Square sensors impossible?

[Michael Fryd]

...

But...

With FF lenses, it is still not optimal !!! Why wasting some part of the image circle for every ratio different from 3:2 ?? So we buy expensive lenses and we don't take full advantage of them because the sensor is not big enough...

...

All sensors waste a part of the image circle when the aspect ratio of the sensor doesn't match the aspect ratio of the final image.

Simoly wrong

There are two types of waste being talked about:

1. Light that's part of the image circle that doesn't hits outside the sensor
2. Pixels on the sensor that don't get used in the final image

Let's assume that our final image will be in a 2:3 aspect ratio. Any sensor that is not in a 2:3 aspect ratio will result in wasted pixels. Therefore we want a 2:3 aspect ratio sensor.

The 2:3 aspect ratio sensor that makes the best use of the image circe is one where the four corners hit the edge of the circle.

If your final image will be in the 2:3 aspect ratio, then a 2:3 sensor will make the best use of the image circle. If you final image will be a square, then a square is better. The sensor aspect ratio that is best in the general case is the one that matches the typical image.

Not correct.

Just use an oversized sensor. Besides it does not require to be much bigger.

A full frame camera has a sensor size of 24 by 36mm, and a 2:3 print uses all of it. If we want an oversize sensor we could use 36 by 36mm. However this does not allow a 2:3 print to use any more of the image circle, and it results in wasted pixels in that 2:3 print.

In terms of efficiency, it may seem that for a variety of aspect ratio prints, that a the largest square that fits into the image circle provides the least waste overall. However it doesn't work out that way. Square is the most efficient when the prints are square, and inefficient for wider aspect ratios. The most overall efficient is the one that somewhere in the range of aspect ratios. If you typically shoot anywhere from square to 16:9, then you want something between the two.

 

[Christof21]

...

But...

With FF lenses, it is still not optimal !!! Why wasting some part of the image circle for every ratio different from 3:2 ?? So we buy expensive lenses and we don't take full advantage of them because the sensor is not big enough...

...

All sensors waste a part of the image circle when the aspect ratio of the sensor doesn't match the aspect ratio of the final image.

Simoly wrong

There are two one types of waste being talked about:

1. Light that's part of the image circle that doesn't hits outside the sensor
2. Pixels on the sensor that don't get used in the final image

Let's assume that our final image will be in a 2:3 aspect ratio. Any sensor that is not in a 2:3 aspect ratio will result in wasted unused pixels. Therefore we want a 2:3 aspect ratio sensor.

Fixed now

Sorry, I will not enter in your very weird logic

 

[Michael Fryd]

...

But...

With FF lenses, it is still not optimal !!! Why wasting some part of the image circle for every ratio different from 3:2 ?? So we buy expensive lenses and we don't take full advantage of them because the sensor is not big enough...

...

All sensors waste a part of the image circle when the aspect ratio of the sensor doesn't match the aspect ratio of the final image.

Simoly wrong

There are two one types of waste being talked about:

1. Light that's part of the image circle that doesn't hits outside the sensor
2. Pixels on the sensor that don't get used in the final image

Let's assume that our final image will be in a 2:3 aspect ratio. Any sensor that is not in a 2:3 aspect ratio will result in wasted unused pixels. Therefore we want a 2:3 aspect ratio sensor.

Fixed now

Sorry, I will not enter in your very weird logic

It's not weird logic.

What is the problem you are trying to solve and why?

If the problem is how to make the most use out of an existing image circle for any aspect ratio, then the answer is to use a round sensor that matches the image circle. You could also use the smallest square that encloses that image circle if that's easier to manufacture.

But the answer is why? What are the advantages and disadvantages? Overall, would you be better off with simply a larger 2:3 sensor?

It's clear, that if you only produced 2:3 images, that a 2:3 sensor is best. You are not paying for pixels that are not being used, you are not paying for the circuitry and horsepower to drive those unused pixels, and you are not waiting around for them to be transferred.

So the question is, are you better off

• making a larger sensor that extends past the image circle
• making the image circle larger
• having a rectangular sensor whose corners just hit the image circle

 

[Tom_N]

If the sensor is oversized, or has a different AR/orientation than the image, there will be two kinds of wasted photosites:

1. Those that fall outside the image circle. These will ALWAYS be wasted.

2. Those that fall inside the image circle, but outside of all of the different rectangular or square capture areas. These, too, will ALWAYS be wasted.

There will also be temporarily-unused photosites:

3. Those that fall inside the capture areas for some ARs/orientations, but that do not fall inside the capture area for the current AR/orientation.

