BorisK1

Lives in United States MI, United States
Works as a Software engineer
Joined on May 7, 2004

Comments

Total: 430, showing: 1 – 20
« First‹ Previous12345Next ›Last »
In reply to:

dan pv: There is no such thing as "perspective distortion" either. The perspective is correct in all cases. But both terms, "perspective distortion" and "perspective compression", refer to how the image created by wide and tele lenses is different of the image expected by our brain. They don't describe real physical phenomenas. The perspective of the image seems, or looks like, distorted or compressed compared with the image our brain expects due to the difference between the lens angle of view and the human eye angle of view. Both terms are equally legitim as they describe sensations and not physical phenomenas. This article is misleading.

> The front and rear focal points of an optical system are cardinal points that exists without any implication or evaluation made by our brain/senses. They can be entirely computed based on the characteristics of the optical system alone and therefore they are part of a "real physical phenomena". The cardinal points can be fully defined and mathematically explained even to a blind person.

The viewing distance that removes perspective distortion, is a physical characteristic of the printed image. I can express it mathematically.

Let me try again.

Suppose, you know an object's dimensions.

If you measure its angular size, it's trivial to calculate the distance to it.

If you measure its perspective (say, nose-width to ears-width ratio), you can also calculate the distance.

If those two distances aren't equal, that means you're looking at a perspective-distorted image of the object.

Link | Posted on May 26, 2018 at 00:31 UTC
In reply to:

dan pv: There is no such thing as "perspective distortion" either. The perspective is correct in all cases. But both terms, "perspective distortion" and "perspective compression", refer to how the image created by wide and tele lenses is different of the image expected by our brain. They don't describe real physical phenomenas. The perspective of the image seems, or looks like, distorted or compressed compared with the image our brain expects due to the difference between the lens angle of view and the human eye angle of view. Both terms are equally legitim as they describe sensations and not physical phenomenas. This article is misleading.

> So far I can see that our only desagrement is about the terms "real physical phenomena" and "sensation".

Let's say, we take a real-life projector, that projects an image onto a wall.

What I understand as "perspective distortion" in photography is similar to that image being "in focus".

It's true that we have a "sensation" of the image being fuzzy, and without this sensation, there's nothing inherently "wrong" with an out-of-focus image.

And yes, it could be measured without the brain (there are projectors with autofocus).

If you consider that image being "in focus" as not "real", then the perspective distortion is not "real" either.

Link | Posted on May 25, 2018 at 22:16 UTC
In reply to:

dan pv: There is no such thing as "perspective distortion" either. The perspective is correct in all cases. But both terms, "perspective distortion" and "perspective compression", refer to how the image created by wide and tele lenses is different of the image expected by our brain. They don't describe real physical phenomenas. The perspective of the image seems, or looks like, distorted or compressed compared with the image our brain expects due to the difference between the lens angle of view and the human eye angle of view. Both terms are equally legitim as they describe sensations and not physical phenomenas. This article is misleading.

> The term "Perspective distortion" means 2 completely different things in Photography and Perspective Geometry.

Agreed.

> In Perspective Geometry "Perspective Distortion" is called the difference between the central projection on a flat plan (as is the Perspective in Photography) and and the image created by the eye (=central projection on a curved plan). Because of this differences between the 2 types of central projection a photography will never perfectly resample an image seen by the eye.

> There is no such thing as the "Perspective distortion" with the meaning it has in Photography.

Yes, in photography, "Perspective distortion" means something very different, and I can't find a good definition.

This type of perspective distortion is not a property of the image.

What *is* a property of the image, is a specific *viewing distance*, from which it looks undistorted.

In photography, perspective distortion is the property of the image's angular size, as seen by the viewer.

Link | Posted on May 25, 2018 at 21:48 UTC
In reply to:

dan pv: There is no such thing as "perspective distortion" either. The perspective is correct in all cases. But both terms, "perspective distortion" and "perspective compression", refer to how the image created by wide and tele lenses is different of the image expected by our brain. They don't describe real physical phenomenas. The perspective of the image seems, or looks like, distorted or compressed compared with the image our brain expects due to the difference between the lens angle of view and the human eye angle of view. Both terms are equally legitim as they describe sensations and not physical phenomenas. This article is misleading.

> If you make 2 photos of a person from a tripod without moving the camera or the person with a tele and then with a wide the perspective won't change. That's a fact.

Up to this point, we're in complete agreement.

> The tele image will be a crop of the wide one. The measurements of the head of the person in bot images will be identical. That means NO REAL DISTORTION induced by changing the FOV (=the lens). This is a well known fact in Perspective.

