Is it just tradition? It seems that square would be a better fit for round (cross section) lenses. Round sensors would be even better, but I doubt that I'd enjoy round photos, except planets, moons and flowers...well maybe I would.
Why are sensors rectangular?
- Thread starter chrysoberyl
- Start date
Because a rectangle is a more interesting compositional shape than a square. That is why almost all drawings and paintings are rectangular, too. Round? Consider trying to make round paper or trimming paper to be round to fit the round photo. Not likely.
Upvote
0
Hi Chrysoberyl.
I would suggest it is because our field of view is basically rectangular, try it, fix your view on a point and see where a finger disappears from your view, I think my range is closer to 16:9 than 4:3!
Cheers, Graham.
I would suggest it is because our field of view is basically rectangular, try it, fix your view on a point and see where a finger disappears from your view, I think my range is closer to 16:9 than 4:3!
Cheers, Graham.
Upvote
0
Valvebounce said:Hi Chrysoberyl.
I would suggest it is because our field of view is basically rectangular, try it, fix your view on a point and see where a finger disappears from your view, I think my range is closer to 16:9 than 4:3!
Cheers, Graham.
Thank you, Graham. Although all the answers make sense, yours makes the most. I, also, am closer to 16:9. The 4/3 folks must have close-set eyes...
Please call me John.
John
Upvote
0
Because film came in long strips, was expensive and the rectangle was the optimum shape (as is a square) to minimize waste. More often than not a rectangle fits the scene better than a square - again minimizing waste.
Upvote
0
This is a patterned wafer holding AMD Opteron dies before they are diced, finished, tested, and packaged into the chips you're used to putting into your computer. As you can see, a rectangular shape results in a good packing while a round shape would waste much of the silicon and thus drive up the cost per chip substantially. And rectangular makes the most sense for customer experiences since all other tools in the pipeline (screens, printers, etc...) are similarly optimized for rectangular shapes.
While the idea of a round image sounds fun, it does present some large problems for certain kinds of optics--tilt-shift lenses come to mind as an example since they would have to be about twice their current size. Likewise, there is no practical use or need for such sensors.
A square shape would use the most lens surface area in a practical way, and indeed in 6" x 6" cameras you can obtain that. I wouldn't mind seeing a 30mm x 30mm format sensor from Canon (the largest square fitting into the same size camera as a 35mm format) as I often find compositions in the square pleasing and don't always want to crop to 24mm squared.
Meantime, if you occasionally want a round shape use the crop tool in Photoshop.
While the idea of a round image sounds fun, it does present some large problems for certain kinds of optics--tilt-shift lenses come to mind as an example since they would have to be about twice their current size. Likewise, there is no practical use or need for such sensors.
A square shape would use the most lens surface area in a practical way, and indeed in 6" x 6" cameras you can obtain that. I wouldn't mind seeing a 30mm x 30mm format sensor from Canon (the largest square fitting into the same size camera as a 35mm format) as I often find compositions in the square pleasing and don't always want to crop to 24mm squared.
Meantime, if you occasionally want a round shape use the crop tool in Photoshop.
Attachments
Upvote
0
chrysoberyl said:
Many existing lenses have internal baffles which would cause vignetting if used with a 36x36 sensor, as would some petal hoods of course. But in the case of DSLRs there's a more fundamental reason - you'd need a 50% taller mirror and there is no room in the box.
A square sensor in a hypothetical future full frame mirrorless would be excellent, if only so you could shoot portrait orientation without rotating the camera. Legacy lenses used with an adaptor might or might not vignette but that wouldn't be an obstacle, you'd just use them in the traditional way.
Upvote
0
scottburgess said:
Well-stated, very informative! A 30mm x 30mm (or even 35mm x 35mm) format sensor would be lovely, but I'm sure it is not to be. I would really like it for astro.
John
Upvote
0
scottburgess said:
That's geometrically true, but would have consequences on other aspects of camera performance – e.g. a larger mirror/shutter which would certainly be slower (frame rate and Xsync) and may not be compatible with current flange focal distances.
Also, as stated above the rectangle is more compositionally pleasing than the square (in general), and with a square sensor you'd have a maximum long side of 30.4mm, compared to the 36mm with the current 3:2 rectangle. While the ability to switch from landscape to portrait orientation without loss would be nice, in most cases you'd be throwing away a large number of pixels, assuming you can decide beforehand which orientation you want to use.
