Why not a 3:4 aspect ratio FX *width* sensor in Nikon / Canon etc?

[BJL]

Hi,

The usual problem with such analyses is the disconnect between correlation and causation. ...

How does that explain the centuries of predominance of shapes in the range 5:4 to 3:2, with many artists over that period making their own canvasses and so free to choose their shapes? How does it explain the predominance of choice of shapes in the range from square to about 7:5, on the other side of 3:2 from the Golden Ratio? There are a great many sizes and shapes of canvasses and drawing pads in art supply stores, far more variety than photographic printing papers: are you suggesting that despite an alleged strong artistic preference for about 8:5 image shape, no maker of canvasses or drawing paper has chosen to offer such a shape (or if they do, none of the art supply stores I visit has chosen to stock such an item.)

On the comment about book shapes: most book pages I see are in shapes between 5:4 and a bit under 3:2, so maybe that claim about nearly Golden Ratio shape is reference to open books being in proportions of about 8 wide by 5 high, which would fit the pages being about 5 high by 4 wide (as with 10"x8" prints.)

As a mathematician, I well aware of the fascinating mathematical properties of the Golden Ratio; but between scientists, I ask you to provide an evidence that "rectangular still image artists" actually gravitate towards that shape, because all the evidence I have gathered points in a different direction.  If I had to choose an empirical measure of "central tendency" it would be about 7:5.

[Bart_van_der_Wolf]

How does that explain the centuries of predominance of shapes in the range 5:4 to 3:2, with many artists over that period making their own canvasses and so free to choose their shapes? How does it explain the predominance of choice of shapes in the range from square to about 7:5, on the other side of 3:2 from the Golden Ratio? There are a great many sizes and shapes of canvasses and drawing pads in art supply stores, far more variety than photographic printing papers: are you suggesting that despite an alleged strong artistic preference for about 8:5 image shape, no maker of canvasses or drawing paper has chosen to offer such a shape (or if they do, none of the art supply stores I visit has chosen to stock such an item.)

There is perhaps nothing aesthetically particular about the aspect ratio choices of images dimensions that people make (which is not the same as compositional choices within the frame), if we are to believe this study:
Beauty in abstract paintings: perceptual contrast and statistical properties

Finally, in agreement with previous findings (McManus, 1980; Russell, 2000), we found no correlations of beauty ratings with the aspect ratio of the abstract artworks.

I do not know if aspect ratios of abstract artwork is appreciated much different from those of figurative art. I suppose it's even harder to get objective results due to sympathy/antipathy towards recognizable subjects.

Maybe more mundane factors play a role, such as habit, copycat behavior, or what have you. Also remember that e.g. the famous Night Watch by Rembrandt was trimmed on 4 sides, presumably to fit the wall space at its new location, talk about mundane...

As a mathematician, I well aware of the fascinating mathematical properties of the Golden Ratio; but between scientists, I ask you to provide an evidence that "rectangular still image artists" actually gravitate towards that shape, because all the evidence I have gathered points in a different direction.  If I had to choose an empirical measure of "central tendency" it would be about 7:5.

Apparently there is no conclusive evidence available in modern research, for either or any aspect ratio to be a clear winner. Again, that's not the same as compositional ratios within the frame size.

Ming Thein also has some useful thoughts about aspect ratios and compositional theory. He states that "the aim of this article is to focus on understanding the compositional impact of different aspect ratios, and more importantly, how to pick the right aspect ratio for a given subject".

It's that last part, "how to pick the right aspect ratio for a given subject" that resonates most with me. It is also what I like about stitching, the freedom to fit any aspect ratio frame around the subject. The subject is key, the frame is secondary (not unimportant but secondary), especially in the case of stitching for resolution, not necessarily for Field of View. The framing is used to augment the composition, e.g. by giving more weight to the sides or to the height.

Ming Thein also touches on why images with portrait orientation tend to have a more square aspect ratio, to do with the more challenging composition given our horizontally place eyes that have to scan the image in tandem. Maybe your study included many verticals, which would likely have shifted the weighting towards more squared aspect ratios.

Cheers,
Bart

[BJL]

There is perhaps nothing aesthetically particular about the aspect ratio choices of images dimensions that people make (which is not the same as compositional choices within the frame), if we are to believe this study:
Beauty in abstract paintings: perceptual contrast and statistical properties

Apparently there is no conclusive evidence available in modern research, for either or any aspect ratio to be a clear winner. Again, that's not the same as compositional ratios within the frame size.

Ming Thein also has some useful thoughts about aspect ratios and compositional theory. He states that "the aim of this article is to focus on understanding the compositional impact of different aspect ratios, and more importantly, how to pick the right aspect ratio for a given subject".

It's that last part, "how to pick the right aspect ratio for a given subject" that resonates most with me. ...

Ming Thein also touches on why images with portrait orientation tend to have a more square aspect ratio, to do with the more challenging composition given our horizontally place eyes that have to scan the image in tandem.

