« Open Mike: Smiling About the Weather | Main | Circles »

Sunday, 08 January 2012

Comments

Pretty strange idea there Mike, SQUARE ........it's kinda "out there".

You might want to run that by the guys in marketing, and while your at it, check with the guys in Legal as well !!

I do love me some square. But I think I'd be happy if we just move away from 3:2 and onto 4:3, 5:4 or 1:1. Preferably one of the last two.

If this square sensor camera also had a square tiltable LCD, with an optional slide on/slide off collapsible hood,then it would be a really great "composing on the screen" tripod instrument as well as using it with the EVF. The camera firmware should also allow for easy selection between square, portrait, or landscape formats, with a range of aspect ratios for the latter two, for either the LCD screen or the EVF.

Make it a round sensor and you're making the most of your lenses.

Actually it would obsolete all 35mm lenses, you'd either have to reduce the square to mee the image circle or make larger slower more expensive lenses.

I kind of like square for the reason given but alas the digital era giveth and taketh away.

bill

See the post on the GH2. Some of us already enjoy square (and b&w, for that matter) preview and square RAW files. :)

I have thought this since Panasonic released the multi-aspect sensor in the GH2. Why not go all the way with a square sensor?
The only down side would be the extra cost of the larger area sensor.
I still like shooting portraits in the 1:1 format, but the ability to go from vertical to horizontal with the push of a button just makes sense.

Why not round sensors - the optics are round!

My EP-1 has a setting for square images, but not vertical/horizontal. I think of it as a really poor mans digital medium format camera.

I completely agree with this.
I always found 3:2 to be a bit too "rectangular". I prefer the aspect ratio of the 4:3.

That's why I love shooting with medium format. You can choose between a 4:3 format (645), a square format (6x6) and the 6x7 format.

One of the things that I appreciate in the 4/3 system is, well, the 4/3.


The thing about these rectangular formats is that it is easier to compose an image, so begginers usually like these formats.

Still in the 3:2, another option is the possibility to print an image in a A4, A3 or A2 paper without cropping it, but that's an entirely different question

I'm all for it. Have been for years. But today I think I'll see your square and raise you a circle.

Why not a circular sensor?

A circular sensor has two obvious advantages over a square one: 1) when you wanted to shoot rectangles, you'd be gaining some pixels in the long dimension, instead of only losing pixels in the short dimension; and 2) you could orient your square (or rectangle) at any angle without giving up pixels. (Or shoot the full circle and crop later.)

#2 may sound frivolous at first blush, but think about it. Cameras that don't have to be reoriented could evolve radically in terms of form factor and ergonomics, possibly to the point that they would be awkward to tilt to any appreciable angle. It would also come in very handy for precision tripod work, especially in rough terrain. In the case of shooting full circle and cropping later, it would allow correction of camera tilt with no loss of pixels.

So, to sum up: compared to the square, the circular sensor makes even more sense in terms of lens coverage, and is even more versatile in terms of orientation.

Whaddaya say?

I've wondered about this for some time, except -- unless I'm overlooking something -- I think I'd go a step further. Why not make a circular sensor and see everything the lens sees? Buy having the largest possible canvas at any given focal length, sports and news photographers would, I imagine, occasionally get a shot they wouldn't have caught otherwise. At the other end of the spectrum, affordable circle-format cameras might become a "thing" among the fun-loving, experimental types. It's easy to imagine lots of circle-format Flickr groups, for example. I would especially like some wide-angle photos shot that way.

Or perhaps there is a compromise that keeps us from buying sensor-corner real estate we would rarely use: Sensors in the shape of fat plus signs -- A vertical rectangle and a horizontal rectangle. You could could choose the orientation before you shoot, or shoot now and orient later.

Oh, hell yeah. These days I find myself cropping to square more often than not. But is anybody listening?

Hmm I get the feeling it might raise another issue: the permanent glass half-empty feeling. Cropping is, for some, a hard decision to make because it affects fov and you get that feeling you didn't use those 18 MP fully. Everybody crops but here it would systematize how often one has to crop. Panasonic has somehow 'fixed' this with their multi aspect trick, hiding that decision process in presets conforming to the mounted lens fov.
As far as I'm concerned, I'm not bothered and I can only agree with your point.

