This is a guest post from Zoltán Biró. Feel free to contact me at 43rumors@gmail.com if you want to publish an article on 43rumors!
–
Square sensors for mirrorfree cameras (by Zoltán Biró).
Once imaging technologies have evolved, some interesting projects become achievable far more easily than before. That’s the case with the square-shaped multi-orientation and multi-aspect ratio sensor, which would pose a real problem for DSLR technology but are quite easy to implement for mirrorfree cameras.
What is the square-shaped multi-orientation and multi aspect ratio sensor (SMOAR)?
The SMOAR is a square sensor which – unlike e.g. some medium-format cameras – should cover not just the image circle for 1:1 (square) format but also for 4:3 portrait, 4:3 landscape, 16:9 portrait and 16:9 landscape modes respectively. See the illustration below (the 4:3 modes are omitted for simplicity, they have to fall anyway between 16:9 and 1:1). Somewhat similar to ‘conventional’ multi-aspect ratio sensors – and opposite to classical square formats – there is no mode where all the pixels from the whole sensor would be used simultaneously.
What would be the main benefit of such sensors?
This type of sensor renders the portrait grip and camera rotation obsolete: you can always hold the camera in its ‘natural’ (landscape) position even if you want to shot in portrait mode. The orientation could be set (together with the aspect ratio) using a button or switch – instead of rotating the camera body.
Advantages (beside the main benefit):
- A simple tilting screen/viewfinder would be enough for all situations, no need for the swivelling screen.
- Using a tripod is much easier as there is no need for a special head which lets the camera be used in portrait mode. Moreover, the camera always stays centred on the tripod, its centre of mass would fall much closer to the tripod axis for enhanced stability. Also much more convenient for shooting mixed portrait/landscape from tripod.
- This is even more evident when using a monopod in which case shooting with rotated camera is a real challenge.
- True (not cropped) 1:1 aspect ratio would become available for high resolution square photos.
- When shooting in RAW you can even decide lately whether you want to use portrait, landscape or square crop.
- No need of orientation sensor.
- No need of rotating on-screen information for a better readability.
- No need to ‘learn’ a much less comfortable layout when keeping the camera rotated (besides the normal position where the controls are already optimised for maximum comfort).
Disadvantages:
- Obviously, the sensor has to be bigger, which means less space for additional circuitry, more heat dissipation, bigger cost etc. However, the sensor size won’t exceed even the (Sony) APS-C sensor surface.
- Manufacturers would either accept that the viewfinder/screen image has to be smaller when shooting portraits or should implement square EVF/LCD. Although a square viewfinder may be a bit more pricy, a square rear screen would be even more convenient as it allows more space for additional buttons.
- The mechanical shutter has to be also bigger which would be the only serious drawback (GM1 can be such small because of its hybrid shutter, even a pure mechanical 4:3 shutter wouldn’t fit in the body). The size remains an issue only until mirrorless shutters become exclusively electronic ones (what I hope will happen very soon).
- Some lens hoods are asymmetrical and optimised for landscape use. This is a real problem, but fortunately this is mostly the case of wide angle lenses where portrait shooting is hardly typical anyway.
- The sensor readout time might be increased (but likely just for RAW shooting if the reading is made already optimised for the appropriate jpeg orientation).
Myths:
- The body should be bigger to house the bigger sensor. False, as the sensor would fit anyway behind the (much larger) mount assembly.
- The electrical contacts of the mount may obtrude and obscure the sensor. False, as they are placed way outside the image circle.
- Lenses are more or less optically designed to be used in landscape mode. False, all but a few cine- lenses have circular symmetry alongside the main axis.
Why this sensor would be much easier to implement with mirrorless systems than for DSLRs?
The flange distance should be increased due to the fact that the mirror assembly (mirror, pentaprism) should be also bigger, this would be unacceptable for existing DSLR mounts. Fortunately this problem is actually no problem for mirrorless systems.
If this sensor is such good idea why manufacturers have not implemented it already?
This is mainly because of conservatism. Just take a look e.g. on the lack of success of mirrorless systems despite of their long-confirmed viability. The other reason may be the problem of the mechanical shutter which – at least as I hope – will become history within a few years.
So, dear colleagues, what do you think, is this proposal worth to be focused on by Panasonic or Olympus – at least for premium models, where the size or the cost factor are not so critical?