 

[J A C S]

An 18 mm x 18 mm sensor might allow him to capture a square image with an 18mm diameter, while the current 18 mm x 13.5 mm sensor only allows him to capture a square image with a 13.5mm diameter. That is a gain of 78% in area, hardly nothing. In fact it is about the same gain as the gain of surface area from MFT to APS-C, and should result in a a gain of about 2/3 stop in shot noise performance.

A square image doesn't have a diameter.

Every convex (you may want to be closed and bounded as well) figure has a diameter.

Loading…

en.wikipedia.org

 

[ed_ten_72]

The whole purpose of photography is to be able to look at result. These days people look at them through their phones, monitors, TV's. What is the poing of square photograph, if you can't see it comfortably? One more thing, you don't have a single eye, you have two eyes that are horizontally positioned one next to the other. Maybe that is why we prefer rectangular TVs these days. Remember, old TV's were almost square.

 

[cosmicnode]

An 18 mm x 18 mm sensor might allow him to capture a square image with an 18mm diameter, while the current 18 mm x 13.5 mm sensor only allows him to capture a square image with a 13.5mm diameter. That is a gain of 78% in area, hardly nothing. In fact it is about the same gain as the gain of surface area from MFT to APS-C, and should result in a a gain of about 2/3 stop in shot noise performance.

A square image doesn't have a diameter.

Every convex (you may want to be closed and bounded as well) figure has a diameter.

https://en.wikipedia.org/wiki/Diameter

A square image has a diagonal, not a diameter. In geometry, a diameter of a circle is any straight line segment that passes through the center of the circle and whose endpoints lie on the circle.

 

[J A C S]

An 18 mm x 18 mm sensor might allow him to capture a square image with an 18mm diameter, while the current 18 mm x 13.5 mm sensor only allows him to capture a square image with a 13.5mm diameter. That is a gain of 78% in area, hardly nothing. In fact it is about the same gain as the gain of surface area from MFT to APS-C, and should result in a a gain of about 2/3 stop in shot noise performance.

A square image doesn't have a diameter.

Every convex (you may want to be closed and bounded as well) figure has a diameter.

https://en.wikipedia.org/wiki/Diameter

A square image has a diagonal, not a diameter. In geometry, a diameter of a circle is any straight line segment that passes through the center of the circle and whose endpoints lie on the circle.

Just read the link:

For a convex shape in the plane, the diameter is defined to be the largest distance that can be formed between two opposite parallel lines tangent to its boundary, and the width is often defined to be the smallest such distance.

 

[cosmicnode]

An 18 mm x 18 mm sensor might allow him to capture a square image with an 18mm diameter, while the current 18 mm x 13.5 mm sensor only allows him to capture a square image with a 13.5mm diameter. That is a gain of 78% in area, hardly nothing. In fact it is about the same gain as the gain of surface area from MFT to APS-C, and should result in a a gain of about 2/3 stop in shot noise performance.

A square image doesn't have a diameter.

Every convex (you may want to be closed and bounded as well) figure has a diameter.

https://en.wikipedia.org/wiki/Diameter

A square image has a diagonal, not a diameter. In geometry, a diameter of a circle is any straight line segment that passes through the center of the circle and whose endpoints lie on the circle.

Just read the link:

For a convex shape in the plane, the diameter is defined to be the largest distance that can be formed between two opposite parallel lines tangent to its boundary, and the width is often defined to be the smallest such distance.

A sensor is not a convex shape. it is flat

A square or rectangular sensor does not have a diameter,

 

[Iliah Borg]

An 18 mm x 18 mm sensor might allow him to capture a square image with an 18mm diameter, while the current 18 mm x 13.5 mm sensor only allows him to capture a square image with a 13.5mm diameter. That is a gain of 78% in area, hardly nothing. In fact it is about the same gain as the gain of surface area from MFT to APS-C, and should result in a a gain of about 2/3 stop in shot noise performance.

A square image doesn't have a diameter.

Every convex (you may want to be closed and bounded as well) figure has a diameter.

https://en.wikipedia.org/wiki/Diameter

A square image has a diagonal, not a diameter. In geometry, a diameter of a circle is any straight line segment that passes through the center of the circle and whose endpoints lie on the circle.

Just read the link:

For a convex shape in the plane, the diameter is defined to be the largest distance that can be formed between two opposite parallel lines tangent to its boundary, and the width is often defined to be the smallest such distance.

A sensor is not a convex shape. it is flat

A square or rectangular sensor does not have a diameter,

You really shouldn't be arguing these points.