The key is choosing an optimal square sensor size. The image circle of 135 is ~43mm, but the image circle for a 36x36 sensor would be ~51mm, which rules out everything but MF lenses.

On the other hand, a 30x30mm sensor could use FF lenses, and would certainly be an attractive proposition...

As an old 6x6 TLR & SLR user I would love to have a square sensor.

Give me the whole sensor to use all of the time. Then provide menu choices for electronic viewfinder grid lines that the user could program for common aspect ratios. Then have an easily accessible function button that would quickly switch these grid lines to either portrait or landscape orientation while retaining the image from the whole sensor in the viewfinder. That way the photographer could easily try different compositions with just the flick of a switch.

Actually, there is one. the Panasonic GH2 uses a multi aspect sensor that has a native 1:1 setting. It's not a crop of 4:3 as the sensor is designed to use as much of the image circle as available in each of the selectable aspect ratios.

It's a brilliant idea. Have a sensor that can shoot more than one aspect ratio and maximize tha image circle in any of them. It's a shame that no other manufacturer thinks its a good idea.

Gordon

Hi,
well, the only sensor shape that can give you flexibility as far as the aspect ratio is concerned, is round, not squared :-)
Having a fixed aspect ratio simply saves cost, since you surely get more rectangular chips on a (round wafer) that round ones.
With one exception, of course: if the sensor covers the whole wafer. In that case, round is cheapest (but not in absolute figures, I am afraid :-).
Regards, Jan

I recall that when the Pentax K-7 was about to be launched a lot rumors mentioned that it would have a square sensor. I was shooting 2:3 on a K10d at the time, and the idea of a a square sensor made a lot sense to me then.

But for a while now I've been framing my photos almost exclusively as 1:1 or 16:9, using a 4:3 EVF camera (Oly E-P3). This sensor geometry falls nicely in between those two ratios. If I were using a square sensor, then I certainly would feel too many pixels were "wasted" when framing 16:9.

Square would be more logical, I agree. And I still think the type of viewfinder like the here depicted Hassy would work magic with LCD viewfinders as well. placed on the back, not the top, with folding shades and a fold out loupe inside, they would solve many problems people have with LCD finders...

My GH2 shoots 4x3,2x3,1x1 and 16x9. I tend to use all but 1x1.

Why no square sensors? Four reasons: Visual perception, tradition, "premature optimization", and electronic viewfinders.

Visual perception biases people to wider formats because we can move our eyes more readily side to side so the "image" our brain composes seems to be wider than high. Once we had a choice to make eyeglasses "not round" we made them rectangular.

Tradition plays on this to generate the current popular aspect ratios. The most popular 3:2 so called 35mm "full frame" on a still camera comes from dividing in two (and rotating by 90 degrees) 4:3 cinematographic 35mm film. That 3:2 aspect ratio also matched nicely with 6x9 which was popular at the time when Leica was setting the standard for 35mm. 4:3 returned when Olympus went to "half frame" and divided still camera 35mm in two in the 1950s.

These two aspect ratios still dominate with the video and cinema driven widescreen 16:9 and 2.35:1 aspect ratios haven't made an impact on still photography..

They're all arbitrary choices but people liked to stick closer to what people already knew for marketing and aesthetic reasons.

The final two technical reasons may be on the verge of disappearing.

The first of the technical reasons is optimization: you paid a lot of money for those pixels and by heck you are going to use them in all the images you capture. People like to optimize the expensive pixels they have to match the sort of photos they want to shoot. So they stick with 3:2 and 4:3 aspect ratios and shoot portrait or landscape orientation by turning the camera.

As more pixels become available this starts to look like a premature optimization: if you have pixels to spare you can defer the choice to later. Once we get to the point the customer thinks they have more than enough pixels for cropping later then they might accept a 1:1. Then why fiddle with rotating the camera too?

