Ok, for now it’s “just” a rumor… but a pretty exciting one 😉 .
According to SonyAlphaRumors here, the Sony A7000 (and therefore maybe also the Fuji X-PRO2) could be the first mirrorless cameras of the world featuring a native “on sensor HDR”.
SAR says that “both, stills and video are capable of 15.5 stops of dynamic range!”
15.5 stops!!! That’s more than any Sony Full Frame A7 series camera! In fact, DxO calculates the following DR:
– Sony A7s: 13.2 f-stops
– Sony A7: 14.2 f-stops
– Sony A7r: 14.1 f-stops
– X-Trans 1 (X-Pro1/X-E1) has about 9.5 f-stops – source
– X-Trans 2 (X-T1, X-E2 & Co) has about 13 f-stops according to measurments of Jordan (Admiringlight) and also imaging-resource, which puts it slightly ahead of the A6000.
If the rumor is right, then this feature would make the new APS-C sensor generation even more advanced than the latest Full Frame sensor of the brand new Sony A7rII.
And I’m curious to see if the special X-Trans CFA will give us even more than the rumored 15.5 f-stops of the A7000.
HOW DOES IT WORK?
According to SAR, a similar technology is already used in the Xperia Z smartphone for video recording. Sony describes it here:
HDR for VIDEO: To achieve true HDR video recording, the advanced Exmor RS camera sensor in Xperia Z changes the exposure every two lines of pixels. This creates a spatially varying exposure (SVE) array image, which is then used to make an image that is perfectly exposed in both dark and light areas. To simplify the concept, for each frame the camera shoots one picture with two different exposure levels at the same time, one light and one dark. Then, an intelligent software algorithm is used to combine (synthesize) the two shots into one picture, pixel by pixel. This allows you to see the details of both the dark and the bright parts of the picture.
And for Stills? Maybe it works as described in this Sony Patent here, that shows a sensor, where every pixel is exposed differently. So, the pixels in a dark area would become “long exposure pixels” and those in bright areas “short exposure pixels”. But the rumor published on SAR does not specify if it works this way.
In the past, Fuji used the EXR sensor (which can be found in the X-S1 and X10) to expand dynamic range. But in the battle between EXR and X-TRANS, Fuji went for the latter one.
Fuji currently offers the efficient DR expansion feature (DR200 = 1 stop underexposure / DR400 = 2 stops underexposure). The extra Dynamic Range is stored into the RAW files and allows you to recover a lot of highlight detail in post.
It’s time to update 2 of the 3 polls launched after my Sony A7000 sensor specs rumor of last week.
[poll id=”117″]
[poll id=”116″]
[poll id=”96″]
Push the READ MORE button to read why any Sony A7000 sensor rumor is potentially interesting also for Fuji X-shooters
Here is what we know:
Fuji’s current X-TRANS APS-C sensor is manufactured by Sony (the 2/3 sensors are made by Toshiba)
As soon as Sony implements a brand new sensor in one of their cameras, Fuji has to wait a few months before using it.
Here is what we don’t know:
We don’t know if Fuji will buy the sensor for the Fuji X-PRO2 and X-T2 from Sony.
We don’t know if Fuji is waiting for the new Sony A7000 sensor (according to our rumors with X-TRANS CFA)
Here are the rumors:
The Sony A7000 has been delayed from late spring to late summer. At the same time, the Fuji X-PRO2 has been delayed from late 2015 to 2016.
So, no rumor, it’s all just a more or less plausible speculation that the Fuji X-PRO2 will sport the same new sensor that Sony is developing for the Sony A7000. But if it does, than any rumor regarding the Sony A7000 is interesting also for us X-shooters.
52 Responses
The current XTrans sensor does not have 9 stops of DR. It’s similar to the current a6000 (the RAWs are calibrated differently, so the a6000 has about a stop more highlight detail once clipped, but the X series has FAR more shadow detail available). I’ve done side by side DR tests, and if anything, I’d put the current X sensors slightly ahead of the a6000 in Total DR, which would be up around 13 stops.