Agreed here, too. But it's not about changing the captured FOV.

We print the image, and start changing the *viewing distance*. This changes the *angular size* of the image for the viewer.

When the image's angular size matches the camera's FOV, the viewer's view exactly matches the view of the subject *as seen at the subject distance*.

At other viewing distances, what the viewer sees, does not correspond to *any* subject distance.

It's distorted, because you *can't* see a subject like this. Perspective doesn't match the angular size.

Link | Posted on May 25, 2018 at 20:06 UTC
In reply to:

dan pv: There is no such thing as "perspective distortion" either. The perspective is correct in all cases. But both terms, "perspective distortion" and "perspective compression", refer to how the image created by wide and tele lenses is different of the image expected by our brain. They don't describe real physical phenomenas. The perspective of the image seems, or looks like, distorted or compressed compared with the image our brain expects due to the difference between the lens angle of view and the human eye angle of view. Both terms are equally legitim as they describe sensations and not physical phenomenas. This article is misleading.

Perhaps you're thinking that I'm measuring whether or not the human will get the feeling of distortion, rather than some physical quality of the image?

Here's another way to show that there *is* real distortion.

Let's say you measured the width of my nose (nose_width), the distance between my ears (ears_width), and the distance from my nose to my ears (face_depth).

Then, you take a head shot, print it, and view it.

If you measure the ear-to-ear angle you see *within the image*, you can calculate the apparent distance to the ears:
ears_apparent = ears_width / (2 * tan(ears_angle / 2)).

If you measure the nose-width angle, you can calculate the apparent distance to the nose in the same way:
nose_apparent = ears_width / (2 * tan(ears_angle / 2)).

Now, (ears_apparent - nose_apparent) should be equal to face_depth.

But that only happens if the image's angular size within the viewer's FOV is equal to the captured FOV.

Otherwise, they won't be equal. So the image really is distorted.

Link | Posted on May 25, 2018 at 15:09 UTC
In reply to:

photophile: OK. Interesting video and thank you for taking the time to illustrate - but seriously ? FEET??? What is that? How many feet in a gallon?

He actually asked about "FEET". Which, according to a Google search, stands for "Fellowship of European Evangelical Theologians". The volume of an average human body is 18.41 gallons. Google says that there are about 600 million evangelicals in the world. Assuming 1% of them to be theologians, we get 110,460,000 gallons total.

Which works out to about 0.000,000,009 FEET per gallon.

Is this better?

Link | Posted on May 25, 2018 at 14:26 UTC
In reply to:

dan pv: There is no such thing as "perspective distortion" either. The perspective is correct in all cases. But both terms, "perspective distortion" and "perspective compression", refer to how the image created by wide and tele lenses is different of the image expected by our brain. They don't describe real physical phenomenas. The perspective of the image seems, or looks like, distorted or compressed compared with the image our brain expects due to the difference between the lens angle of view and the human eye angle of view. Both terms are equally legitim as they describe sensations and not physical phenomenas. This article is misleading.

I just described a method to measure the amount of perspective distortion, based on measuring the angles, without referring to any "sensations". That makes it real. Like temperature, not like heat/cold sensation (which can change with humidity, the person's well-being, etc).

You could come up with different temperature scales, too. Doesn't mean temperature isn't real.

There's no reason whatsoever to make this into a separate area of study in geometry. As geometric problems go, it is very basic.

Link | Posted on May 25, 2018 at 00:56 UTC
In reply to:

BorisK1: From the video: "The distance to the subject is creating distorion".
This is incorrect.

Perspective distortion is caused by the mismatch of the angular field of view captured by the camera and the apparent size of the displayed image.

For example, a 25mm-equivalent lens has 81.8° field of view. That means, to remove perspective distortion, an 8x10" print of such an image has to be viewed up close - from about 7.3", if I counted right.

*It does not matter* how far the subject was from the lens. As long as the viewing angle of the print matches the field of view captured by the camera, there won't be any distortion.

"Lens compression" is what happens when you're viewing the print at a *wider* angle than what the camera captured.

"Wide angle distortion" is when you're viewing the print at a *narrower* angle than what the camera captured.

> So the last point is a bit bizarre - are you trying to match something? There is not going to be a significant difference in relative sizes until you get very close.

I'm talking about matching two fields of vision: The field of view captured by the camera, against the viewer's angular field of view of the image.

The image looks undistorted when those two fields of vision are equal.