When might that be? If you use a 3:2 sensor and think 12Mpx is enough (i.e. enough for printing A3 at 250dpi) then a 18Mpx 1:1 sensor should be exactly the same resolution (after a 3:2 crop in either orientation). If you use a 4:3 12Mpx then a 16Mpx 1:1 sensor should suffice. That time doesn't seem to far away with the next generation of cameras are sporting 16Mpx and 24Mpx (and up) sensors.

The second of these technical reasons is electronic viewfinders. With a square SLR design the prism has to get larger to accommodate the bigger vertical field of view. So the mount gets larger and everything gets heavier. The finder would either need to provide optical masking to select the aspect ratio of the view. Or finder lines to indicate the many different aspect ratios you could use. The one upside of that is "overscan".

But if you are using an EVF it's very easy to crop the image so what you see is what you get in whatever aspect ratio you choose (and RAW would always be 1:1 for cropping later). There's no extra weight penalty for the prism.

It shouldn't make much difference to the manufacturer of sensors so long as they can make a sensor with a single step of the mask stepper. APS-H, 29 x 19 mm, designed is the biggest that's accommodated with standard steppers (one reason why "full frame" sensors are expensive). So if you can keep the sensor below 28mm on a side you should be able to make square sensors in a single step. APS-C derived square sensors (24 x 24 mm) should be possible on current equipment and would match 1.5x crop lenses. They would be a bit more expensive than current APS-C sensors (perhaps up to twice the price?) given a lower yield.

The other cheaper route might be to "square up" FourThirds sensors to 17mm x 17mm giving a 2x crop factor option. That would be an interesting evolution for Panasonic to follow if APS-C starts to win out in "the sensor wars" and a natural follow on from what Panasonic is doing with their almost "constant diagonal FOV" even when the aspect ratio changes.

But square "full frame" 36 x 36 mm sensors would be even more expensive than current "full frame" sensors (even lower yield per wafer). That would be a business reason not to push this.

So what does that point to? The future square format camera will be mirrorless with an EVF and an "square" APS-C (or smaller) sensor.

I feel the need to express my disagreement with the proposition. Although I understand and agree with its intent, the unfortunate fact is, the sensor, especially for the larger sizes, is an expensive part of the camera. Having a square sensor would for most photographers, leave a lot of unused pixels, as the desired format is generally a rectangle. So a square sensor adds cost in a highly price-competitive market.

Why not panoramic sensors?
After the new Fuji anouncement, I see the digital X-Pan closer...

One problem with a square (or circular) sensor is that most existing lenses incorporate rectangular 'light blockers' in their physical design to stop non image forming light from bouncing around and degrading the image. Thus the image circle (again for most lenses, not T/S) is not in fact a circle. When the cost/convenience factor of a new range of optics is factored in it is no surprise that the square sensor idea hasn't taken off.

At least one new camera uses a square format sensor, as I recall: the Lytro.

That ancient square-sensor camera has another aspect that I sometimes miss. It is used at a vertical position substantially below that of any eye-level-viewing camera. The difference in point of view can be rewarding.

4:3
3:2
16:9
1:1

The Panasonic GF1 allows all of these, of course - although the sensor itself is not 1:1

Timely topic that I've pondered recently as I worked with my Rolleiflex and considered a minty Hassy 503CW. I wonder how a 6x6 digital TLR would sell if priced around $15,000-$18,000?

Pure fantasy, of course.

Re. round sensors -- The optimal solution is obviously a hexagonal sensor, since it uses the wafer area just as efficiently as a rectangular sensor does.

H'mm. I was surprised to see that I'm not the first one to think about a round sensor.

Most camera formats have been some sort of rectangle, whether or not all the sides are the same length, since photography began. It's been a practical thing, to do with the making of cameras and so that pictures can be made without wasting materials. Imagine buying a box of 100 sheets of 10" diameter printing paper, and cropping it. Rectangular formats are accepted almost without question.

However, if your photos are only ever seen on a monitor screen then it doesn't matter if they are round. Or elliptical.

