did you notice that they releases fw updates deadly rare and all ppl bought any of their products “looks like an idiots”?
I work for a camera retailer in canada, and sony passed by to show this piece off.
The silent shutter mode really was super silent, it made me a bit uncomfortable at times because I was unaware if I actually took a picture. Would be a great stills camera on the film set, as well as for weddings. I was hoping that it would raise the flash sync speed too, but after trying I found out that flash is dissabled while silent mode is on.
For that camera very rarely you’ll need the flash. Silent mode is to avoid interrupting the event, so no logic to use flash in that situation.
I was trying it on a Profoto B1. I use lighting for fashion and editorial work, having a super high sync speed would be wonderful for outdoor photo shoots.
That would be cool but it doesn’t actually work like a leaf shutter (AKA an electronic global shutter) so it if it did work it may actually have more banding a lower sync speed than the regular shutter. Like the GH series electronic shutter this is still a rolling shutter so with lots of limitations and caveats but man I would have killed for this if I were still shooting publicity & archives for theaters.
I was hoping that Sony had secretly implemented a global shutter 🙁
I’ve been wanting a full frame camera with one for years.
Being able to kill for such a little thing. Hope I am not leaving nearby.
Rest easy Frankly, that time has passed and should I get more such gigs in the future I will now be able to acquire the A7S. ;-P
Yes and no, if you want to paint with light, you still need the brush.
Soundless-ness is definitely disconcerting at first! (though I’ve gotten use to it on the Nikon1 over time). Did the A7S provide full/12MP files in silent mode?
Yup, full Rez.
Is the shutter button still one stop top to bottom (like on a7r and most possibly a7)or is there finally a proper half stop to lock exposure and focus?
I set up lock on af, and did the half press like I usually do on the regular a7.
So I’m going to go with half stop.
I meant if the half press had tactile feedback, if it actually stopped you a little when half way pressed. There was non on the a7r I tried but it’s there on most cameras.
May I ask if the ISO is limited? On GX7 silent shutter only extends to ISO 3200.
Hi, did you notice how good or bad was the “rolling shutter” problem in stills when silent mode is on? (I’m not talking about video)…like, was there any kind of vertical distortion when shooting? And was the maximum shutter speed still 1/8000 or was somehow limited? Thank you for your feedback!
I know this has nothing to do with this topic, but I just gotta to tell someone; I just got an email from Adorama informing me that my A77II has been shipped to my house!
Congrats!
Come back and tell us how awesome it is. I just might buy one if the user reviews are good.
I’ll do that guys. This is great timing. Next Saturday the 7th, my local camera club is coming to my house to hike a trial I have cut into the woods behind my house. We photograph wild flowers then jump the fence and take photos of my neighbors cows and horses. I did not buy the lens they bundled with the camera. I plan on taking just two lenses for the hike, my Sony 70~ 300mm beer can with the SSM focus and my old Minolta 35~70mm.
I also have some other high quality FF lenses that I can use to show off the A77II.
A 90mm macro would be the most usefull lens possible for photographing the wild flowers.
Since you just got the a77II the lens I MOST HIGHLY RECOMMEND is 60mm tamron f 2.0 macro.
Its for C only and is therefore 90mm equivelent on your new camera.
Not only is it as good or better than any macro lens made but it is reasonably inexpensive.
The lens has been around a while so you might even find a good used one.
I’m sure that would be a good lens. Thank you. I happen to already have the Sony 100mm f/2.8 macro FF lens. It will become effectively a 200mm macro on my A77II. I’m attaching a photo I took with that lens on my A100 several years ago.
https://www.flickr.com/photos/24969967@N06/5461304681/player/
150mm equivelent.
Yes its a good lens.
Thank you Berry. I’m sure that would be a good lens. I happen to already own the Sony 100mm f/2.8 FF lens. I took the attached photo using that lens on my old A100 several years ago. The A77II should do even better.
https://www.flickr.com/photos/24969967@N06/5461304681/player/
Man.. I kept posting here that Digitalrev has the A77II in stock , a little less in price than official price too (I think it’s $1,196 including shipping and handling) and ships super fast.
Mine is on the way, 2 days or so.
Congrats anyway
Yet again Adorama gets stock before Sony does. I pre-ordered on the first day it was available, and it is still telling me June 9th.
Ugh…
But……congratulations on that! I envy you! 🙂
Good decision
Enjoy
There’s also official confirmation available: http://presscentre.sony.eu//content/detail.aspx?ReleaseID=9810&NewsAreaID=2
Gotta LOVE those HUGE photosites. The “photosites size doesn’t matter” crowd must be feeling pretty tight today. 😉
There’s a crowd?
Yeah…There is a crowd alright. I don’t know how many people kept swearing to me that “sensor” size matters not “pixel” size. I’m sure there are still people today that believe that a FF 36mp sensor will have the same iso and noise performance as a 12mp FF sensor when scaled down to 12mp.
