Sony sensor engeneers told at Imaging Resource that “the A7 use the same sensor as the Sony A99, which does in fact have offset microlenses.”
Honestly, didn’t know the A99 had offset microlenses. I have searched the web and found no info about the A99 having that. Did you know it? Would like to find some hard evidence somewhere…
Note: Siammirrorless site crashed after me linking to his site. His Legacy Test will be posted soon on another site with stronger server. Will link to it as soon as it is there!
60 Responses
Most modern sensors do have offset but it is optimized for the angle of incidence you get from a 35mm mirror-box flange distance not the shorter flange distance in a mirrorless design. The A7 can likely get away with the non-optimized offset of a conventional DSLR type sensor due to the large pixel sites. The smaller pixel sites of the 36mp sensors likely are not able to collect enough light for good image quality in the corners unless a larger offset is built into the sensor.
Yeah good point. That makes sense.
You sir, win the physics answer of the day, my friend.
Offsetting a microlens is a simple manufacturing step.
By varying the temperature at which the mirolens is manufactured the size will vary…therefore the viewing distance will vary.
It is simply a matter of determining the correct temperature for the desired viewing distance.
It’s not as simple as that 😉 Sony is a a master of ultra fine injection molding. I guess there are just different molds for the different microlens sheets… It’s likely, that they saw no reason to change the microlens arrangement of the 24 MP A99 Sensor for already told reasons…
Changing the viewing distance is that simple.
I wonder if they could make a non-planar (curved) sensor to better angle the pixels towards the angle of incidence.
This would make a nice Sensor, difficult with existing though as they could not be focussed at the same plane over the sensor area. New Lenses would be necessary, so I think we might see this first (if ever) in something like a fixed lens Camera like the RX series.
It will be quite expensive.
A curved sensor would be a really silly idea when existing camera lenses are all designed to focus an image onto a flat plane. You would have to come up with an entirely different optical design. It also has a misconception that current lenses produce rays which somehow radiate out at a fixed angle from the centre. They don’t – on conventional lenses at least, the rays come in at many angles from the back of the lens, it’s just that the extremes of these are more pronounced at the sensor’s extremities. (Note that true rear-telecentric design lenses are different, but then the problem doesn’t exist).
It would also be extremely difficult to manufacture and impossible to produce on existing semi-conductor manufacturing facilities, all of which are geared to essentially flat surfaces.
So, in short, it’s not a good idea at all short of a complete redesign of optics and manufacturing processes.
Sony does have a patent for both a curved ff sensor and a 4 element 35mm lens for it.
This technology is best suited to fixed lens cameras.
Due to the immenent demise of the outdated cmos technology it is highly unlikely sony will develope this until after 2015 when the next new sensor technology will be introduced.
Also gapless microlens technology has been around a while and is unlikely to be only for the A7r
See here for the D7000 same as the NEX 5n
http://landingfield.wordpress.com/2012/04/
It’s actually the lens design that dictates the angle of incidence of light rays, not the flange distance. Of course a short registration allows for more extreme angles, but it doesn’t enforce it. It’s true that an SLR mirror box has the (somewhat fortuitous) side effect of forcing the rear focal length of the lens to be fairly high so limiting the extremes of angle of incidence, but the reverse is not true, albeit that the most compact lens designs will have this unfortunate feature. However, it is possible to design reasonably compact wide angle lenses that achieve a reasonable compromise. (Also worth noting that there tends to be a convergence of rear focal length on DSLR designs as they tend to be shortened for longer lenses using telephoto designs to achieve compactness whilst the opposite retrofocus design is used to allow wide angle lenses to be mounted – a camera designed to allow for both short and long rear focal length design lenses is a much more difficult proposition).
The danger of designing for the extremes of angle of incidence arrival is that they will tend to be sub optimal in other circumstances. The A7/A7r will still be required to mount lenses designed for the greater registration distances, especially for the very fastest and longest lenses which are likely to be designed for traditional SLR mounts (not to mention legacy).