I would love a square sensor. As for all the people mentioning that you would toss pixels away because you need to crop, well that happens to me with my Nikon on most shots. I don't really like the 3:2 aspect ratio, so I crop it down. Maybe a compromise of a 4x5 sensor would be ideal. For me it would result in the least cropping averaged per image.

The requirement is for the diagonal of the 1:1 sensor to match the diagonal of the existing 3:2 or 4:3 i.e. the sensors have to have the same image circle as pre-exisiting lenses to make sense.

So

a "full frame" 35mm image circle needs a 30.6mm square sensor.

APS-C image circle needs a 21.3mm square sensor.

FourThirds image circle needs a 15.9mm square sensor.

Each of these is (of course) smaller in one direction and large in the other direction than the original 3:2 or 4:3 sensor but this results in an overall larger area than the original sensor e.g. the 35mm 3:2 is 36 x 24 == 864 mm^2 but the equivalent image circle 30.6mm square sensor is 8% larger at 936mm^2.

That 30.6mm square sensor is not much bigger than APS-H longest dimension and might be capable of being made in a single step. That might be an interesting development.

Of course this all makes a mess of peoples "feel" for their crop factor (i.e. these square sensors would have have the same crop factor as their related sensors but somehow it "feels" different).

I messed this up in my previous coment!

Kevin Purcell: cinema driven widescreen 16:9 and 2.35:1 aspect ratios haven't made an impact on still photography..

I know it's only anecdotal, but I, for one, have started favouring 16:9 for a lot of my wide(ish) landscape work. (I'm another of these pesky GH2 users that seem to have popped-up on here of late, and I seem to be composing to a standard ratio these days, whichever it may be per circumstance.)

I love square (a lot of my "art/nature/stuff" uses it), truly hate 4:3 (so *FAT*!), tolerate 3:2 (it feels like a generic lingua franca) and (inadvertently) treat 5x4 as a safety margin over something visualised as square.

I gather 16:9 is a fair approximation to how one's eyes see; having the LCD previewing that way makes composing landscapes feel very cinema-esque (unsurprising, but pleasant nonetheless).

Ramón Acosta: if your lens's coverage was just enough to circle the square (as happens in 6x6 MF) then actually what you gain is shift movements relative to 645. There's more to life than megapixels.

Now, for those shooting with magazine covers in mind, A4 size is closest to 5x7. :)

I'm fine with a rectangular format as long as I can have a square LCD monitor for photoshopping and viewing pictures. I hate, hate, the fact that my vertical pictures need to be viewed at a lower resolution on a monitor than my horizontal compositions.

Alas, even 5:4 monitors, so common a few years ago, are nearly an extinct species now.

As a long time Hasselblad user and lover, I love the square. BUT, I gotta say, almost anything I ever shot with the square was cropped by the art director for oblong uses. I used the 'blad, because it was smaller and lighter than the Mamiya 6X7, which was important on location, but if I went all the way to 645 (which was even lighter and smaller than the 'blad), if I had to crop to a square use, I'd be cropping to 1 5/8th's by 1 5/8th's, which would be unacceptable. I'd like to see a Rollei (or something cheaper like a new Yashica) twin lens, that was a stand alone square digital. They could make it small, with a 36mm by 36mm square sensor, like the old digital back for the 'blad, and price it afforably!

Hexagonal sensors should be the optimal compromise between the production wafer coverage and the circular projection given by the current lenses. Just a thought. Crop as you wish. Remember, nature is hexagonal for a reason...

I've been asking for that for years. I wrote Olympus and Canon about it. Olympus replied that my suggestion had been forwarded to their engineers. Canon ignored me. Years later when 4/3rds came along camera manufacturers started including the option to crop a square image from the rectangle in camera but only in JPG. Please, make a square sensor 16-20 MP. Put it in a camera with basic camera functions. I don't need "face detection", panoramas made by holding down the shutter release while 'sweeping' the scene, etc., etc. I know how to do the stuff those features do. I don't need a camera that does it for me.

Pro DSLR's will always be 3:2. This is due to the shape of the human body (DSLR in vertical orientation). The need for fashion editors and for US Weekly to analyze the shoes of a prospective starlet on the red carpet trumps the need for a square sensor, where the subject might not be isolated.

