34
\$\begingroup\$

What settings of my DSLR camera will emulate the scene exactly as I can see through my naked eyes?

I think it's not possible to get the contrast and colors exactly as my eyes can see, and it may vary from person to person. So, I am more interested in focal length of the lens. If anyone can give more insight beyond focal length, I will be glad to know.

For example, if I am standing on the seashore and want to capture the sunrise, what should be the focal length so that I can cover the angle of view which my eyes can see, and so the size of the objects in the photo will be precisely like my eyes perceive it?

My camera is an APS-C Canon EOS 1000D. I have 50mm 1.8 and 70-300mm Sigma , Can it be achieved through this equipment lens? Till now, I have not achieved or been satisfied with what I see and what I get.

\$\endgroup\$
12
  • \$\begingroup\$ Do you have a full frame (FX) of crop (DX, 1.6x) DSLR ? \$\endgroup\$ Feb 17, 2011 at 2:26
  • \$\begingroup\$ sure I will edit my question , forgot that factor. I have canon eos 1000D , which is not a full frame. crop factor of 1.6x \$\endgroup\$
    – sat
    Feb 17, 2011 at 2:30
  • 1
    \$\begingroup\$ What about the media intended to deliver the picture to other's eyes? Big or small? Screen or paper? ... \$\endgroup\$
    – mouviciel
    Feb 17, 2011 at 20:02
  • \$\begingroup\$ Interesting -- this ended up as a more general form of photo.stackexchange.com/questions/8043/… in the sense that our viewers' brains play a huge part in their perceptions of photos. Great answers here. \$\endgroup\$
    – D. Lambert
    Feb 17, 2011 at 21:22
  • 1
    \$\begingroup\$ @mouviviel -- Media is intended to deliver to my eyes after taking the picture , About , Screen and paper, If screen is calibrated properly I think we can expect almost the same thing on screen into Paper , Basically Screen and paper both :D \$\endgroup\$
    – sat
    Feb 18, 2011 at 3:17

9 Answers 9

44
\$\begingroup\$

Well, I hate to break it to you, but you can't exactly emulate your eyes. There's a few reasons, let me explain.

  1. Humans see much higher resolution in the central fovia (center part of our eyes) than near the edges. Cameras have uniform resolution everywhere.
  2. The dynamic range is handled differently between cameras and humans. I can't explain it, but a scene appears to have more dynamic range to a human than a camera, although technically a camera has more dynamic range.
  3. Humans see in 3 dimensions.
  4. Humans change their focal points very quickly, to the point that we don't actually notice the out of focus portions of most scenes.
  5. The shape of human vision is very different than a photograph. Photographs come out rectangular typically, with some dimension, human vision is more of a curved shape, although it is difficult to quantify by the way our brain manages the signals.

Notwithstanding all of that, let me just say that it depends if you are wanting to focus on a specific area, or on the larger scene around. If you want the specific area, you probably should go about 150mm or so. As for a dramatic landscape, something more like a 24 will get your entire field of view. A commonly cited number is 50mm, which will let you see the higher resolution portion of your eyes and then some, but not the entire field, and is usually a good compromise. (All of these assume you have a full framed camera, if yours is a crop sensor, please divide by the appropriate factor)

\$\endgroup\$
6
  • \$\begingroup\$ Ok , I was not satisfied with this situation , where I took a pic of sunrise using 50mm 1.8 , Size of the sun was much smaller in pic compared to what I saw in real , So I think I will have to adjust my focal length around 90-100mm to get the proper size. Thank you. \$\endgroup\$
    – sat
    Feb 18, 2011 at 3:29
  • 1
    \$\begingroup\$ Like I said, the eye has various regions of sensitivity. It's a bit of a pity you can't really get the FOV and the perception, it's really amazing what our eyes and mind can do, so... \$\endgroup\$ Feb 18, 2011 at 3:33
  • 1
    \$\begingroup\$ you forgot one: the human eye(s) has a different field of view from the image generated by most cameras (highly specialised cameras might exist). We don't see in 4/3 or 2/3 form factor. \$\endgroup\$
    – jwenting
    Mar 28, 2011 at 12:12
  • 1
    \$\begingroup\$ Also the image you "see" in your brain is basically run through a stack of 1000 filters, post-processors and a decent photoshop session. \$\endgroup\$
    – Hobbamok
    Aug 14, 2019 at 14:30
  • \$\begingroup\$ For the second point - basically, the human vision works like HDR. See here: cambridgeincolour.com/tutorials/… \$\endgroup\$
    – jng224
    Aug 11, 2020 at 10:10
41
\$\begingroup\$

The eye moves

I'd like to add to the answer of @Pearsonartphoto, that human eyes do not see a still scene, instead they scan it continuously, adjusting their “aperture” and refocusing as they scan it. So thanks to this scanning process we can perceive:

  • the higher dynamic range,
  • an infinite depth of field,
  • a wider scene as a whole,
  • much more detail than a “resolution” of the still eye can see.

So what we “see” is the result of “post-processing” of a stream of images, 30-70 images per second. See Eye movement.