Maybe your study included many verticals, which would likely have shifted the weighting towards more squared aspect ratios.

the lack of evidence for a shape reference with abstract images seems to undermine any claim about a preference in abstract for a particular shape such, be it 4:3, 3:2 or the Golden Ratio.  That leads me even more to agree with Ming Thein that shape choices are likely based on subject matter: head shots favor a squarish shape like 4:5, distant landscapes without high mountains can favor wide shapes like 2:1 and beyond, etc.  That is why I surveyed actual representational art, both paintings and photographs.  If so, choosing a shape based on one's own history of shapes when cropping seems a good approach.

On the question of a bias towards verticals in my survey, I think I already mentioned (but in less detail) that even looking at shapes only for horizontals, the mode and median came out to the square side of 3:2: about 7:5 median; 4:3 mode.  The mode however is prone to be skewed by practical considerations like being a convenient shape for manufacturing and thus more commonly available, so the median might be more enlightening.

[Rory]

Maybe the sensor should be circular...  the circular mirror could split in the middle to keep the camera box depth the same.      

[Martin Wouterlood]

...  If I had to choose an empirical measure of "central tendency" it would be about 7:5.

...and that's near enough to 1:root 2, which fascinated me when I read about it once (for the life of me I cannot recall where) as a special ratio in the classical arts...easily done by a camera maker, I would guess - nip a mill off the side, add it to the top, there you go - 25x35mm. A subtle but tangible difference which I was drawn to when I briefly toyed with a 5"x7" many yrs ago. Personally I'd love to see it happen but sadly, the emphasis to include video features these days will mean I won't hold my breath.

[dwswager]

All images have their own built in aspect.  That is, there is an aspect at which the subject is shot that includes the entire subject as needed while cropping out extraneous information.  Of course, this is all based on subjective preferences of individual.  Some subjects lend themselves to a specific aspect ratio.  Not sure I would every shoot a landscape in square aspect or a head and shoulders shot in 3:2.  Hence, to make a general purpose camera, it requires that the sensor/film aspect account for the most options.  I never saw a single 2 1/4" Hasselblad out shooting landscapes, but they were the definitive standard for weddings in the day.

My D810 works fine for whatever aspect I want.  It comes with a built in 5:4 mode too.  There are plenty of pixels for however I wish to shoot most scenes.  Below are different aspects of the same basic photo.  While they are certainly different, they all came from the same 3:2 start.  Of course, once we get to "enough" pixels, it won't really matter.

[AFairley]

Nikon in a d810 lets you choose between  36 x24mm and  30x24mm ( 5:4) ,
then they have a 30x20mm and a DX mode 24 x16mm - so enough to choose from...

I hate the implementation of this in the D800 (I don't know if it's the same in the 810) because it is a hard crop of the RAW, so if you misframed slightly side to side you are SOL re ability to tweak in post.  Compare to, say, the Sony RX100, which has an aspect ratio option but soft crops the raw file so you do have the outer edges available in post.

Since I tend to work close to 4:5 or 3:4, I have masked the edges of my D800 focusing screen with transparent tape which gives me a hard framing guideline for the cropped format but let's me frame to 2:3 if needed.  That way I avoid the hard crop problem.

[dwswager]

I hate the implementation of this in the D800 (I don't know if it's the same in the 810) because it is a hard crop of the RAW, so if you misframed slightly side to side you are SOL re ability to tweak in post.  Compare to, say, the Sony RX100, which has an aspect ratio option but soft crops the raw file so you do have the outer edges available in post.

Since I tend to work close to 4:5 or 3:4, I have masked the edges of my D800 focusing screen with transparent tape which gives me a hard framing guideline for the cropped format but let's me frame to 2:3 if needed.  That way I avoid the hard crop problem.

It is a hard crop, but does give guides in viewfinder for 1.2x, 1.5x and 5:4 modes.  And if you turn custom setting a6 Focus Point Illumination to OFF it greys out the outside part of the frame as well.  I shoot 1.5x mode sometimes and don't have too much issue.  I do use the DK-17M 1.2x magnifying eyepiece.

I can see both sides to the hard/soft crop.  Hard crop makes smaller files and why would you want all the data if you don't want all the data.  Soft crop preserves more options if you make an error and actually did want all the data that could be captured.  With 36MP to work with now, I'm a little less anal about filling the frame and giving myself some options so I would probably shoot full FX and frame for another crop unless I new for sure I only needed 5:4 or something else.

[AFairley]

It is a hard crop, but does give guides in viewfinder for 1.2x, 1.5x and 5:4 modes.  And if you turn custom setting a6 Focus Point Illumination to OFF it greys out the outside part of the frame as well.

That's exactly why the hard crop gripes me so much.  With soft crop (or an option for either) the viewing system would be ideal for me.