It's simple economics. Teenage girls buy magazines, and shoes.

2/3 of the medium-format cameras I have owned have had square sensors. I never liked that about them. I felt like I was wasting a big chunk of the film I ran through them. And if they influenced me to squarer compositions, then I ended up wasting paper in the darkroom instead, and making smaller prints (since the paper sheet sizes I started from were all various rectangles).

6x6 is not actually "nearly 6x7". 6x6 is an expensive way to shoot 6x4.5.

(Your arguments on camera holding and complexity, to which you could have added rotating backs on the RB-67 and many view cameras, are spot-on.)

Sensor size is expensive like film was, so I'm not sure I'm ready to pay for a 43 x 43mm sensor to most often use only 24x36mm of it.

But that would be so... square.

There is a reason why movies, TV screens, books and magazines, and most of the great historical paintings in museums are not square. Square is boring dynamically. It is symmetrical and static, unless one throws part of it away.

When I used a 2 1/4 square 40 years ago, the extra real estate on the negative was an improvement over 35mm for cropping and making a larger print. Similar technical or optical advantages for lens coverage on a digital sensor. Yada yada. But, when taking pictures, I felt constrained, a little suffocated by the square format. Just no joy for me. YMMV, but I think I am not alone on this.

Thanks, but no thanks.

I'm not quite old enough to remember this myself, but I seem to remember hearing or reading somewhere that those original Kodak cameras produced round images. Didn't want to waste any of the available pixels, I guess.

And the first Kodak camera took a round picture 2 1/2 in diameter. Reading these comments we don't seem to have got very far!

Dave,
I'm not saying you'd have to SHOOT square. I'm saying the sensors should BE square. And that the rectangles should then be selectable and customizable and switchable according to each person's tastes and needs and preferences.

Mike

Kodak actually makes (made?) square scientific grade CCDs. This camera is available with monochrome or color versions:

http://www.ccd.com/pdf/U4000.pdf

A bit tough to use if you don't have a computer though. No shutter release onboard.

Trouble is if you are not going to end up with a square picture, you have wasted more photosites than you would have done starting with a rectangular sensor and re-orienting as required.

I love the multi-aspect sensor of the Panasonic GH2, since it gives me a real choice.

It's not often, I use 1:1, but 4:3 is frequently used, and my preferred setting is 16:9 (which gives an extra wide view compared to only "sensor cropping", as used in most other cameras). In some situations - mainly the historic quarters of i.e. italian citys - I even use 16:9 in the vertical direction in combination with a 9-18mm Olympus M.Zuiko lens. This usually gives me the "extra height" needed to capture a church tower, a special building or other such thing, whith my back solidly planted against some house wall opposite the target.

I've even used 16:9 vertically to good (dramatic) effect with my LX3 in the 24mm (eqv.) position in New York for some night images.

A circular sensor with circular RAW could be all things to everybody. Whether we tend to go in the 16:9 direction, 9:16 direction or 1:1 or whatever...

Kurt Friis Hansen

Regardless of whether a physical rotation is needed to achieve a different photo orientation, I think this would still be my preferred way of doing it. Imagine having to dive into a menu to choose landscape or portrait mode!

I assumed that Mike thought of a square sensor that covers more than the image circle, for example a 36mmx36mm sensor for a full frame camera. So if some suggest a circular sensor, this is just an implementation detail. A square sensor would be simpler, which would be less expensive I do not know.

And it was also my first thought that the problem is that such a sensor would be more expensive. On the other hand, if you consider the price difference between a D700 and a D3, then the price of the sensor would not be such a problem for a pro camera.

One variant of this idea would be easy to implement in some current cameras: When using a Nikon FX camera with a DX lens the image is cropped anyway, and the sensor is exactly the right size to allow for a vertical 3:2-crop.