[poll id=”242″]
Kesztió
102 Responses
Thanks for that post! I’ve always wished for a square sensor too…
Although not a square, this reminds me how I miss the multi-aspect sensor of the Panasonic GH2. This allowed us 3:2 SLR shooters to get more resolution out of m43 without just cropping a 4:3 image. (Please don’t tell me how the 3:2 ratio sucks and 4:3 is better; I’m sick of that discussion.) A true multi-aspect sensor would be a lot of fun. 1:1 really gets you thinking differently about composition.
The GH2 sensor was great. I loved switching aspect ratios in different situations. I really miss it in the GX7. I wish Panasonic would bring it back!
Shooting with the 7-14mm @ 7mm in 16:9, landscape or portrait mode, is just great!
That’s effectively putting an APS-C sized sensor into m43 body. It would also allow to use native FOV with Sigma DN lenses.
>However, the sensor size won’t exceed even the (Sony) APS-C sensor surface.
That’s utterly wrong… The sensor diagonal would be slightly smaller (as for 1.62 crop) than Sony’s APSC, but the area would be much larger because of it’s square.
Well, GH2 uses 19mm of the sensor horizontally. So a square version would probably be 19x19mm, which is less than 16x24mm. 361 vs 384 mm2.
Not done because of conservatism of manufacturers… What did you expect when everyone makes cameras with fake pentaprisms, and those fake pentaprisms are praised by users?
pseudo SLR does have upside to it : its tried tested and true. The EVF recesses out at a very convenient place and it works just fine.
Thinking outside of the box is a good thing and I applaud every attempts, but the results aren’t always happy.
Guess what, rectangular 3:2 sensor is tried tested and true.
Happens to be my favorite ratio for paintings 😉
I’m not against a square-sensor option at all, but if it was more expensive than the same camera with a traditionally 4:3 shaped sensor, I would always choose the traditional option. Turning the camera for portrait mode was never a problem for me, and I like 4:3 better than 1:1, so a sensor I have to crop for getting 1:1 makes more sense for me than one I have to crop to get 4:3.
APS-C cameras don’t seem to be pricier than 43 ones.
Don’t worry, I am extremely confident that you would get all the classic aspect ratios to choose from. 1:1, 4:3, 3:2 and 16:9, if not custom with electronic shutter…
But it does makes me wonder too. Like seriously? It really is a struggle to hold a camera in portrait mode?
I’ve been for this all along. I loved shooting the square format Hassy and Rollei cameras.
Now I have an authoritative article to back me up.
Maybe that is why the LX9 will not use the Sony 1″ sensor? ….. Just hoping. 😀
I think that idea would be very nice, but I still would like to have a swivel screen which can be folded to protect the screen itself (but no must have at all).
Looks like Admin has cut a deal with kesztió! No more campaigning?
Not that I was complaining, I think it’s a great idea. Keep the raws somewhat circular, just have masking for previews/jpegs. Not sure about a square screen though..
At least we would need new lens hoods for some lenses. Maybe that’s the reason Olympus sells it extra… 😉
The lens hood issue only applies to shaped hoods, a conventional circular hood would result in no difference, other than dimension, it would need to be of a slightly larger diameter, in keeping with the image circle.
Why not take it a step further an make it a hexagonal sensor? That’s better use of sensor area and you could make more sensors out of a wafer, making it cheaper.
round sensor ?
Hexagonal sensor allows a better use of the wafer surface :-p
It is time to make a patent on wafer cutting along hexagonal patterns.
The main advantage of hexagonal pixel is that instead of 4, they have 6 neighbor pixels. That would means that green pixels would always have at least one green neighbor. That would allow for some of the X-Trans magic to happen: use the difference between the values of adjacent pixels of the same color for noise reduction.
At least on paper hexagonal pattern is not that hard to accomplish: you simply shift every second row of pixels down by half height. (Or make smaller pixels, then bin every 4 of them in groups, but in odd rows, you start not from the first pixel, but the second: odd rows would be shifted by half pixel height.)
An interresting question, would that (tend to) reduce the need for low pass filter?
(edit):
Not that we really know yet.
A thing to consider for early adopters. I am not sure Adobe and co would have their algorithm quite ready and optimal on the first year.
> the need for low pass filter?