Upvote
0
neuroanatomist said:
Thank you. This is starting to sound like something Sony would do, mirrorless and electronic shutter, rather than Canon.
And a benefit would be less coma and vignette.
Upvote
0
tr573 said:
IIRC, m43 corners are supposed to be closer to the center than 3:2 sensors, which improves image quality.
Upvote
0
What has been the most photographed subject in the early times? I am not shure / have no statistics, but I think it was
- people shots (portrait orientation)
- landscape (landscape orientation) : combined view of both eyes (as Valvebounced mentioned)
- larger groups of people (landscape orientation)
Maybe this historic startup of photography has influenced the formats - at least the naming of orientation which is common today.
On the other hand there was a large number of cameras like Rolleiflex, Hasselblad which sported the 6x6 format which makes at least a good starting point for later cropping into portrait or landscape.
About square shaped sensors: I would like it but I would like a square shaped display / beamer too. A long time ago I had seen a square shaped TFT display with 2048x2048 pixels for air traffic control - I really liked that thing and had some ideas of misusing it to view portrait and landscape oriented without size difference.
- people shots (portrait orientation)
- landscape (landscape orientation) : combined view of both eyes (as Valvebounced mentioned)
- larger groups of people (landscape orientation)
Maybe this historic startup of photography has influenced the formats - at least the naming of orientation which is common today.
On the other hand there was a large number of cameras like Rolleiflex, Hasselblad which sported the 6x6 format which makes at least a good starting point for later cropping into portrait or landscape.
About square shaped sensors: I would like it but I would like a square shaped display / beamer too. A long time ago I had seen a square shaped TFT display with 2048x2048 pixels for air traffic control - I really liked that thing and had some ideas of misusing it to view portrait and landscape oriented without size difference.
Upvote
0
Hoo boy, where to start? Artists for hundreds of years have been attracted to the Golden Rectangle, with an aspect ratio of 1.6180:1 (16:9 is 1.777:1). You can find this in all kinds of paintings, and more than a few PhDs have been launched describing the Golden Mean as used compositionally and architecturally in the Renaissance; it was practically a fetish. When photography was standardized, sheet film in the United States fixed on old paper sizes; 8x10 and 4x5 (low aspect ratios of 1.25:1) which are descended from the quarto size used in Shakespeare's time. (8.5x11 paper size is a modern standardization.) Eventually, Eastman introduced all kinds of different film sizes for various Brownie cameras, too numerous to list here, but a crazy-quilt melange of squares and rectangles, eventually settling on 120 in the 30s. Different cameras used this film for different aspect ratios, but a 6cm x 7cm image scaled up almost perfectly to an 8x10 sheet of paper, but a 6x6 image yielded 12 exposures instead of 10. For most hobbyists, and plenty of professionals, the square 6x6 image on a 120 roll was pretty standard through the 50s. So, historically, photography was initially somewhat low aspect ratio: square, or 1.25:1.
Motion picture film started with all kinds of different "standards" until settling on 35mm in the early silent era. The Academy ratio was initially 1.33:1 (and this is where the 4:3 aspect comes from), but because of exhibition problems eventually settled on a masked 1.375:1 in the sound era. (For our purposes I'm ignoring widescreen innovation in the 50s.) So, many, many viewers became accustomed to this particular aspect ratio for the next 100 years and more, first in motion pictures and then in television. (Early TV [50s-80s] was 1.33:1 and slightly matted the Academy aperture motion pictures it broadcast, or worse in the case of widescreen. Later TV [90s-00s] settled on NTSC 720x480 pixels or 1.5:1 before the advent of recent 16:9 1920x1080 pixels HDTV.)
Still with me? More important for our purposes was Oscar Barnack inventing the Leica. The Leica was designed as a compact landscape camera. Barnack ran standard motion picture film through the camera horizontally for a 3:2 aspect ratio, or 1.5:1, kind of a happy compromise between Academy aperture and the golden rectangle; not too wide (more exposures per roll), not too square. Many glass plate manufacturers in Europe had used a similar aspect ratio, and in general, Europeans incline to higher aspect ratios in their TVs and standard paper sizes than the U.S. The Leica was very successful and spawned a number of early imitators, notably Canon and Nikon, and by the 60s, 35mm photography had supplanted 120 certainly as an amateur format and as a professional format in many instances, particularly a result of the advent of the SLR. By then, most users were accustomed to the 3:2 aspect; it was easier to compose for and was more visually dynamic. (It's a good exercise to compose for a square; it's much more difficult to pull off successfully. The 5DIV will allow Live View masking for 1:1 among other aspect ratios. Try it sometime and get some other synapses working!)