The camera is still

The camera cannot simulate what the eye sees, because a large part of what the eye sees is our process of seeing and thinking about the scene, and a lot of it happens in the brain, not in the eye alone.

But you can take photos which produce a certain, desired impression in the person who sees them. And this is where the art begins, because you have to provoke certain feelings in another person by means of a flat coloured object.

Multiple image techniques

P.S. Some popular techniques in digital photography actually try to recover more from multiple sequential images as the human vision does. These techniques include:

  • panorama stitching (mosaicking multiple images to cover wider angle of view and increase resolution),
  • exposure fusion and HDR techniques (stacking multiple images to increase the dynamic range),
  • focus stacking (stacking multiple images to increase the depth of field) (more).
\$\endgroup\$
1
  • \$\begingroup\$ You have summed it up well, the best reply by far. \$\endgroup\$
    – labnut
    Apr 1, 2011 at 5:47
29
\$\begingroup\$

Capturing an image exactly as your eyes see it? No problem, but it's not all in-camera, you'll have to do some work in post. And it almost certainly not what you really want.

First, you'll need a circular fisheye lens that vignettes badly with which to take the picture(s). (You'll actually need several pictures and a viewing system with built-in eye tracking that can flip images as your eye moves.) Nobody makes a lens that works quite as badly as the human eye (at least, nobody makes one like that on purpose and offers it for sale as a photographic lens), so the rest will have to happen in post.

First, you'll need to cut off the image at the top. Exactly where will vary, but that's where your eyebrow gets in the way. Then you'll need to run an edge finder -- in Photoshop, the Sketch->Bas relief filter will do nicely. Put the result of that on a separate layer set to multiply -- you'll merge it down later, but you can hide the layer for now.

Next, select a section in the middle of the picture -- about one or two percent of the whole image -- feather the selection by twenty or thirty pixels, then copy that selection to a separate layer.

Now run a heavy blur filter over the original image. You want to be able to see the colours, but no real detail. Merge the bas relief layer down onto the blurred image. Then merge down the layer you made from the selection. You should now have a circular image with the top cut off, ill-defined edges, no real picture detail over most of the image, but with the edges of the original elements crying out for attention, and a small area in the middle (about ten percent of the width and height) in relatively sharp focus.

You're not done yet. You need to create a circular selection about half the size of the area that's in focus, just below and to one side, and overlapping the area of focus slightly. Feather the selection well, then fill it with black.

Create a new overlay layer, fill it with 50% grey and add about ten percent monochrome Gaussian noise, than give it a good blurring. Merge that down.

You now have an image that approximates what you actually see with your eye fixed. It's actually a lot less distorted than what your eye sees, but adding all of the little convolutions that happen because of the differing thickness of the fluids on the surface of the eye at any given time is a lot of trouble to go through for this exercise.

Anything else that you think you see is coming from your brain, not your eye. That black spot? You have one in each eye, and your brain fills in the missing data. The vast sea of detail? Panoramic stitching in the visual cortex. Sharpness of any sort? Averaging of multiple "exposures". We also have an automatic aperture system and do "ISO shifting" on the fly depending on what we're paying attention to at any given moment (our static dynamic range is not too terribly different from what the camera captures -- we do HDR "in post").

What you want is probably to capture what your mind sees, not what your eyes see, and that's different for every picture. That's why they make all of those different lenses. Sometimes you want to show a vast panorama because that's what you saw. But even with a huge landscape, it may have just been a little copse of trees that caught your attention. A portrait may be a tight close-up, or it may involve the sitter's entire environment -- again, it all depends what you saw in your mind's eye, not what you saw with your actual eye. The same thing goes for depth of field -- were you looking at the whole scene, or just paying attention to one small detail within a larger context?

In other words, this isn't a question that can be answered directly. The answer is, and always will be, "it depends". Experiment. Practice. Get to know your tools and what they do, and eventually you will know exactly what you need to do to capture not what was actually there, but what you saw. Ansel Adams called it previsualizing, and he spent a lot of time developing a formal system that would allow him to capture on film what he saw with his mind and his heart, not with his eyes.

\$\endgroup\$
4
  • 4
    \$\begingroup\$ While this perhaps starts out a little flip, it ends up being a beautiful and wonderful answer. Kudos. \$\endgroup\$
    – lindes
    Feb 17, 2011 at 19:53
  • 1
    \$\begingroup\$ Wow , I got lost reading the photoshop stuff then at the end it was good advice :) TY. \$\endgroup\$
    – sat
    Feb 18, 2011 at 3:23
  • 2
    \$\begingroup\$ What, no floaters?! \$\endgroup\$ May 19, 2015 at 20:57
  • \$\begingroup\$ The Photoshop process is an excellent description of what poor image quality the eye actually sees, fixed in place. But it's not fixed and the brain stitches it all together. \$\endgroup\$
    – doug
    Jan 28, 2019 at 5:45
12
\$\begingroup\$

One of the issues with trying to emulate human sight in a photograph is the field of view.