One way or another the filter has to be there: either in hardware or software form.
X-Trans removed the OLPF, but images are still “soft” in pixel-peeper eyes, since the X-Trans pattern is 3×3, compared to 2×2 of Bayer sensor.
The only way to remove OLPF is to go with the true color pixels, like Foveon. For as long as the pixels are monochrome, you need some form of anti-aliasing.
old Fuji like S 602 Z ?
They experimented with similar ideas.
The S 602 Z was pretty good
> old Fuji like S 602 Z ?
No. In case of the Fuji the pixel layout was basically the same checkers, but turned by 45 degrees.
Note that a hexagonal pixel pattern is different from a hexagonal sensor. You could have square pixels on a hexagonal sensor or vice versa.
I was going to suggest this too. I’m not actually sure how wafers are cut.
Either way, you wouldn’t necessarily have to put working pixels outside the image circle, although it might be useful with some lenses.
I’m not sure that this is feasible, given the potential issues with the other electronics. (You need to do something with the charge held in the sensor.)
Good idea in terms of economics, though, given that the wafers are also circular.
Square sensor are better infact imaging scientific astro sensor are all squared and less megapixel with magical lowlight performance.
“scientific astro sensor”
They have a completely diff goal. They want to exploit the lens completely the picture composition is not important.
Interestingly, thats more or less the upside I saw to such sensor.
Not for composition, that would be a glorified crutch to those who somehow can’t hold a camera upright. Sorry to say. But for pushing the envelope and getting everything out of a lens, it would be interesting to see what a square sensor brings to the table. Would it be worth it? Maybe.
It’s a great idea, but it’ll never happen because output media (printing papers and computer monitors) stubbornly remain rectangular.
And this will be the multiaspect FT sensor of a surprise camera by Panasonic… is a dream
You do not print RAWs. Image is PPed first. Square sensor simply provides for more PP possibilities.
You can crop physical media. Monitors and other screens come is a variety of aspect ratios — 16:9 is most common now for ‘consumer’ products, but 4:3 is still common and 16:10 is available. In future, it may be cost-effective to build display devices to random aspect ratios, or even non-rectangular shapes.
Artists used to select the canvas to suite what they required (within limitations), and it seems obvious to me that allowing for a choice of aspect ratios in a camera could be a worthwhile trade-off.
First of all, I’m all for this idea; it allows you to fully exploit the lens’ imaging circle regardless of the aspect ratio you choose. A major challenge will be finding sufficient room for a rear screen where the preview image isn’t smaller than it is now on the typical, horizontal, 3:2 or 4:3 screens. I’d rather rotate the camera than end up with a tiny vertical preview when shooting portrait photos with the camera in landscape orientation. Nor do I want to go back to a larger camera…
Let’s consider the larger body you think of as no more than an E-M5 with the single part of the grip as an integral part of the body, lens can drop down about 4 mm to allow for the larger sensor and what becomes a quite body space for any extra electronics required or possibly larger heat sinking for the sensor.
It is a great idea, though hardly a new one. Personally, I find the square format the most natural for creating photographic compositions. However, the m4/3 standard was designed with software-implemented geometric correction in mind. Stretching the image circle would surely emphasize the weakness of the optics towards their periphery. Hence new, larger lenses would have to be designed for the square format, a proposition which the manufacturers are unlikely to be keen of.
That said, I am all in for a new mirrorless system based on a 32X32mm sensor (mini-MF). Panasonic and Olympus must already possess the experience and the expertise required for developing it, but clearly the current climate in the market prohibits such ventures.
There are already Sigma DN lenses in m43 mount for APS-C image circle.
True, and there are also a few native m4/3 lenses which rely on optical correction (the 25, 45 and 42.5 Panaleicas, Olympus 25, 45, 60 and 75 and perhaps one or two more). Nevertheless, the zooms and wide lenses would show their weakness to a greater extent.
No, you don’t need larger lenses, the image diagonal is the same. Software correction is slightly easier with a square format as less of the most-distorted periphery is in the image
r
@Kesztio….Did you send it to camera company?
I’m pretty sure all mirrorless camera reps are reading the blog to check on the leaks. So he had, in a way, sent it to them 🙂
was just curios what they officialy reply to such idea …normaly they at least reply
That proposal is one step short from square sensor covering the whole, full image circle. If one advocates for a square sensor, IMO it makes more sense to advocate for the full image circle sensor. To capture everything lens projects as.