Link | Posted on May 24, 2018 at 22:34 UTC
In reply to:

BorisK1: From the video: "The distance to the subject is creating distorion".
This is incorrect.

Perspective distortion is caused by the mismatch of the angular field of view captured by the camera and the apparent size of the displayed image.

For example, a 25mm-equivalent lens has 81.8° field of view. That means, to remove perspective distortion, an 8x10" print of such an image has to be viewed up close - from about 7.3", if I counted right.

*It does not matter* how far the subject was from the lens. As long as the viewing angle of the print matches the field of view captured by the camera, there won't be any distortion.

"Lens compression" is what happens when you're viewing the print at a *wider* angle than what the camera captured.

"Wide angle distortion" is when you're viewing the print at a *narrower* angle than what the camera captured.

> there is only one kind of compression that happens when capturing an image and that is the result of your distance to the objects, not the lens.
> It is the relative sizes of the objects to each other in the captured frame. It has NOTHING to do with the field of view of the lens.

The ratio of the sizes of near-to-far objects within the frame is called "perspective". Yes, it's determined by subject distance.

> Once you've printed that captured image, that size ratio between objects in the frame does not change on the flat image.

Yes, the image captures *perspective*.

Changing viewing distance affects the viewer's *angle of view*, not *perspective*.

> But get very close, you will start to see size mismatches [...] Pull back, that is reduced [...].

This depends on the captured FOV. For a 90° FOV (a 21mm lens), you get close to the image for the mismatches to disappear. For a 25° FOV (100mm lens), you pull back.

Either way, you're matching viewer's FOV to the camera's.

Link | Posted on May 24, 2018 at 21:31 UTC
In reply to:

BorisK1: From the video: "The distance to the subject is creating distorion".
This is incorrect.

Perspective distortion is caused by the mismatch of the angular field of view captured by the camera and the apparent size of the displayed image.

For example, a 25mm-equivalent lens has 81.8° field of view. That means, to remove perspective distortion, an 8x10" print of such an image has to be viewed up close - from about 7.3", if I counted right.

*It does not matter* how far the subject was from the lens. As long as the viewing angle of the print matches the field of view captured by the camera, there won't be any distortion.

"Lens compression" is what happens when you're viewing the print at a *wider* angle than what the camera captured.

"Wide angle distortion" is when you're viewing the print at a *narrower* angle than what the camera captured.

@stevo23

> @Borisk - you mean perspective exaggeration? Viewing distance of a flat printed image has no impact on compression or even straight line distortion.

Sorry, I don't know what exactly you mean by "perspective exxageration" or "straight line distortion".

I'm talking about what's commonly known as "perspective distortion" or "telephoto compression". For example, when viewing an image taken by a wide angle lens, it's common to notice that the parts of the image that were close to the camera, appear to be "larger than they should be".

*This* type of perspective distortion depends on the viewing distance. It disappears when the image viewing angle matches the viewing angle captured by the camera.

For example, an 8x10" image taken by a 21mm lens on a 35mm camera, will look distorted when viewed from 12", less distorted when viewed from 8", and undistorted when viewed from 6" away.

For a 45mm lens on a 35mm camera, that distance is 12". For 90mm, it's 24".

Link | Posted on May 24, 2018 at 20:41 UTC
In reply to:

dan pv: There is no such thing as "perspective distortion" either. The perspective is correct in all cases. But both terms, "perspective distortion" and "perspective compression", refer to how the image created by wide and tele lenses is different of the image expected by our brain. They don't describe real physical phenomenas. The perspective of the image seems, or looks like, distorted or compressed compared with the image our brain expects due to the difference between the lens angle of view and the human eye angle of view. Both terms are equally legitim as they describe sensations and not physical phenomenas. This article is misleading.

No, I'm talking about about real ratios of real angles, that you can measure with a real protractor. It's *geometry*, not some nebulous "sensation".

You're right that *perspective* - as captured by the image - doesn't change.

But there's also "expected" perspective - the perspective the viewer *should* see, given the *angular size* of the subject.

For example, if a human head takes up 20 degrees of your vision, you expect it to be about 3 feet away.

You can measure perspective, for example, as a ratio of the nose width to the face width.

If you measure the *actual* perspective (using camera-to-subject distance), and the *expected* perspective (using "expected" distance based on the viewer's angle of view of the subject), then the amount of perspective distortion is the ratio of these two numbers.

If it's 1, you have no distortion. If it's greater than 1 (nose is too wide), you have wide-angle distortion. If it's less than 1 (nose is too narrow), you have telephoto compression.