I’ve updated the article with your considerations.
The important question for me is: Will this work in RAW mode?
Obviously yes, provided your chosen RAW developer can handle it.
Another change to conversion algoritms, and support probs.
As it is said that Fuji made an option between EXR and X-Trans before, if this rumor materializes does it mean X-Tans color array will be left behind and the new sensor will show either Bayer or something else?
Yes, that’s what it would mean. X-Trans and this technology wouldn’t really work together.
What it would also mean is that Sony will run into the same problems as Fuji: Adobe being slow and crappy at implementing proper algorithms to make use of the full potential of such technologies.
Probably yes, but will be supported by Phase One C1 as it is the official Sony converter.
You can’t just implement a 2 line skip with X-Trans. X-Trans would require a 6 line skip.
Btw, the concept is very familiar, that’s basically how Fuji’s (now abandoned) EXR sensors worked. The difference is that Sony does it w/o sacrificing resolution by interpolating the two exposures.
If you have to interpolate, you’d still lose some resolution compared to normal sensors, wouldn’t you?
IMHO not really – you have the “resolution information” for all areas, weather they have been exposed low or high – at least for the overlapping part of DR.
For the extremes you are correct – you have half the resolution information in highlights or shadows – but that’s still better than no information.
Didn’t know, that Fuji used this trick already… could be an interesting dilemma for Sony to file a patent, which main innovation is Fujifilm’s prior-art… maybe some kind of “we won’t attack your patent, if we also can use your implemented sensor later coopetition?” 😉
It makes sense.
Thanks for the answer.
The Sony approach is probably sufficiently different to qualify as patentable (look at all the patented hand dryers running around in public restrooms) but even if not, they may have worked out a licensing deal with Fuji.
Added in edit: In particular, if you look at an EXR filter array, you’ll see that it’s essentially two Bayer arrays offset by one pixel, at a diagonal, so that you can use each capture to correct the other. The Sony array here looks to be a basic Bayer array; rather than two captures offset by one diagonal, they have two captures offset by two horizontal rows. That’s both doable and intriguing, but you face exactly the same problems in resolution that Fuji faced. The algorithms they use will be interesting, to say the least.
I think you can skip just two lines. The algorithm will take care of the rest.
But look at the Sony image: it’s an GRBW sensor. I’m interested to see what the white pixel’s role is in this story..
This needs a meme, methinks.
http://cdn.meme.am/instances2/500x/646023.jpg
Even though this rumour is just a rumour, I find it very interesting. It has always been my experience that people will tend to refine and adapt existing technology, meaning that, as is illustrated here, more photosites does not imply more megapixels but rather providing the ability to do other stuff with the extra photosites. As said, this configuration is a rumour, but it might also be why we’ll never see the mythical “organic” sensor; as it is not enough of a leap in technology over what can be done with existing sensors.
Admin, please don’t use SAR as a source. Friend of yours or not. The guy puts more effort into bad mouthing other companies and starting gear wars, than he does into actually posting articles about real photography.
bad mouthing other companies? Gear wars? Never noticed that
SAR and FR have the same owner and probably the same administrator.
Check IP address for both domains, you will see that both are hosted on the same sever (IP: 84.244.183.82).
Patrick, can you confirm? 😉
SAR = Andrea
FR = Patrick (me)
And, unless I’m not Dr. Jekyll & Mr. Hyde, this makes us two different persons 😉 .
Believe me, one blog + a day job are truely more than enough work for me… so let Andrea do the (great) job on SAR.
We are using the same platform (same server, same sys-admin and so on), to save a bit on the costs of the sites (CW, SAR, FR, 43…)
SAR, CW and FR sitting together at dinner: https://www.facebook.com/fujirumors/photos/pb.267787456620125.-2207520000.1436362792./652327204832813/?type=3&theater
Great job at SAR? Why are so many people making jokes about it then? Posting every wild speculation, hyping up every single improvement as THE NEW REVOLUTION in camera imaging, the notorious “enthusiasm” of SAR readers (they are everywhere, posting on Nikon, Fuji, Canon, dpreview) – who’s responsible for that?