In geometry, for a shape to be convex, a line segment drawn between any two points on the shape must always lie inside the shape.

 

[cosmicnode]

An 18 mm x 18 mm sensor might allow him to capture a square image with an 18mm diameter, while the current 18 mm x 13.5 mm sensor only allows him to capture a square image with a 13.5mm diameter. That is a gain of 78% in area, hardly nothing. In fact it is about the same gain as the gain of surface area from MFT to APS-C, and should result in a a gain of about 2/3 stop in shot noise performance.

A square image doesn't have a diameter.

Every convex (you may want to be closed and bounded as well) figure has a diameter.

https://en.wikipedia.org/wiki/Diameter

A square image has a diagonal, not a diameter. In geometry, a diameter of a circle is any straight line segment that passes through the center of the circle and whose endpoints lie on the circle.

Just read the link:

For a convex shape in the plane, the diameter is defined to be the largest distance that can be formed between two opposite parallel lines tangent to its boundary, and the width is often defined to be the smallest such distance.

A sensor is not a convex shape. it is flat

A square or rectangular sensor does not have a diameter,

You really shouldn't be arguing these points.

In geometry, for a shape to be convex, a line segment drawn between any two points on the shape must always lie inside the shape.

as in a lens?

 

[Iliah Borg]

An 18 mm x 18 mm sensor might allow him to capture a square image with an 18mm diameter, while the current 18 mm x 13.5 mm sensor only allows him to capture a square image with a 13.5mm diameter. That is a gain of 78% in area, hardly nothing. In fact it is about the same gain as the gain of surface area from MFT to APS-C, and should result in a a gain of about 2/3 stop in shot noise performance.

A square image doesn't have a diameter.

Every convex (you may want to be closed and bounded as well) figure has a diameter.

https://en.wikipedia.org/wiki/Diameter

A square image has a diagonal, not a diameter. In geometry, a diameter of a circle is any straight line segment that passes through the center of the circle and whose endpoints lie on the circle.

Just read the link:

For a convex shape in the plane, the diameter is defined to be the largest distance that can be formed between two opposite parallel lines tangent to its boundary, and the width is often defined to be the smallest such distance.

A sensor is not a convex shape. it is flat

A square or rectangular sensor does not have a diameter,

You really shouldn't be arguing these points.

In geometry, for a shape to be convex, a line segment drawn between any two points on the shape must always lie inside the shape.

as in a lens?

Please have a look at the definition of a convex polygon https://en.wikipedia.org/wiki/Convex_polygon

 

[petrochemist]

You really shouldn't be arguing these points.

In geometry, for a shape to be convex, a line segment drawn between any two points on the shape must always lie inside the shape.

If the edges are straight a line segment between two points on the same side does not lie inside the shape. It lies directly on the boundary of the shape.

The normal understanding of a convex shape is one that bulges outwards, like a square with barrel distortion. The definition you've quoted applies to these, but not to a square.

 

[J A C S]

An 18 mm x 18 mm sensor might allow him to capture a square image with an 18mm diameter, while the current 18 mm x 13.5 mm sensor only allows him to capture a square image with a 13.5mm diameter. That is a gain of 78% in area, hardly nothing. In fact it is about the same gain as the gain of surface area from MFT to APS-C, and should result in a a gain of about 2/3 stop in shot noise performance.

A square image doesn't have a diameter.

Every convex (you may want to be closed and bounded as well) figure has a diameter.

https://en.wikipedia.org/wiki/Diameter

A square image has a diagonal, not a diameter. In geometry, a diameter of a circle is any straight line segment that passes through the center of the circle and whose endpoints lie on the circle.

Just read the link:

For a convex shape in the plane, the diameter is defined to be the largest distance that can be formed between two opposite parallel lines tangent to its boundary, and the width is often defined to be the smallest such distance.

A sensor is not a convex shape. it is flat

LOL! It is rectangular. The third dimension does not matter for this discussion. Even with it, it is still convex.

A square or rectangular sensor does not have a diameter,

Sure, if you says so.

Why don't you just admit that you were wrong?

 

[Iliah Borg]

The normal understanding of a convex shape is one that bulges outwards, like a square with barrel distortion.

You can use the terms as you like, but you don't argue when somebody is using correct terms, especially not argue based on "common sense". Remarks made by JACS are correct.

--
http://www.libraw.org/

 

[J A C S]

You really shouldn't be arguing these points.

In geometry, for a shape to be convex, a line segment drawn between any two points on the shape must always lie inside the shape.