I believe there is a conspiracy to promote rectangular format :-)

Seriously though, there's probably quite a bit of marketing involved, at least revolving around the idea that the image will always be cropped to a 4:3 rectangle and then those pixels will be wasted. In reality, whether cropping is done or not and how is not something to determine when buying a camera. Also, only small format users are paranoid about wasting pixels anyway. For an example, when I shoot portraits, I rarely find 3:2 to be the optimal format, so I crop. The results count more that losing a few pixels.

At this point it's probably obvious that I like 6x6 very much, although in my defense, most of my shooting is with rectangular formats.

And there you have it: http://www.dpreview.com/news/2012/01/09/CanonG1X_Preview - image ratios: 1:1, 5:4, 4:3, 3:2, 16:9.

Kevin Purcell has most of it right. To his comments I'd add:

1. A square format makes optimizing bandwidth of moving data off the sensor slightly problematic.

2. The notion that all still cameras must also do video means that #1 is important and you will have a wide crop pixel subset.

A better question would be this: why aren't all cameras 16:9 aspect ratio? First, there's that side-to-side parsing thing Kevin mentions, but second, virtually all electronic display (and we're moving to electronic display) is 16:9.

If you shoot a square format, you will unconsciously compose your shot to fill the square space. You'll lose your option to crop later.

As an ex Mamiya RZ67 user, I wholeheartedly agree with the proposition of a square sensor - not that I've ever wanted or needed a square image that I can remember, but for the option of selecting a vertical or horizontal rectangle at the flick of a switch, just like in the good old days of the RZ.
Bearing in mind that 99% of all my RZ shots were tripod mounted, and never the need to drop the camera over on its side, or remount it with an L bracket etc, just flick a switch, rotate the back and voila! Literally a one-second job, and the camera remains aligned with the subject.
I've got an L bracket on my DSLR but OH! how I miss the rotating back solution.
The addition of different shape rectangles, 3:2, 4:3, 6:7, 16:9 both horizontal or vertical would give even greater joy.

The TV camera I use everyday can select either 16x9 or 4x3 output and can scale the resolution to suit your needs.
I believe the 39mp Hassy back allows you to select 4x3 or 1x1 aspect ratio.

Bad idea. Might as well as "why not circular?"

How about a big round one? Square is well, just square.

For those who miss a rotating back solution, Leaf actually has an answer with the Aptus-II 10R. It's a 56MP back that lets you rotate it between horizontal and vertical.

Cost more than the house I grew up in, or any car I've ever owned, but there it is.

I'm making do with an RB-67 and film, at least until my treasure fleet comes in.

Since relatively few images are displayed with square proportions it seems fairly obvious why square format sensors aren't in common use. What seems more strange is that when camera manufacturers made the transition from film to digital, most didn't take the opportunity to ditch the over-long 2 x 3 format. Relatively few images are shown in those proportions either, and the 3 x 4 format common to compacts and micro four thirds cameras would have made much more sense as this is nearest to how most images are reproduced in magazines and other printed media.

As I understand it, the other advantage of a big, heavy body is that it helps to counterbalance the large, heavy telephoto lenses used in professional sports and wildlife shooting, leading to less camera-induced vibration when using them on a monopod or other less-than-perfectly stable platform. However, as someone who has never shot this kind of camera (or lens, for that matter), I can't personally attest to this being accurate.

As far as the square sensor is concerned, makes sense to me provided you have enough megapixels on offer that throwing a significant number of them away doesn't matter for your intended application.

Dicomed made a 2 inch sensor on a capture back for the hasselblad if I remember right? back in the mid 90s--the Bigshot. they had a one shot and three shot (higher res. used an LCD filter for separate RGB exposures) version. $55,000 for the back only. they had a scanning back as well-it was 35K. think Phase 1 wound up buying them out eventually. could be wrong about that though.

After I got my firrst Hasselblad, I became amazed at the number of things I was printing square. I came to quite like it and I miss it.
But what I really really really want is a FF digital back for my Xpan.

Looking at that Hassy, I feel more than a pang of guilt, of remorse. I am fundamentally a 3:2 shooter, but when I shoot portraits I want to work within the square. Selling my V system is the biggest mistake I've made since being assimilated by the digi-borg. Square sensor or 120 film, I think I need to pick up some Hassy gear again, just for the pure aesthetic pleasure of working with the camera and screen.