There are many people making jokes also about FujiRumors. So I started to track all the rumors, and it turns out almost 90% are correct.
Guess what happened?
There are even more people making jokes about FujiRumors (like: He posts fake rumors to collect clicks / He posts every cheap anonymous rumors to get traffic / He just wants to harvest your money / He…. bla bla bla….)
So, the fact that people make jokes on SAR, doesn’t mean that SAR isn’t doing a great job, but just that he has way more readers and therefore way more trolls, and that he has competitors that systematically spread shit on him, because they do not work a fraction of what the SAR admin works, but would like to have the same success.
I know that guy very well. We sit together and drink a beer now and then. And I know what is going on behind the scenes of SAR… and I know that he is honest and works very hard for the rumors. I know it, because I know him.
But since you don’t know him, and base your judgment on comments of anonymous readers, I can fully understand your scepticism.
If you want, you can go through all the rumors SAR published in the last 12 months, track them, and see what his hit rate is. Probably around 75-80% I’d guess.
But the point is, our brain remembers just the 20% wrong rumors. It’s evolutionary, because remembering our faults and errors helped our ancestors to survive in the old wild world… but that’s another discussion.
Anyway… as said, I can fully understand that, with the information you have now, you don’t trust the SAR admin. All I can do is to recommend you to track his rumors.
Ever heard of shared hosting? ))
It is possible to do that with a Canon device and magic lentern intalled (custom firmware).
When you take a shot, the camera produce 2 photos with 2 differents ISO (some pixels line at 200 and some at 800 for exemple). You have to combine them with a computer software. The result is good, the DR is big but you loose in resolution and you produce some aliasing.
“The extra Dynamic Range is stored into the RAW files”. That is a misleading sentence. DR expansion is a feature of the JPG engine and doesn’t affect the RAW file at all (that’s why you can change DR when you reprocess a RAW file).
As far as I know, the DR function in Fuji cameras affects also the RAW file.
Are you sure? I may be worng but in Rico’s book about X-E2 he said the opposite if I remember correctly
At least X100 RAW files darken according to DR adjustments. You get JPEG files showing extended DR effects but to convert RAF files you start with a file showing a histohram curve pushed to the left.
afaik it doesn’t change the raw per se.. it just affects the iso of the raw file. without dr the raw would be iso 800 with dr. it is taken at iso 400 for example
DR200 underexposes the RAW by 1 stop, while DR400 underexposes the RAW by 2 stops. So it does affect RAW.
@Patrick If that was the case, you wouldn’t be able to change the DR setting retroactively by reprocessing a given RAW file.
I know it’s a tricky subject (it’s basically related to that ISO-less sensor thing). Took me a while to fully understand it, too. No offense! 🙂
Does it mean it’s possible for X-Pro2 (if it uses this sensor) to have about 9 stops of DR with 6400iso RAW? That would be exellent. But can CFA of X-Trans be used with this sensor at all?
Fuji must compete on portability/minaturization, quality control/customer service, price point and critically… affordable world class glass. Growing these digital photo/video platforms take huge amounts of man-hours, the “smaller” teams are at a disadvantage when compared to the driven armies at canon and sony/minolta, problematic when the flagship camera release is delayed per contract stipulation with sony, a competitor. I don’t know answer, maybe fuji can just make fabulous glass for sony, canon etc ?? Or solve the holy grail of blending pro camera with smart phone ?? Is fuji fighting unwinnable war ??
“when compared to the driven armies at canon and sony/minolta…”
1. Minolta’s photo division is no more.
2. Fuji is in MUCH better shape financially than Sony, and diversified in much more profitable market segments. And they are at no disadvantage when it comes to their imaging division, which has 80+ years of experience … more than Sony will ever have.