If the edges are straight a line segment between two points on the same side does not lie inside the shape. It lies directly on the boundary of the shape.

Who said "inside"? The boundary is included in those examples.

The normal understanding of a convex shape is one that bulges outwards, like a square with barrel distortion. The definition you've quoted applies to these, but not to a square.

 

[Iliah Borg]

You really shouldn't be arguing these points.

In geometry, for a shape to be convex, a line segment drawn between any two points on the shape must always lie inside the shape.

If the edges are straight a line segment between two points on the same side does not lie inside the shape. It lies directly on the boundary of the shape.

Who said "inside"?

I said, and it's not totally correct (depending on how we understand "inside"). I should have said "inside or on the boundary".

The boundary is included in those examples.

Yes.

 

[J A C S]

You really shouldn't be arguing these points.

In geometry, for a shape to be convex, a line segment drawn between any two points on the shape must always lie inside the shape.

If the edges are straight a line segment between two points on the same side does not lie inside the shape. It lies directly on the boundary of the shape.

Who said "inside"?

I said, and it's not totally correct (depending on how we understand "inside").

Yes, "inside" means in the set, not away from the boundary. Your definition was fine, that is why I put quotes.

I should have said "inside or on the boundary".

There is no problem even if the set is open (the interior is not included). Then those two points you take would not be on the boundary in the first place, so there is no problem with the definition. An open rectangle is convex, and a closed one (the boundary included) is convex as well.

Besides, convexity is not really needed. A diameter of a set is often defined as the supremum (something like a maximum) of the distance between pairs of points (maximized over all pairs). If the boundary is included, the supremum is actually a genuine maximum.

 

[cosmicnode]

An 18 mm x 18 mm sensor might allow him to capture a square image with an 18mm diameter, while the current 18 mm x 13.5 mm sensor only allows him to capture a square image with a 13.5mm diameter. That is a gain of 78% in area, hardly nothing. In fact it is about the same gain as the gain of surface area from MFT to APS-C, and should result in a a gain of about 2/3 stop in shot noise performance.

A square image doesn't have a diameter.

Every convex (you may want to be closed and bounded as well) figure has a diameter.

https://en.wikipedia.org/wiki/Diameter

A square image has a diagonal, not a diameter. In geometry, a diameter of a circle is any straight line segment that passes through the center of the circle and whose endpoints lie on the circle.

Just read the link:

For a convex shape in the plane, the diameter is defined to be the largest distance that can be formed between two opposite parallel lines tangent to its boundary, and the width is often defined to be the smallest such distance.

A sensor is not a convex shape. it is flat

LOL! It is rectangular. The third dimension does not matter for this discussion. Even with it, it is still convex.

A square or rectangular sensor does not have a diameter,

Sure, if you says so.

Why don't you just admit that you were wrong?

https://www.mathsisfun.com/definitions/diameter.html

 

[J A C S]

An 18 mm x 18 mm sensor might allow him to capture a square image with an 18mm diameter, while the current 18 mm x 13.5 mm sensor only allows him to capture a square image with a 13.5mm diameter. That is a gain of 78% in area, hardly nothing. In fact it is about the same gain as the gain of surface area from MFT to APS-C, and should result in a a gain of about 2/3 stop in shot noise performance.

A square image doesn't have a diameter.

Every convex (you may want to be closed and bounded as well) figure has a diameter.

https://en.wikipedia.org/wiki/Diameter

A square image has a diagonal, not a diameter. In geometry, a diameter of a circle is any straight line segment that passes through the center of the circle and whose endpoints lie on the circle.

Just read the link:

For a convex shape in the plane, the diameter is defined to be the largest distance that can be formed between two opposite parallel lines tangent to its boundary, and the width is often defined to be the smallest such distance.

A sensor is not a convex shape. it is flat

LOL! It is rectangular. The third dimension does not matter for this discussion. Even with it, it is still convex.

A square or rectangular sensor does not have a diameter,

Sure, if you says so.

Why don't you just admit that you were wrong?

https://www.mathsisfun.com/definitions/diameter.html

Keep digging. This is diameter of a circle (it should be a disk, actually). I gave you are link for a diamater of a general "shape".

The definitions given above are only valid for circles, spheres and convex shapes. However, they are special cases of a more general definition that is valid for any kind of n-dimensional convex or non-convex object, such as a hypercube or a set of scattered points. The diameter of a subset of a metric space is the least upper bound of the set of all distances between pairs of points in the subset.

Look, you cannot google your way to knowledge, at least not the way you are doing it. You have to at least know what to google.