And that is why we simply need an RX2 with a wide angle lens.
Indeed, if you want the best match, then design the lens, shutter and sensor as one.
It’s also possible to produce an interchangeable system where the sensor, lens, shutter and image processor is combined into a module. It’s the approach that Ricoh took with the GXR. However, it’s an expensive route to take, although it’s notable that one of Ricoh’s modules is designed for M-Mount interchangeable lenses. I suppose an equivalent system could be designed for FF, even with modules that supported some interchangeable mounts as well as fully integrated ones.
It is indeed expensive (GXR), but I think many would be interested. Add a service where people can send the lensor module back for a sensor and electronics update every couple of years and it could become quite interesting. There is certainly some merit to such an approach, having modules covering the basics, say 24,35,50 and 85 and an interchangeable lens one.
Sony is already producing two lensors which can be used with celphones.
The success or lack of success of these will determine the future of the concept
Same as A99, which was to be expected, does that make it an “old” sensor?
The sensor is still newer than the 5D mark 3. The new processor is creating significantly better detail and sharpness.
yeah but 5d3 outshine in high iso … and ofc lenses
sort of?
Nice to have this “sort of” confirmed, this explains the better performance Steve Huff got off the A7 with WA M lenses, compared to the A7r.
I remember when the a99 was new, one of the feature points was that the AA filter changed intensity as you moved away from the center of the sensor. I expect that this is what they are talking about in the case of the a7.
The question really is a question of how to define when a sensor is new.
We clearly have 99 OSPDAF-Points on the a99 and 117 on the A7. Does this not count as a new design? Where does ‘new design’ start?
If it goes directly from the factory to the camera it is new.
If they take it out of an old camera and reuse it is used
Smartass :p
I wrote ‘new design’ in the end, the question is still valid!
As silly as my statement is it seems to match sonys definition of new.
There different amount of OSPDAF sites on A99 and A7, that would be a clear indication of whether the sensors are same or not.
It could be the manufacturing process is the same for both, but different designs.
It’s new but that doesn’t mean much. Chip design, a realm within which sensor design definitely exists, is a highly iterative industry. In this case, most likely all the technologies in the a99 and a7 sensors are exactly the same, and it’s just a very minor adjustment of pixel arrangement for PDAF or slight addition of a few extra pixels or whatever. They probably kept as much the same as possible including the microlens/AA setup. To go any further in such a short time would be prohibitive in terms of design risk as well as manufacturing ramp-up.
One thing to note, there are only 15 more AF points on this sensor. If I was betting, I would bet that adding those few additional AF points was the one and only thing they changed from the a99 sensor. Any other improvements are in the JPEG engine given life from the faster processor.
Well, it is pretty clear. Any iteration is called a “new design” for promotional purposes.
Maybe now the sensor in the a99 (and the camera itself) will get some proper recognition as one of the best sensor’s to date. The images that come from the a99 are IMPRESSIVE ! Even compared to my RX1R theres hardly and difference.
Looks OK, apart from the terrible ISO performance. http://www.dxomark.com/Cameras/Camera-Sensor-Database/Sony/SLT-Alpha-99
Why is it so terrible bad compared to 5d Mark III and D800 in raw numbers?
I was also told by a Sony executive that the a7 uses the same sensor as the a99 & RX-1.
I remember a comment from a while back (I think it was here at SAR… could have been elsewhere) where someone had asked Sony if the A7 had the new offset micro lens design.
They said Sony’s reply was ‘the mirco lenses were already optimally placed on the A7 sensor’ (implying that the sensor does in fact use micro lenses but that no change was required to the placement as used in the A99?)
Sorry I can’t recall/cite the reference.
Sony could now give us an upgraded firmware for the A99 based on the progress obtained on IQ for the A7, of course taking into account that they have different image processors.
The oft-requested firmware improvements on the a99 aren’t particularly sensor-centric. What is it that you would like improved that related to the sensor plus firmware of the a99?
There will noy be any major new advances to the cmos sensor.