"Why not circular?"

Because that leaves wasted space as you take out pieces of a silicon wafer to create the imaging chip. As others have pointed out, a hexagon is the best approximation of a circle that doesn't leave any open space.

Personally, I wish I could use a camera with a 16x9 sensor, but I suspect that I'll either keep stitching digital images or perhaps try out a 4x5 with a 6x12 film back. But I'm probably in the minority...

When I wrote a TOP article analyzing preferred aspect ratios, I looked at 72 paintings (not photographs, for reasons explained in the article.) Of the 72, 47 were horizontal, 21 vertical and only four were square. In other words, 68 were some sort of rectangle, and only four were square. That's because (I believe) we experience our vision as horizontal -- the major exception to that is our continuous inspection of other people, which is why we call vertical pictures a "portrait" format.

I think a square sensor would wind up being like a Leica camera or a Fuji X-10 -- you'd have a small group of people who are wildly enthusiastic about it, but most people would opt for the rectangle.

Bugger "wasted" pixels. Every time you save a raw file to JPG, do you know how may innocent bits/bytes of information you ruthlessly MURDER?

Jeez, complaining about wasted pixels is like harpooning pregnant whales to save the fricken krill, I swear. You gotta break eggs to make an omlette, and yeah, you sometimes that to crop to improve an image.

A square ratio sensor would be a great way to make 'togs think outside the rectangular box, IMO.

Number of pixels does NOT mean better art, and all too often it means bigger bad art.

Erlichda, boys.
Edie

oh, for a square sensor that fit the box of an existing camera, and only cost about 20 bucks, remove to download, show to a screen on the back, or on top in the case of a TLR Rollei or Mamiyaflex... that WOULD be nice...

I kind of thought you were joking with the "why not?", but your subequent comment seems to indicate otherwise. Okay then. I think several posters have provided good reasons why not (square): because the display media aren't square, basically, which entails inefficiency and waste at various points in the chain; and also because square sensors would not in fact permit optimal use of our lenses imaging circles. That is, the sensor would need to be significantly shorter than the long dimension of current frames to be used with a given lens (that's assuming the lens hasn't been baffled). So, i'd once again lose my fast wide lenses--and i had enough of that the first time round with digital. I honestly don't see what the big deal is about rotating the camera as needed. And be serious; those dual grips on flagship battleships aren't the reason for the bloat, they at most are an excuse for it. You have seen the batteries that go into those things, haven't you?

At a general level, i absolutely hope that a variety of square format cameras come to market, and are devoured by their fans. I won't buy one, though. I personally just don't relate to square photos. I've seen superb examples by other photographers, of course, but i just don't like to shoot that way. In fact, the squat 4/3 format is a major reason i don't own a micro 4/3 camera--i intensely dislike that ratio. (actually, i don't hate it for vertical orientation, but almost all my photos are horizontal.) that sweet little 20/1.7 is actually closer to a 50 in coverage than to a 40, once you crop to 3:2.

My own predilection for the classic 35mm frame, uncropped, is no doubt mainly a matter of habit and tradition. But that is part of the point: photography relates to a distinct, unnatural, mechanical way of seeing, and i like to see its traditions and points of references in photographs. Sure, some great photos are in other ratios, but the photos which i would argue are the most iconic and "photographic" in our world are presented as they were shot, in 3:2. Consider a ken burns zoom/pan over a photo: the very intent, as it fills your screen, is to erode its status as a photo, and to make of it instead a scene. Only if you zoom back out to see the frame of the photograph inset into your viewing screen does the picture once again take on the appearance of a photo.

Anyway, i truly am happy for anyone who likes other ratios, but i thought the good ol' barnack frame deserved a little support, too. (incidently, one of my favorite things about the s2 is that, unlike most medium format, it shoots 3:2.)

Not having square sensors is a simple answer. They cost more. Lots more. It has to do with the number of error free sensors that can be taken from a silicon disk. A 35mm sensor costs about 10 times what an APS sensor costs even though it is only physically about twice the size.