With current tech IMO it would be too risky and expensive (in relative terms) for manufacturers to justify.
Otherwise, for a top-of-the-line, pro-centric camera this is not a bad idea. One more way to differentiate from competitors.
P.S. Combine that with the bent sensor from Sony, and you might have something truly interesting on your hands.
P.P.S.
> “No need of orientation sensor.”
Not true. It records the orientation of the camera when image was taken. Do you want the camera to display the image the way you saw the scene when you captured it? Apparently not.
> “the mirror assembly (mirror, pentaprism) should be also bigger”
The non-mechanical assembly – SLT, translucent mirror – might be capable of pulling it though.
“No need of orientation sensor – not true, it records the orientation of the camera when image was taken …”
You forget that the camera will always be in landscape orientation, as there will no longer be a need to move it 90 degrees into a portrait position.
Why do you think people only rotate camera for landscape or portrait orientation?
There are plenty of angles between the 0, 90, 180 and 270.
Round pixels are not yet invented 😉
Cameras don’t record the random angle you may be holding them at (or none that I know of). You only get the ‘landscape’ or ‘portrait’ mode info recorded.
Another advantage: when mounting lenses made for a larger image circle (such as FF lenses) and shooting raw, you can keep the whole sensor output.
Good job, Zoltan!
This could be THE tech that allows the camera manufacturers to truly break free from conventional ergonomics. If the tech could be worked out in a way that wouldn’t be detrimental to IQ, I’d definitely consider buying it when I’m in the market for another body.
It reminds me of the idea of Maitani of a modular system such as used by conventional MF systems. An idea which I feel would please a great many users.
I actually suggested this very idea in a DPreview forum more than a year ago, complete with hand-drawn and almost identical diagram!
4/4 system? http://www.dpreview.com/forums/post/41007919
P.S. There were so many talks about square sensor – on all forums – that it is impossible to find anything behind the static noise.
The earliest posts discussing this in the Olympus SLR Talk Forum are from 2001.
It is an old idea, but has only really become viable with the shift to EVF.
I would welcome it just for the full resolution of the 1:1 aspect ratio.
My thoughts exactly, sometimes I like to compose thinking on a 1:1 ratio but I end up no doing it because I would need to crop a lot.
It is something that I’ve argued for since we lost the mirror. However I have two points of criticism against this article :
1) Since every penny counts, I’m pretty certain people wouldn’t miss 16:9 in portrait orientation. 3:2 should cover the needs of 99.99% of the population. Therefore a not-so square sensor that goes from 16:9 landscape to 3:2 portrait should be enough, save cost, weight (good for IBIS), reduce the mechanical shutter travel, etc.
2) This implementation is heavily dependent on a well-designed user interface to change formats, otherwise part of its advantage (no lost time changing grip over the camera) is lost. IMHO from a conceptual point of view, something that works is to have several “aspect ratio memories” (call them whichever way you want), which number may be user-defined (you can choose to memorise two, three, four, etc. aspect ratios) depending on how the input is implemented to switch between them (can be either a marked input, in which case their number is fixed, but can be seen when the camera is off or in sleep mode, or a soft input, in which case their number can be user-defined), and then let users change aspect ratios within those memories at will. Something that may work is a sort of four way lever. Going from right to left and vice-versa would switch between aspect ratio memories, going up/down would change the aspect ratio within this memory. There may be other, more clever implementation. But the gist of it is that it should be extremely fast to go from, for example, 16:9 landscape to 4:3 portrait. A button + dial won’t do I think. Also, it might be interesting to have a way to quickly revert back to a “home” or “base” aspect ratio, just as it’s convenient to be able to quickly revery back to a home AF point position.
“we lost the mirror”
??
We have got rid of it 🙂
Good points. I would probably drop the portrait 2:3 too, since we’re on a MFT forum anyway.
There is a trade-off where the higher the aspect ratio you want, the more sensor area is ‘wasted’ — so 3:2 is a more efficient cut-off than 16:9, 2:1 is less efficient, etc.