3. There is zero hard evidence out there of any contract stipulations concerning competing cameras not being allowed to use new sensors manufactured by Sony Semiconductor (which is an entirely different division form Sony Imaging).
4. This is not a war, Fujifilm’s imaging division is doing just fine, and I’d be far more concerned about Sony’s health, if I were you. Not to mention Nikon’s.
that’s just sniping and nitpicking, the imperative idea is that sony/canon are running 100 miles-an-hour, fuji continues to fall farther behind and imo needs a survive’ability angle outside of the current box
Sony alone is selling more mirrorless cameras than all the other manufacturers together. Mirrorless is a way smaller market than DSRS; but inside mirrorless, Sony rules. I shoot FUji X, but they are a niche player, whereas Sony always is mainstream….
With each new sensor generation, motion artifacts are getting weirder and weirder.
This looks like an advanced extrapolation algorithm to me in terms of increased distance between each color block.
X-pro1 and X-E1 have both 13 stops of DR. As tested on Imaging-Resource. And both have better DR then x-trans II cameras in Jpeg
But it drops quickly as you turn up the iso. However, new working tech is always welcomed.
It seems something as “on the fly” bracketing or instant bracketing. But I don’t understand if resolution remains the same and probably the exposure time will be a bit higher (to expose the +1 frame, for example).
I don’t love that new poll, Patrick. 🙂 You’re asking us to choose what aspect of a modern sensor design we most want. What I want is a sensor with *all* the latest aspects of modern sensor design. I will be disappointed if the new sensor doesn’t take advantage of all the latest Sony tech.
To be honest this sounds like a recipe for disaster if the CFA has anything to do with the image above. So many problems I can think of that this would create. And Adobe is only one of them, the rest lies with the sensor itself 😉
If this is true, then Apple’s approach to on-sensor HDR seems much more intelligent to me. That’s where I see the future. Alternatively, a sensor that registers the time till it is saturated. That would give us some serious DR…
A “sensor” doesn’t saturate; a photosite does. I think you mean a sensor whose photosites register the time that they saturate, perhaps even reset at that time?
Yeah, obviously. Mr_Nitpicker 😉
What ever happened to the ‘organic’ sensor that Panasonic and Fuijifilm were jointly developing?
Has Panasonic dropped out of the race and conceded the sensor field to Sony?
Still in development as I understand it. This other new technology discussed here would serve as the next stop-gap until the organic sensor is ready for primetime (assuming something better doesn’t usurp organic’s capabilities in the meantime).
If it produces higher dynamic range images without motion artifacts or compromised resolution, then it may be a great idea. I have heard good things about Fuji’s EXR sensors of the past, so perhaps this would be an improvement over those sensors.
If so, I hope Fuji will provide its own bare-bones raw converter or the ability to convert the RAW files to 16-bit tiff using the in-camera raw converter. Currently, I am using the in-camera raw converter for my X-Pro 1 files, and the results are better than using Capture One, which is also pretty good. I wish they took the in-camera converter and provided a similar Windows/Mac-based converter.
To be honest, I wish Fuji had kept EXR and developed it rather than creating all the problems associated with X-Trans. I actually hate X-Trans at this point. And Adobe.
Let’s see if I understand the idea well: Samsung makes the best APS-C hybrid sensors out there (see for example http://www.dxomark.com/Reviews/Samsung-NX1-review-Hail-the-new-king-of-APS-C-hybrids), Fuji can build their own sensors and X-Trans keeps coming up in all the polls of pro and amateur users alike as one of the reasons people adopt Fuji, hence we are discussing of using a non existent Sony sensor and dropping the X-Trans? Wow, way to go!
This is close to what Fujifilm did more than 10years ago in the S3pro and S5pro: Sensor was called sCCD SR. It still producest some of the smoothest files I have seen out of a camera…. Nice to see that Sony catches up with what Fuji did in 2004-06.