These people only care about resolution and dont understand how photosites are like “photon buckets” and how charges are collected at the pixel level.
Bigger buckets offer more photons per pixel. More photons offer a hotter charge and more dynamic range potential with each pixel. Their “voltages” shoot up higher off the noise floor.
Smaller buckets saturate faster and produce smaller charges and potentially offer less dynamic range and operate at relatively low to the noise floor of the circuit.
Some people love to preach against simple physics.
Why are you talking about a noise if you know nothing about it?
Yeah? Sony knows what noise is. That’s why they made a 12mp sensor with very high dynamic range and a high signal to noise ratio. Based exactly on the principles I just laid out. But hey…what do I know? LOL. 🙂
I repeat, why you, who knows nothing about a noise are talking about it? And now are trying to hide behind Sony’s back. So, you seem to be some of the Sony’s decision makers?
BTW make the SNR value rationale (since you have just used this term) then and explain how did you lead to the conclusion this huge photosite sensor has higher SNR then, say, 36Mp. I mean, describe the comparison procedure, step by step.
You’d better invest your time in education than trolling. English language training would be a good start. Just saying..
Forget about me. Just go read Sony’s explanation as to why the A7s is so sensitive. If you still don’t accept it, then send them an email and tell them they are all wrong. Tell them your 36mp sensor is just as bright and clean as this A7s 12mp.
Can you really azure with the end results?
Sony made a 12 mp sensor to allow for 4k readout. For stills, if you can wait for a 1/8 of second to retrieve data from 36 mp sensor, 36 mp sensor is as fine as 12 mp sensor.
Yes, Sony did make this specifically for video. And, as a side affect of that goal, they ended up with an extremely sensitive, lower resolution phtography sensor. Even Sony admits that the reason its so sensitive is based on its photosites size. Their own press is very very clear about that fact.
That’s marketing. They also have said that SEL16F28 has good IQ typical to primes. Believe them, don’t you?
Ooh sweeeeet! Look at that: An amateur troll making his first baby steps..
Cliff – you are correct!
For most scenarios my 36mp D800e down-sampled and with a touch selective NR { which the extra detail allows for } output at the same size as my D4 will deliver near identical results up to and including 6400ISO. However they are different tools I use for their respective strengths.
SNR is near identical { at output size , print option DXO}
http://www.dxomark.com/Cameras/Compare/Side-by-side/Nikon-D4s-versus-Nikon-D800E___945_814
DR even at 6400ISO is under a stop. { at output size , print option DXO}
http://www.dxomark.com/Cameras/Compare/Side-by-side/Nikon-D4s-versus-Nikon-D800E___945_814
tonal range and colour sensitivity are also near identical.
Look here for your bucket scenario explained better:
http://www.luminous-landscape.com/essays/dxomark_sensor_for_benchmarking_cameras2.shtml
A big bucket gets more photons, but 4 smaller buckets together have almost the same number of photons (minus the ones being lost at the borders. So at very high ISO the lower MP sensor has the advantage of having less photons lost at the boundary). One may not get the same high ISO capability (but comes close), but can trade resolution for DR and lower noise, but not the other way round.
What is it… like 8 microns or something like that? That’s insane (love it!)
NEX7 3.9 microns
A7R 4.9 microns
A7S 8 microns??
Holding out for the 1sq ft
8 microns can be seen with naked eye! 🙂
Do you even own a camera ?
Please give us some RAW sample files please. Please!!!!!!
Ahhh…. the sweet sound of silence – just what I’ve been waiting for! I wonder if there’s a drawback?
I’ll be meeting a Sony camera product tester this weekend, and will be sure to ask! 🙂
Make sure to ask him about the A99II- we’d all appreciate it.
He’s an A99 product specialist. Anything specific you’d like to ask?
I’m sure they won’t give details. However, I’d be interested in knowing what their product vision is for Sony’s full frame alpha mount line. Maybe that would give us a clue if we could expect a similar A7S, A7R series of FF cameras. Just a thought.
Sensor specs
Timetable for organic sensors.
How about why Sony is neglecting a few key A-mount lenses? 400/4.5, wide angle primes, 70-200/4 for starters.
A-mount Zeiss or G 24-105mm f4? Sigma version is too heavy.
I’m also interested in the A99ii sensor. It would be great if it uses the 12mp sensor. Of course, it’s even better if it has the rumored organic sensor…
To me the a99 , is a bit of a mystery, not really strong at anything, a jack of all trades. ( Not super hi-res, not super fast etc. )
What is the primary customer group for the a99?
I’m fine with a leaf shutter…as long as they’ve got one.
1080P 10bits – 4:2:2 Please Mr. Sony !