Which brings us to digital. When camera companies switched to digital, they already had been through the aspect wars and existing SLR customers were pretty fixed on 3:2 for the many reasons listed above. Additionally, the camera companies already had all the tooling (and lenses) for the 3:2 ratio and the 36x24 size; they weren't about to start all over again when they could just order up sensors in the same sizes (or similar aspect ratios) and drop them into existing frameworks with slight alterations and more or less meet their customers' expectations and existing inventories.
So, tl:dr -- Generally, humans are comfortable with what we know. Our lives are filled with many various rectangles and few squares. We are accustomed to looking at rectangles in our printed and projected matter. So, we use rectangular sensors as a matter of historical precedent, user expectations, and camera manufacturers' convenience. And it should be noted that the drool-worthy Hasselblad X1D shoots old-school Academy (1.33:1). It should also be noted that some of my favorite personal photographs were shot square.
Now, aren't you sorry you asked?
Motion picture film started with all kinds of different "standards" until settling on 35mm in the early silent era. The Academy ratio was initially 1.33:1 (and this is where the 4:3 aspect comes from), but because of exhibition problems eventually settled on a masked 1.375:1 in the sound era. (For our purposes I'm ignoring widescreen innovation in the 50s.) So, many, many viewers became accustomed to this particular aspect ratio for the next 100 years and more, first in motion pictures and then in television. (Early TV [50s-80s] was 1.33:1 and slightly matted the Academy aperture motion pictures it broadcast, or worse in the case of widescreen. Later TV [90s-00s] settled on NTSC 720x480 pixels or 1.5:1 before the advent of recent 16:9 1920x1080 pixels HDTV.)
Still with me? More important for our purposes was Oscar Barnack inventing the Leica. The Leica was designed as a compact landscape camera. Barnack ran standard motion picture film through the camera horizontally for a 3:2 aspect ratio, or 1.5:1, kind of a happy compromise between Academy aperture and the golden rectangle; not too wide (more exposures per roll), not too square. Many glass plate manufacturers in Europe had used a similar aspect ratio, and in general, Europeans incline to higher aspect ratios in their TVs and standard paper sizes than the U.S. The Leica was very successful and spawned a number of early imitators, notably Canon and Nikon, and by the 60s, 35mm photography had supplanted 120 certainly as an amateur format and as a professional format in many instances, particularly a result of the advent of the SLR. By then, most users were accustomed to the 3:2 aspect; it was easier to compose for and was more visually dynamic. (It's a good exercise to compose for a square; it's much more difficult to pull off successfully. The 5DIV will allow Live View masking for 1:1 among other aspect ratios. Try it sometime and get some other synapses working!)
Which brings us to digital. When camera companies switched to digital, they already had been through the aspect wars and existing SLR customers were pretty fixed on 3:2 for the many reasons listed above. Additionally, the camera companies already had all the tooling (and lenses) for the 3:2 ratio and the 36x24 size; they weren't about to start all over again when they could just order up sensors in the same sizes (or similar aspect ratios) and drop them into existing frameworks with slight alterations and more or less meet their customers' expectations and existing inventories.
So, tl:dr -- Generally, humans are comfortable with what we know. Our lives are filled with many various rectangles and few squares. We are accustomed to looking at rectangles in our printed and projected matter. So, we use rectangular sensors as a matter of historical precedent, user expectations, and camera manufacturers' convenience. And it should be noted that the drool-worthy Hasselblad X1D shoots old-school Academy (1.33:1). It should also be noted that some of my favorite personal photographs were shot square.
Now, aren't you sorry you asked?
Upvote
0
P
Pookie
Guest
JMZawodny said:
Nope... squares fit into film strips quite easily and depending on the format you get more squares than rectangles.
Upvote
0
Pookie said:
Pretty sure he meant that rectangles and squares are both optimal for fitting on a film strip, compared to circles.
But regardless, film ≠ digital. With film, you pay for every shot. You pay to buy the frame of film. You pay to develop the frame of film. Then you can choose to print a landscape- or portrait-oriented rectangle.
With digital, you can just shoot both ways and select later, for no extra cost.
Upvote
0
Similar threads