The UI could be sorted out in various ways — perhaps simply rotating left/right through the selected aspect ratios or holding down a button while doing so to get to ‘intermediate’ settings would work OK. For example, set 3:2, 4:3, and 1:1 as your ‘normal’ AR set (from a standard selection of e.g. ’16:9, 3:2, 4:3, 5:4, 1:1′), and starting from 3:2 — rotating the dial clockwise yields 4:3 -> 1:1 -> 3:2, rotating anti-clockwise yields 2:3 -> 3:4 -> 1:1 (useful redundancy), holding a button and rotating CW yields 4:3 -> 5:4 -> 1:1 -> 16:9 -> 3:2, etc. Add the ability for ‘user selected aspect ratios’ and whether they should be available in landscape or portrait and you’d have a fairly quick way to select the most appropriate framing for what you are shooting.
The biggest advantage is that you don’t have to think about composing within the sensor aspect ratio constraints (or accept losing much more sensor area by in-camera ‘crop’ modes).
The first rule in production is::
YOU WANT TO SELL THE SAME THING TOMORROW
v1 v2 v3 v4 v5 v6…and so on
and then you invent the show and all the marketing and what ever is…foolish.
you find some easy to sell ” points”
and then after 5-6 years you return to this point.
that is it.
Technical:: do you want a plane verry fast…a plane with long range..same 100 liters kerosine?
1:1 would be a big benefit.
but….( timemechine on)
YOU WANT TO…
but what would be a ROUND SENSOR made of a FLOID?
max lowlight
no pixel we know
the floid always makes the crop to zerro.
absolute and adequade to future and come back in line we startet
long ago ( we live now in the geizhals-liky-like-lowtec-world = the
most income for producers ever…. mouuuuuuuuuuuuuuuuuuuuuuh…[ sheeps]…:)
What’s with this site’s Admin ? now he even helps trolls publish their own articles ? This guy was SPAMMING the comment section for weeks with this square sensor nonsense and he’s awarded a whole article now ? I guess next week we’ll see a full featured artcile on why FF is the best and m43 are just toys by FF Troll Nº675
Seriously – I’m really disappointed in Admin. If this is how he deals with spammers and off-topic posters, I’ll be finding someone else to click to Amazon through.
No more of my money for 43rumors — you’ve been added to my adblock list.
What? you’re on a 43 Rumours site! If someone puts together an interesting piece that provokes conversation about the future of the format, Then why shouldn’t it be posted? It’s not like admin forgets to tell us all and any info he can get on camera development and release etc.
I think your weird for getting so angry. And I say this having already posted a mildly scathing post about my lack of interest in the idea suggested here. But I’ve got no problem with it being on the website. (had a bit of a problem with the way the guy started spamming the posts section with this info, but not too bothered really, just skipped over it)
Keep calm and carry on.
The topic is not the issue here, discussion about the format is always a welcome break from all the trolling that this site gets daily, the issue is with the way a notoriously annoying spammer gets awarded by Admin. Sure, someone could post the theoretical cure for cancer on 43rumors comments section and that would wonderful, but if that person posted that same post verbatim in the most crude cut&paste manner possible on and on and on repeatedly on every single news post and article on this website, as awesome as that info is the guy is still a spammer.
What worries me most is the attitude Admin has taken with the site in the last few months, I used to respect his effort and care for this site but the lack of moderation and the rampant trolling is a HUGE disappointment, this site has turned into a playground for trolls and Admin keeps doing nothing except profit from the traffic this generates.
Negative, a spammer is one who primarily posts w/o revealing its real identity and makes this thing for an obvious commercial purpose.
You can imagine that I’m just an ordinary user exactly as you also are, my identity is really easy to be found out and posted this so-called spam just 4 (four) times, one bigger post and 3 smaller and identical ones (no way harassing for weeks and so on). That’s only because I wanted people to be also ‘prepared’ for discussing the subject until Andmin has finally posted the article.
To all you stupid kids: you don’t like it? Don’t read it. It is not acceptable to abuse whoever writes things you dislike.
PS being on a dead format will get you more things you don’t like.
@kesztió
Obnoxiously repeating the same post was bad and annoying just by itself but now you’re telling us you had already sent the article to Admin beforehand ? You needed to ¨prepare¨ people for the discussion ? what kind of nonsense is that ? Couldn’t you just have waited instead of derailing all those numerous threads for no reason? Call it what you want, spamming, trolling, it doesn’t matter, it wasn’t the right way to go about it. You’re a regular poster I know that, but dude, check that attitude of yours, your repeated posts were unnecessary and did not contribute to the threads they were posted in. In fact, your annoying¨preparatory¨ postings did exactly the opposite and just created more critics towards your ideas.