Lots of cameras now provide a square option in the formats which throws away pixels. Where are the square pictures? I like square too, but obviously not many do and the camera industry will add anything to gain an extra customer without spending any money. Square is too specialized, as is black and white.

I like "golden mean" 2x3 format too. 4x3, not so much. And why would anyone like that old CRT format? Boring.

For my personal work I prefer to shoot square, or 6x7. For my wedding work it's always 3:2 and I sell a lot of prints for the families to frame.

Guess what ? They never, ever buy frames with a 3:2 ratio and their fave pics are always going in 10x8" mounts, so I'm always cropping to suit and I can confirm that their frame is far more important to them than the crop on the image - cut the feet off, that's OK !

The photo paper industry and the frame makers, sort of ignored the 35mm image for half a century and more. Time to drop the 3:2 format ?

I'm a long term Hasselblad user and cannot bring myself to adopt a 645 digital sensor. A square crop from the H4D-40 is good enough in terms of quality, but it's not the same as using a square ground glass to compose. I should probably be more flexible in my approach, but I do love making pictures with my 203FE and 503 Cx. It's not just the end result, it's also about involvement and craft.

I think the market has spoken, and going back at least to the European renaissance, a rectangle is the prefered format. A common format is classical Chinese and Japanese painting is of course the vertical scroll, even farther from square than most European art.

Personally, I shoot 6x6 (cm) often and like working in the square format. I typically print a 7x7 (inch) area centered on 8x10 paper. Film and wet darkroom.

Unless I'm trying to conserve space on a memory card, I rarely see a need to shoot at anything less than full resolution and crop or scale to whatever format and size I want later. A digital camera option to save square images is little more than a gimmick to me.

from above:
« I would feel like I was wasting all those pixels I paid for. »

this reminds me on the early days of CDs, when the first CD-Singles appeared and there were lists of "CDs that are too short" (in duration), and people not wanting to buy something that filled up the CD to its 74 mins. it is the quality, and intent of the release, that matters, not something that is "wasted."

I would think that a square design would still retain the same MP count, just distributed differently, say in square form, because of the marketing might of the MP number still has a grip on people.

all other matters being equal: MP count fixed, and the pixel size not made smaller to regress into smaller/noisier sensors.

while we are at it... a BW sensor too — yes, though the caveats mentioned in a recent TOP article are sensible.

Many of my final images are cropped 1:1.

So now how about a square monitor so that the portrait modes are the same size as Landscape.

Re: 4/3 and m4/3 standards, IIRC, the original 4/3 spec called for sensors to cover the image circle equivalent of 4/3" video capture tubes. Unfortunately, I can't point to documentation of that factoid. The spec didn't specify a format, just the size of image circle. So a manufacturer could release a 1:1 sensor without changing the spec at all. I'll trust the math from others above that put it at ~16mm.

One thought I had re: the Canon G1X sensor being, essentially, 4/3 sized is that Olympus or Panasonic could just buy those sensors (ok, Olympus, not Panasonic) and drop those into new models. I'd bet Canon's sales volumes will make that camera a better seller than the combined Oly/Pana cameras and thus open the door to more competitive sensors at lower cost. As well, I wonder whether Canon or someone else is making the G1X sensor?

And where I really go down the rabbit hole is with the idea of Olympus just buying APS-C sensors from Sony as-is and electronically neutering them to be 'square' sensors. I'd bet a good sandwich that in volume, Sony's 16mp APS-C sensor would cost less than Panasonic's 16mp 4/3 sensor, and perform better, besides. Even if the Sony cost more, it is my belief that the Sony is sufficiently superior in performance that Olympus could offset a higher price in the bill of materials by putting the sensor only in a high-end model.

Patrick

Several people said things pointing this direction, but it just clicked for me -- a square sensor in a world of rectangular photos (as Crabby said, "almost anything I ever shot with the square was cropped by the art director for oblong uses") gives you an inflated megapixel number.

So that's a good thing, right? A way to inflate the megapixel number that doesn't directly harm image quality?

The comments to this entry are closed.