You’ll need at least that to capture all the dynamic range at the base ISO but I fear it will not happen. We’ll probably see this sensor in a “professional” camera cost 2-10X as much and one of the big selling points will be 10bit (or higher) video to capture all that dynamic range. I’m not sure about processing requirements but 10bit 1080P would only bump up the bitrate to 65mbps—still very doable.
All we need is Vitaly, Someone 1.00 and their crew to break Sony’s software encryption, repackage new firmware and boom, considerably less compressed higher bitrate output from Sony cams.
My GH2 still rocks the pants of any AVCHD cam out there.
+++++ 1 !!!!
I’m with you on this. Wish we could get some help migrating S-LOG and XAVC over to the other A7 models (yes, they are much weaker in video but it would help to cut footage between them).
However, moving from 8 to 10 bit is a lot more complicated. No hacked camera has ever done that. Non of the GH hacks changed the bit depth and the ML Canon hack enabled RAW by finding a way to bypass the encoder altogether—not altering it. My understanding is that the bit DEPTH is a much more fundamental variable than the bit RATE. It is likely only changeable if someone can load software for a differenct codec into the firmware and/or if full XAVC is secretly embedded in the firmware an just not enabled.
I’m hoping it happens but I’m skeptical
Yes all this is really doable … 😉
15.3 stops at the base ISO? What happens over ISO800 will be another story.
So, now it is only “expandable” to ISO409600?
Like Freddy said, it would not be asking too much if there is 1080P at 10bits, would it?
BTW, I got used to the sweet sound of A7R. A silent shutter would be very boring….
landscape and studio work don’t worry about anything above base ISO any how.
Who would buy this camera for landscape and studio? Choose A7r for landscape or studio. You are not “loosing” that much DR.
This camera is not for that.
i use the d700 for landscape
It was always ‘expandable’ to 409600. The change is that now you can have ISO 100 in movie mode.
“BTW, I got used to the sweet sound of A7R. A silent shutter would be very boring….”
Then disable it and don’t use it.
Continuous shooting does that on its own and will default to the shot gun shutter. 😉
“Sounds terrific!”
Technically it doesn’t sound like anything… 😛
Silent slow rolling shutter (1/20s real speed).
Oh please, this would imply that all video and live view can only do a max of 20fps. Yes there will be rolling shutter, but this is just being silly.
Video uses a crop of the sensor, it does not read all sensor. Do the math yourself: 8 mpix crop of 12 mpix sensor at 30fps, means full 12 mpix at X fps. What is X? 😉
Right concept but wrong values. Video is a 10-megapixel crop: a full-width 16:9 crop of the 3:2 sensor. That translates to 1/25 s or faster.
Only 8 mpix are processed. And how do you know that complete lines are being read out? Sony A7s cannot even do full 4k (4096 pixels wide), only UHD (3840 pixels wide), I guess there is a reason for that.
And don’t forget that video reads out at 10 bits (if not 8 bits in case of A7s) per pixel, and still at other FF cameras work at 14 (and to reach their claimed DR without compression/posterization they better be 16).
I’d love to have a silent mode for travel photography (on countless occasions), but why put it only into the Movie Camera and not into the actual photo cameras A7 and A7R (especially the A7R, where the vibrations is causing problems); this could have been better managed…
The amount of pixels to make this happen and the energy needs to make it happen may be mass-produced-possible on a 12mp FF sensor right now and considerably more difficult on 24 and 36mp sensors of the same size.
you’re probably right. Hopefully they are able to do this on the next camera generation…
This is TOTALLY a stills camera. Ultra crops and 4×6′ prints aside this camera will produce a better quality photo with deeper dynamic range, less noise (so better, truer colour), with less blur (faster shutter speed) vs 36 mp of blur/ghosting/noise in all conditions except where your subject is still, your in bright light, or your using artificial light (flash) where it will produce comparable quality with a little extra DR headroom to deal with mixed lighting. The a7s shoots video? Who cares – this is a PREMIER event stills camera. The only areas I see the a7r trumping it is if you have to crop ALL THE TIME, your aiming to shoot exclusively static landscape on tripod or in good light or your a studio photog. I would also say that the deplorable low light AF on the a7 and even worse on the a7r actually means that those cameras cannot even function as fully functional stills cameras in many event parts such as the reception of a wedding. I bought the a6000 just so I could get the better AF in low light. The a7s will photographically slay the other two in most real world conditions. Just use superb glass and milk that iso for higher shutter speeds. .. as for why the OTHER stills/video hybrid cams don’t have the function see Hendrick’s response below
Of course the A7S is also a stills camera and you are totally right with what you say, but the marketing of the previous month emphasized heavily on the movie aspect of this camera, that doesn’t interest me at all, and very little on the photo aspect. I should have made it clearer, that my critique wasn’t aimed at the camera, which looks very interesting, but at the Sony marketing department.