[dwswager]

That's exactly why the hard crop gripes me so much.  With soft crop (or an option for either) the viewing system would be ideal for me.

Yeah, my preference would be full image area (FX 1.2X or DX) and the ability to put up crop guides in the viewfinder/Live View of a user defined specification.  Hence something like DX and 1:1 or FX and 5:7.  The 1st would give all 15.6MP of DX mode, with square viewfinder crop rectangle.  The 2nd would be a full 36MP file with 5:7 viewfinder crop rectangle.

[Jack Hogan]

the lack of evidence for a shape reference with abstract images seems to undermine any claim about a preference in abstract for a particular shape such, be it 4:3, 3:2 or the Golden Ratio.  That leads me even more to agree with Ming Thein that shape choices are likely based on subject matter: head shots favor a squarish shape like 4:5, distant landscapes without high mountains can favor wide shapes like 2:1 and beyond, etc.  That is why I surveyed actual representational art, both paintings and photographs.

I agree that the subject, as well as viewing distance, would ideally determine framing.  Alas it is not always simple to specify what frame means so aspect ratio is sometimes determined by other parameters as well: do you count the area surrounding the picture?  Does the size of the wall or the height of the ceiling count?  What about if in an alcove?  Each setting may well demand a specific ratio, lest the picture be out of proportion or 'not sit right'.

An informal survey around my house reveals a multitude of shapes, anywhere from oval, to square, to 3:2.  It does occur to me though that one framing decision that was given a lot of thought in the last couple of decades, independently of cript or credenza size, approaches and even exceeds the golden ratio: TVs.

Jack

[stamper]

Crop to content or frame for content. Forcing a ratio means that you will have to include something that "harms" an image or leaves out something that "helps" an image. Get specially made frames for prints if needed.

[dwswager]

Crop to content or frame for content. Forcing a ratio means that you will have to include something that "harms" an image or leaves out something that "helps" an image. Get specially made frames for prints if needed.

On one hand, each subject 'projects' it's own aspect ratio, while on the other, some aspects are just more naturally pleasing (familiar) to the human brain.

While I have already stated that I agree with this point, there are lots of real world factors that can intercede.  Everthing from ability to print the perfect crop to availability of matting and framing materials, to the space in which the image is to be displayed.  The framed print may look great by itself, but might look peculiar in the overall display space.

 

[BJL]

Crop to content or frame for content. Forcing a ratio means that you will have to include something that "harms" an image or leaves out something that "helps" an image. Get specially made frames for prints if needed.

On one hand, each subject 'projects' it's own aspect ratio, while on the other, some aspects are just more naturally pleasing (familiar) to the human brain.

... there are lots of real world factors that can intercede.  Everthing from ability to print the perfect crop to availability of matting and framing materials, to the space in which the image is to be displayed.

Even if some shapes are inherently more appealing that others, I would think that only a photographer with a highly rigid and narrow esthetic sense would always prefer the same shape for every subject, and the need to adapt to the intended display space is yet another reason not to be locked into one shape for all images.
Also, for modest deviations from common shapes, you can cut your own mats to the shape and size of the image that you wish to display; such matting also avoids the need to print to exactly the desired framing.

[dwswager]

Even if some shapes are inherently more appealing that others, I would think that only a photographer with a highly rigid and narrow esthetic sense would always prefer the same shape for every subject, and the need to adapt to the intended display space is yet another reason not to be locked into one shape for all images.
Also, for modest deviations from common shapes, you can cut your own mats to the shape and size of the image that you wish to display; such matting also avoids the need to print to exactly the desired framing.

The point of this thread was since the lens projects an image circle, why not a bigger sensor in the vertical direction which makes a more square frame.  Lots of technical reasons that is not viable.

The discussion of crops is more to do with standard imaging projects.  Not only are certain types of images more pleasing with a more square crop like 5:4, but some of them are required to be a certain aspect, head and shoulder portraits, for example.  The project might start with a given image size and aspect and you are required to do the best you can to put the subject into that frame.  A 3:1 landscape aspect doesn't work too well when you are placing it on a 2'x3' banner hung vertically.  Hence, with a natural tendency to fill the frame and do it in the most visually pleasing manner, crop marks in the viewfinder are helpful to remind you that while given the aspect of the sensor that is all great, but your gonna screw yourself when it comes time to execute the intended output of this image.  Some would prefer a soft crop so you still get the whole file because it makes re-purposing the image easier.

[BJL]

... crop marks in the viewfinder are helpful to remind you that while given the aspect of the sensor that is all great, but your gonna screw yourself when it comes time to execute the intended output of this image.  Some would prefer a soft crop so you still get the whole file because it makes re-purposing the image easier.

Yes; I occasionally use the in-camera shape-changing crop modes of the EM5, where the EVF and LCD preview the intended crop.  The crop is "soft", as it is only applied in the JPEG conversion, so in raw+JPEG mode, the raw file still has the whole image seen by the sensor, in case I change my mind.