Its a dead technolgy soon to be replaced with high resolution organiic stacked sensors in 2015.
Mo one is going to waste money to make major omprovementd in it anymore.
We will see minor improvements and yhey will call it all new.
No mew cmos will be produced with over 36mp.
Sony does not really want to replace canon or nikon in the # 1 or 2 camera slot because then someone like samsung will come in to fill the void and kick sonys butt again.
Sony is yhe leader in sensor sales and only wants to keep the pressure on the competition so they will buy newer sensors.
Sony wants Canon and Nikon right where they are in the. #1 and #2 slots.
Thats why sony will wait till 2015 to introduce thier organic sensor so Canon and Nikon can indroduce thier own version and keep up.
This opinion is based on my interprtation of a top sony official as recently posted on sar…I believe it was a reposting ofvan interview from another website.
How do you say everything so matter-of-factly with a serious lack of facts?
The proof can be given by himself so he likes discussing with himself the most as you can see 😉
Footnotes
He’s not the only one who likes discussing things with himself.
I agree with you. There are other people who reply to themselves.
+1 very good
It’s called “deductive method”. I like it.
Makes sense. You probably aren’t very far from how it’s going to be.
I thought it was a totaly New sensor?
I guess that would depend on who you ask. I could care less though because the A99 & rx1 has a fine sensor.
hahahaha supposedly, but once again, the a7r performance was too suspiciously similar to d800, both said to be “new” sensors
This is the first time they confess about 24mp sensor. The 36mp is presumably “new” too.
Therecwill not be another trul new cmos senso from sony EVER.
Cmos is dead. If you want new wait till 2015 and toally new ( this time for real) sensors will come out…layered/stacked/organic/high resolution?
I just really wish IBIS was included… After having taken apart/repaired an a900, an a33, 2 a77s, and an a99, the additional sensor package that constitutes sensor stabilization doesn’t seem particularly large.
See this image:
http://goo.gl/a7ITIm
Its the old double screw ypu sony user job.
Sony doesnt offer ibis in e mount an tamron leaves vr out of thier lenses because sony users diny need them plus charges us the same orice for a lesser product.
Tamron also doesnt offer their teleconverters in sony mount…forcing us tobuy inferior kenko converters if we want one.
As impalatable as this situation is…to add insult to injury….Sony is the major stockholder for tamron stock and basically controls wjat tamron does!!!!!!!!!!
1) It is true that Tamron charges the same for their lenses with A-mount lacking VC, as they charges for their Canon and Nikon versions which include VC. However, this does not necessarily implies that Tamron is overcharging us Sony users. We can be pretty sure that they are selling several times more of the Canon and Nikon versions than of the Sony version. Hence, the Sony A-mount version will be more costly per item for Tamron to produce.
2) Sony may be the largest shareholder in Tamron, but they are certainly not a major shareholder. They are still a minority shareholder – as far as I know Sony owns 11% only of the Tamron stock).
.
11% and they are the largest shareholder? Sound pretty major to me.
As for the no vr issue…some small token discount would go a long way towards costomer satisfaction without adversly effecting tamron profits.
Lastly…offer tamron best teleconverters in sony mpunt…even if they need to charge a little more.
If tamron shoued a little more concern for its sony costomers then they would sell them more lenses.
Words have often a bit different meanings in different languages. !1% is certainly noting major in my language. Sony may get 89% votes against them at Tamron, right?
A majority sharholder has to have MORE than 50% of the stock, usually 51% to be regarded as such. Provided that all shares have the same voting value, which is not always the case. In those cases their different weights are of course also taken into the calculation.
here is the site, that tells A7 has offset microlense. It shares the same sensor with A99.
http://www.imaging-resource.com/news/2013/11/08/sony-a7-microlenses-are-offset
Mr.Dave Etchells interviewed Sony sensor engineer.
Sounds like another Mark Weir white lie, not full truth. Guy does that all the time.
Sony Photo courses (in Latvian) http://foto-kursi.lv/wp/