Link | Posted on May 24, 2018 at 20:26 UTC
In reply to:

dan pv: "The compression you get using a long lens isn't a result of the lens, so much as the distance between your subject, your background, and the camera."

That's not true. If you keep unchanged the distance between your subject, your background, and the camera and use a tele lens instead of a normal one , assuming theoretical lenses with infinite DOF, the perspective will seam compressed. The perspective itself doesn't change. What changes is the crop factor of the image. Our brain is used with normal angle of view images and it interprets the cropped image as a normal image. This creates the sensation of compressed perspective. And this sensation is entirely the result of the crop factor induced by the tele lens.
Of course that by changing the distance between your subject, your background and the camera one can produce a sort of "compressed perspective" image, too. But it has nothing to do with "lens compression".

> I have never heard of someone looking at pictures from different distances depending on the FL of the lens that took the picture.

It happens more often than you'd think.

What's traditionally considered "normal" viewing distance, is fairly close .

The classic example of a different viewing - a portrait that's hanging on the wall - is definitely further away than a "normal" viewing distance.

Holding a small smartphone that close would be just plain uncomfortable.

Link | Posted on May 24, 2018 at 19:55 UTC
In reply to:

bives: Actually, there is not distortion, either. It is all simply the perspective of the lens from the distance of the shot to the subject. We tend to think of “distortion” as anything that looks unusual to us. But what is usual to us is simply due to the optics of our particular eyeballs. In my view, it is all neutral in the sense that the way the image looks is due to the focal length and the distance. It is just optics. Why call something distortion?

bives:
> But what is usual to us is simply due to the optics of our particular eyeballs

The "perspective distortion" effect is not due to the optics of our eyeballs. It's a mismatch between the field of view captured by the camera, and the angular size of the image within the viewer's field of vision.

It's a real geometrical effect, and can be measured independently of the human vision. The amount of perspective distortion depends on the viewing distance.

For example, a 21mm lens on a 35mm camera has 90 degrees field of view. To view such an image without perspective distortion, the viewing distance has to be equal to 1/2 of the print's diagonal.

Link | Posted on May 24, 2018 at 19:20 UTC
In reply to:

BorisK1: From the video: "The distance to the subject is creating distorion".
This is incorrect.

Perspective distortion is caused by the mismatch of the angular field of view captured by the camera and the apparent size of the displayed image.

For example, a 25mm-equivalent lens has 81.8° field of view. That means, to remove perspective distortion, an 8x10" print of such an image has to be viewed up close - from about 7.3", if I counted right.

*It does not matter* how far the subject was from the lens. As long as the viewing angle of the print matches the field of view captured by the camera, there won't be any distortion.

"Lens compression" is what happens when you're viewing the print at a *wider* angle than what the camera captured.

"Wide angle distortion" is when you're viewing the print at a *narrower* angle than what the camera captured.

@dr8 When I take a photo of a persons face from about 1ft with a 21mm wide angle lens, the resulting image is in no way the same as if I were to put my eye 1ft from their face.

That image captures perspective you'd see from that distance.

> I have tried this and I can honestly say that I do not see the facial features pulled, stretched, and "distorted" as in the captured image. The photo & reality are not the same.

The amount of distortion you see in the captured image depends on how close you're viewing the image.

A 21mm lens covers roughly 90 degrees diagonally. To remove distortion, you need to match the camera's field of vision *when you're viewing the image*.

To match 90 degree field of vision, you need to view the image from a distance equal to half its diagonal.

For example, if you print that image as 8x10", you need it to view it from about six inches.

Link | Posted on May 24, 2018 at 18:42 UTC
In reply to:

dan pv: There is no such thing as "perspective distortion" either. The perspective is correct in all cases. But both terms, "perspective distortion" and "perspective compression", refer to how the image created by wide and tele lenses is different of the image expected by our brain. They don't describe real physical phenomenas. The perspective of the image seems, or looks like, distorted or compressed compared with the image our brain expects due to the difference between the lens angle of view and the human eye angle of view. Both terms are equally legitim as they describe sensations and not physical phenomenas. This article is misleading.

Perspective distortion is very much a physical phenomenon. It's independent of the eye/brain, and you can measure it, using a tool that measures angular sizes - a sextant, or a protractor.