The major advantage of the df, d800 (e) in comparison to the sony a7r and s is the AF SYSTEM why they havent got implemented the af system of the a6000 or a77II in the a7s………
The same thing why not the a7 af system in the a7r???
The hell , why not? Sadly, its a shame on you sony….you lost time and a bunch of money/sales…..
So we have to wait for the second generation of sony a7(r,s) or the a8 or a9 series…..
Time will tell…
Bear in mind that, when you compare Sony to Canon, what the shadows giveth the highlights taketh away to some extent… Having used both the A7r and several Canons I remain unconvinced that Sony sensors are as wildly better than Canon’s as many people make them out to be.
Even the 5D II has better highlight recoverability that the A7r and the 1Dx blows it away. The A7r, for example, clearly underexposes by around half a stop to preserve avoid highlight clipping. That’s actually a big advantage to canon IMO as abrupt clipping in highlights is probably worse than a bit of shadow noise.
The vast majority of “real world” shots are taken at low ISO and in decent light. The significant advantages in resolution of the 36mp cameras and even the 24mp options combined with the ability to downsize and add some NR { compensated for by the extra resoloution} can give very good results up to 6400 ISO. Above this you are into tiny niches which I am sure though vital to some are almost totally irrelevant to most.
The D800e and A7r already bring us DR over 14, and frankly the ultra high ISO samples I have seen from the A7s look dreadful. Just because a camera has an ISO setting on it doesnt make it good. As for AF are you certain that the A7s will be a low light star , personally I would wait and see
http://www.dxomark.com/Cameras/Compare/Side-by-side/Sony-A7R-versus-Nikon-D800E___917_814
Despite the hype I will be genuinely shocked when honest tests are done if the A7s genuinely advances high ISO performance compared to the current leaders { DF,D4s ?} by even a stop and I expect less. The A7s is a very specifically niche camera as you suggest { assuming good AF} it could indeed be an excellent event camera. Once the native mount lens chocie gets better.
12Mp is quite low and I am almost sure the AA filter on this cam will be a too much for stills, otherwise camera will be prone to moire in video. 36Mp will likely have better noise figures up to ISO12800 combined with sophisticated NR and downsampling to 12Mp, just like Nikon D800 has an advantage over D4 up to ISO12800 with DxO PRIME.
no, it hasn’t actually *been* released so it is a revision to the specs, not an upgrade. I agree that I’d love some more updates to the A7 but Sony’s already issued one within six months of release. That’s pretty good I think.
It was a bit of tongue and cheek response. I love my 7R. I didn’t realise the A7 had received an update, does that mean A7R? I should check up on it.
Nah, they should have brought out the FW update AFTER release like Fuji does so users can write glowing praise about how great they are for foxing the camera and adding new features AFTER they bought it 😉
May I ask if the S-Log2 can be applied on still photo mode? I am looking for one-shot HDR solution
just shoot raw
You can use the DR Mode. In this mode the camera Highlights the shadows. So you can use with this a one shot HDR. But the DR stops wide range is not so high as a real HDR shot. The A7s has a lot stops dynamic range, perhaps she can make a one shot HDR with the DR mode better then others. The HDR with the DR mode looks natural.
So manual lenses will give you a silent aperture too.
I wonder if they’ll also include an option for native lenses to work similarly and so affect the aperture immediately when changing it in aperture priority mode?
Native lenses could then also give a permanent DOF preview in the same way that manual lenses do.
They already do in ‘some’ lighting situations.
On my A7r (should be the same on NEX) in Aperture mode the aperture is changed immediately when changing it. It makes no noise when focussing and so would be fully silent on the A7s
I confess I’ve not tried it with a native lens, but using the LA-EA1 on my a7, it definitely doesn’t do that; there’s an additional noise made by the aperture, that is very noticable as I’m normally using electronic first curtain shutter.
It only works with native e-mount.
They already do so… on all E-Mount cameras with E-Mount lenses (NEX, …)
With the silent shutter, the FE lenses are a no-no. That annoying beep from the focus will spoil the whole purpose.
Anyone knows how to disable that?
???
settings (on the right, page 1) > audio signals > off
Thank you! 🙂
I often read in the discussions about a7s and very High ISO shots from the Graphene-based image sensor.
I have a question to the experts here:
What you think. How much ISO (Real ISO not expanded) is with a next generation camera with a graphene sensor and approx 8-12 MP possible?
Graphene based image sensor for photography- does not exist!
Hi,
In the labor exist this sensor here the article:
http://www.gizmag.com/graphene-imaging-sensor/27718/
The problem is- the press release tells not details. The 1000x more sensitive or 10 Stops more light.are difficult to compare. 10 Stops more to what? To a APS-C camera with Iso 25000 or to a other camera (FullFrame?) with more or less Iso value….
Read further down to see it’s just a mis-wording case.
Go, Sony!