@ToleranceAndCivilisation
Oh, the irony…Anyway, in case you didn’t notice, the poster of this article didn’t limit himself to just one post, he derailed numerous previous unrelated threads for the sake of shoving his ideas into people heads resulting more in an imposed monologue than his so-called future ¨debate¨ .
@Admin, if what kesztió said is true, then I’m utterly disappointed in your role in all of this. If you were preparing a full featured blog post on this subject, why didn’t you tell kesztió to stop with the non-stop derailing ? Or was this extended blog post actually the fruit of all that labor ? Is this attitude a guaranteed method to get a featured post on this site now ? what’s next, FF trolls articles on the front page ? Seriously, how much more ground are you going to lend to trolls on your site ? this use to be a community-oriented news site, it was an oasis for m43 users now it’s just a troll’s nest.
Its an option that may be suitable for use by many who visit here.
You have your say. You don’t like it OK.
Some may do. Let them see the idea for themselves.
Clear?
This proposal solves nothing. Why should I pay for sensor pixels that I never use? Just so I don’t have to rotate my camera? With a rectangular sensor, I use ALL the pixels, ALL the time and there is absolutely NO user interface issue. I ALWAYS know what orientation I am in because it’s controlled by HOW I HOLD THE CAMERA. Nothing could be easier!
This is a stupid idea.
As I understand you are against conventional multi aspect ratio sensors as well since there are also some pixels there you never use.
It looks like lots of people agree with you in the Poll. Though I have chuckled when wondering if you have managed to vote multiple times? I don’t even know if this is possible, but you seem keen enough to have probably tried. hehe . .
Many thanks for making me a cheater.
With “convential multi aspect ratio sensors” I have the ABILITY of using all the pixels I have paid for and that is, in fact, what I do 99.9% of the time. With your proposal I have NO ABILILTY of using all the pixels I have paid for. Further, I gain NOTHING for this penalty.
Face it: Your idea is stupid. It’s a solution looking for a problem. 43rumors.com should be ashamed of themselves for posting this nonsense.
This is quite an old idea (I have put it forward myself many times). The advantages, as kesztio above is alluding to, are that you can select the appropriate ratio for what you are shooting — 3:2 or 4:3 is not appropriate for everything, so you’ll end up cropping and ‘wasting’ those pixels that you’re so enraged about in any case.
The area of sensor in the corners wasted depends on where you want to put the maximum ratio cut-off. Set it at 2:1 instead of 16:9 and you’ll have more unused area.
If you have ever tried using on-camera flash in portrait mode then you should be aware of a practical issue with rotating the camera other than handling. If you can’t figure out the ‘orientation’ from the EVF or LCD image then you have larger problems.
You’re ignoring my point: With this ridiculous proposal, you are paying for pixels in a square sensor that YOU CAN NEVER USE. With a 2:3 or a 4:3 sensor, you buy a sensor that you CAN use fully. You don’t HAVE TO waste pixels, UNLESS you have a square sensor.
So what is the benefit for buying pixels I can never use? Not having to turn my hand 90 degrees? A slightly less crappy on-camera flash experience in portrait mode?
Give it up, guys. It’s stupid.
Deear @disqus_gJQb8IUosu:disqus ,
1. you probably also wouldn’t pay for a lens whose image circle has portions you never use them, but unfortunately you do as conventional sensors where ‘100% of pixels are used every time’ use just a relatively small amount of the light the lens captures.
2. Why you don’t bother about cropping (dropping) a huge amount of pixels from 4:3 when you shot e.g. 1:1, instead of having native 1:1 ratio sensor?
I thought I always buy camera with all it’s features. I guess I was wrong I was always buying just pixels.
exapple, I hope you use all of the features that your camera has. If not you just wasted your money.
As MAR sensors always have pixels outside the image circle you indeed have unusable pixels e.g. on GH2. MOREOVER, my G5 has (a just slightly enhanced) GH2 sensor, with 16:9 disabled so I really have lots of unusable pixels in my camera. Hence I’ll drop it as I’ve just found what a bad thing it is.
Congratulations, you haven’t undestood anything.