1. Measure the nose-width-to-face-width ratio at the "subject distance". Let's call this "true perspective".
2. Measure the "angular size" of the subject *as visible to the viewer*. This will depend on the print size, and the viewing distance.
3. Find the "viewer's apparent distance" - that's how far the subject needs to be from the viewer to match that angular size.
4. Measure the nose-width-to-face-width ratio at the "viewer's apparent distance". Let's call this "expected perspective".

The ratio of the "expected perspective" to the "true perspective" is the measure of perspective distortion. If it's 1, there's no distortion. If it's greater than 1 (the nose looks too wide), you have wide-angle distortion. If it's less than 1 (the nose looks too narrow), you have telephoto compression.

Link | Posted on May 24, 2018 at 18:26 UTC
In reply to:

photophile: OK. Interesting video and thank you for taking the time to illustrate - but seriously ? FEET??? What is that? How many feet in a gallon?

A liquid gallon is 0.133681 cubic feet.

Well, you asked...

Link | Posted on May 24, 2018 at 17:20 UTC
In reply to:

rc53: There is one point missing from this demonstration; the viewing distance between the eyes and the final image or computer screen. This is usually reckoned to be about 20" for normal sized images.

If you take an image with a 16mm lens, as in the first images here, and then project it onto a wall, so that it measures several feet by several feet, and then go to 20" from it, the 'distortion' will not be apparent.

It's common to express the viewing distance in image diagonals. This way, the size of the image doesn't matter - as long as the viewing distance is 1 diagonal away from the image, it's the "normal" viewing.

In cinema, "normal" viewing distance is traditionally about 2 image diagonals.

Link | Posted on May 24, 2018 at 17:18 UTC
In reply to:

dan pv: "The compression you get using a long lens isn't a result of the lens, so much as the distance between your subject, your background, and the camera."

That's not true. If you keep unchanged the distance between your subject, your background, and the camera and use a tele lens instead of a normal one , assuming theoretical lenses with infinite DOF, the perspective will seam compressed. The perspective itself doesn't change. What changes is the crop factor of the image. Our brain is used with normal angle of view images and it interprets the cropped image as a normal image. This creates the sensation of compressed perspective. And this sensation is entirely the result of the crop factor induced by the tele lens.
Of course that by changing the distance between your subject, your background and the camera one can produce a sort of "compressed perspective" image, too. But it has nothing to do with "lens compression".

> If you keep unchanged the distance between your subject, your background, and the camera and use a tele lens instead of a normal one , assuming theoretical lenses with infinite DOF, the perspective will seam compressed.

This only holds true as long as the print is viewed at the normal viewing distance (about 1 diagonal of the image away from the viewer).

Move the print further away, and perspective becomes less compressed.

When the image takes up the exact same angle within viewer's vision as the angle of view that the camera captured, perspective distortion goes away.

> Our brain is used with normal angle of view images and it interprets the cropped image as a normal image.

Not all images are viewed at "normal" viewing distance. For example, a portrait on a wall angle of view is narrower than "normal".

Our brain can estimate subject distances based on subject's perspective, and also based on the subject's angular size. When those distances don't match, we see distortion.

Link | Posted on May 24, 2018 at 17:02 UTC
In reply to:

BorisK1: "Eco-friendly" mode, huh? My car has an "eco" button - to make the AC run at half-power, and make the acceleration lethargic at best. Which is, apparently, what it takes to approach the advertised gas mileage.
I wonder what an "eco-friendly"-mode does on a camera...

I'd think the primary ways to save the battery have to do with the screen. Reduce brightness, resolution, and refresh rate - and you can cut power quite a bit.

Link | Posted on May 24, 2018 at 15:26 UTC
In reply to:

BorisK1: From the video: "The distance to the subject is creating distorion".
This is incorrect.

Perspective distortion is caused by the mismatch of the angular field of view captured by the camera and the apparent size of the displayed image.

For example, a 25mm-equivalent lens has 81.8° field of view. That means, to remove perspective distortion, an 8x10" print of such an image has to be viewed up close - from about 7.3", if I counted right.

*It does not matter* how far the subject was from the lens. As long as the viewing angle of the print matches the field of view captured by the camera, there won't be any distortion.

"Lens compression" is what happens when you're viewing the print at a *wider* angle than what the camera captured.

"Wide angle distortion" is when you're viewing the print at a *narrower* angle than what the camera captured.

> oscarvdvelde
> even though it is unusual to have such close view of a face.

Well, it's not *exactly* unusual - it's just when you see a face that close, you're not necessarily thinking of photography ;-)

Link | Posted on May 24, 2018 at 15:20 UTC
Total: 430, showing: 1 – 20
« First‹ Previous12345Next ›Last »