The mirrorless system can be cheaper than DSLR – but it is not yet. It may offer higher quality images (smaller flange distance) – but it does not yet. It can have a square sensor – but it does not yet. Boooring!
Thanks, Admin! 🙂
Ban this moron from posting here.
I don’t understand, but GH2 has already a muti-aspect senor, isn’t it? I don’t see any value to have it squared.
Lower crop factor when using manual lens!!
Oh god, not this again. Who are these people that have such issues rotating the camera?
I guess they all have arthritis.
Plus the EVFs and screens would all have to be square unless you like looking at tiny images when shooting portrait or square with the camera in a landscape orientation.
It’s not where I’d spend the $ if I were Pana or Oly.
@boggisthecat:disqus figured another noticeable advantage as either in-camera flash or on-camera TTL flashes poses big problems when rotating the camera, e.g. bouncing. Not speaking that it’s not really confortable to keep a camera rotated with a bigger system flash mounted on the hotshoe.
It’s an excellent idea, as a matter of fact I thought of it years ago. Just consider the advantages when using a flash or a tripod. Of course the viewfinder must show the chosen aspect. We need square monitors and projectors with a square image too, to view or show our landscape and portrait pictures with the same size.
Well, guys, it‘s the hight time to apologise, if you do consider my previous post really as harrastment.
So, please forgive me for these just 4 (four) previous ‘spams’ which had the only purpose to have the subject prepared for discussion by the time admin posted my article. And believe me, I’m just an ordinary user like many of you, have no commercial gain for bringing this subject in focus.
I just wanted to pay this subject the attention it deserves. Sorry if I made this thing in an inappropriate way.
Olympus had a project of putting in a square sensor into cameras. It was published in 2008. I read it on dpreview – they called it ‘whole frame’ sensor.
Not sure this will catch on any time soon, but maybe? Especially as sensors get cheaper and cheaper, then it might very well make sense.
Multiaspect sensor was already present in GH2 and LX7, although in a bit more rudimentary and cheaper form. The cost is probably the reason why it didn’t caught on. Square sensor is technically easier to make but it has a larger area – and the cost is determined mainly by silicon wafer utilization. I’m sure manufacturers contemplated this idea for like a decade.
I don’t think this is usefull enough compared to the efford.
A multi aspect sensor like in the GH2 compared to the GH4 gives the following results:
4:3 4.608 x 3.456 (no gain)
3:2 4.752 x 3.168 / 4.608 x 3.072 (gain: 6.3%)
16:9 4.976 x 2.880 / 4.608 x 2.592 (gain: 20%)
The multi aspect sensor is 8% larger (17.2 MP) and gives 20% more area on 16:9.
The proposed sensor would have a size of 4976×4976 (24.76 MP) which is 55.5% larger. The only advantage over the multi aspect sensor would be, that you don’t have to rotate the camera and a little bit area between 4:3 and 1:1.
Instead of building such a monster, it would be much better to think about the multi aspect sensor which is only 8% larger than the plain 4:3 sensor and delivers almost the same final result.
I like the the author’s name. Way nicer than his screen name. 🙂
Why not allow all the pixels to be read out at once? The full imaging circle, vignetting and all. Capture all the pixels and then choose the shape during developing.
Of course this would be the best way, moreover if the sensor’s readout speed permitted that (which shouldn’t be an issue with nowadays’ processors).
As I see the market evolving, ppl is aiming at shoot&share products (smartphone -> fb)… Do they really want to specify pic orientation before sharing it? Do they really care about the aspect ratio?
Who’s the market for this? Mass “don’t event care” market or pro “don’t even dare to change my habits”?
// Or… a rotable sensor //
A rotable sensor after we’ve finally got rid of all movable parts (mirror, shutter. etc.)
In principle, I think it would require cameras to be a little taller than with the current ‘wide’ sensors.
Other than that, I have long suggested round sensors — like the image circle. Like you say, the big advantage is shooting portrait or ‘tall’ photos without turning the camera around, and in fact not even having to make the orientation decision at the time of shooting.
If using pure electronic shutter the camera has to be no taller at all. Just think about it, the sensor should fit in the mount circle’s boundary rectangle (even for extreme implementations like your suggestion about round sensors) and the camera should be anyway and obviously bigger than the mount boundary itself.