The perspective we see, which is a function of relative focal length is by most accounts roughly the perspective of a 50mm lens on a full frame sensor or 32mm on a DX sensor, but then the problem isn't perspective, it's field of view. Think of the picture you see at 50mm and now extend the field of view such as you would when taking a panorama.

If you took a "human" photograph, you would see almost 180 degrees horizontally and approximately 120 degrees vertically, yet still maintain the perspective of the medium focal length.

Take this crude diagram of the eye (green) and a digital SLR sensor (blue). You'll note that the focal length is exactly the same for both mediums, 17mm approximately, but the angle that the retina extends round to is much more than that of the sensor.

Eye field of view

It sees a larger field of view, with the same focal length. This is why a DX sensor equates to the field of view 1.6 times smaller than that of the 35mm sensor, yet at the same focal length, the perspective does not change. It simply captures a smaller area of the scene.

Panoramas are a way of emulating the field of view of something like the human eye whilst retaining the flatter and more realistic perspective.


The next issue is dynamic range. An average dynamic range of digital SLR sized sensors is roughly 11 stops of light. This means it can record the difference between 11 doublings in light intensity. 1, 2, 4, 8, 16, 32, etc. How accurate is another story. 14-bits is better than 12 and 12 is better than 8 bits, but analog is even better than 14-bit digital.

Whilst a full frame professional body capable of capturing up to and over 13 stops on a good day is considered impressive by modern standards, it doesn't even come close to the human eye.

The human eye is at times capable of distinguishing the difference between roughly 18 to 20 stops of intensity, in a very crude measurement. This means that the black shadow that your camera sees could be quite easily seen in detail by the human eye, at the same time as seeing bright details in the scene. This is where the dynamic range of a digital sensor currently falls down.

It simply cannot distinguish such wildly different light intensities at the same time. Sometimes it's so bad that you have to either expose for the highlights or expose for the shadows and suck it up, even when your own eyes can see both fine.

HDR is a way of emulating the dynamic range of the human eye, but is still limited by the mediums on which it is viewed as well as the way it is processed.


Another issue is that whilst it's a walk in the park for us as our brains are designed to see this way, only the fovea sees in great detail. Peripheral vision is rather undetailed, and is primarily there to see motion, which can help us identify that things are happening around us or warn us of danger and trigger the fight or flight response.

If you were to simulate this in a photograph, the image would have a small in focus area in the center and the image would quickly become blurred as you move toward the edges.


There are other issues which I have either not touched on or do not know about myself, but I think at the end of the day, a photograph is not really meant to "emulate" the human experience, it's meant to capture a moment, to create a response or an emotion, or to capture memories, or to get a big paycheque from a client :)

\$\endgroup\$
4
  • \$\begingroup\$ So u mean lens baby gives better or say real effect to what we see ? because it will focus at one point and other things blurred ? \$\endgroup\$
    – sat
    Feb 18, 2011 at 3:31
  • \$\begingroup\$ What do you mean? \$\endgroup\$ Feb 18, 2011 at 3:33
  • 1
    \$\begingroup\$ I suppose so. I guess the real question is whether you should or shouldn't do it. Pure simulation would mean you would blur the outer area of the photo, but if you managed to blow up final photo to fit your vision, then... your vision would take of the blur, If you know what I mean. \$\endgroup\$ Feb 18, 2011 at 3:37
  • \$\begingroup\$ ok. Will experiment a bit next time :) \$\endgroup\$
    – sat
    Feb 18, 2011 at 5:02
1
\$\begingroup\$

Your brain can adjust to the signals coming from the light sensing cells in your eye so that you can have a higher dynamic range from your eyes than is possible from your camera cells. Until there are intelligent camera sensors that can shut off one they have had enough light on them you will struggle to get the range that you need.

I would suggest making a HDR of a scene, but only pushing the under and over exposure slightly so it still looks natural.

\$\endgroup\$
1
  • \$\begingroup\$ Well, there is Highlight Tone Priority (as Canon calls it; probably also exists with different names for different brands), but it's not exactly what you're talking about here. I'd love a sensor that somehow had a logarithmic response curve per photo site after getting to a certain exposure... \$\endgroup\$
    – lindes
    Feb 17, 2011 at 20:14
1
\$\begingroup\$

Well, why not just put your 70-200mm lens on and zoom until the image on your lcd is the same size as your eye view without the camera. The lens will then read what focal length your at. Or just use the view finder. Close study of actual view, LCD view and Viewfinder view will get you there but you will need a tripod and at least 10 minutes. You may find your set up only excels at certain distances..i.e. the 3 views match up satisfactorily. At that point you can consider a lens purchase. Another thing after scale matching will be tonal matching to your eye. Cameras suck at seeing shadow generally. It seems a good idea to add a neutral darkening filter and add a few stops to compensate with the aim being to lighten the shadows while keeping the highlights tame. This might lead to the highlights not being so "blown out" while the shadows being completely black if the exposure was simply stepped up. A UV filter seems important too. After that one can adjust processing features in camera to tweak and make these settings default. I just mention here what I would try. Hope this helps.

\$\endgroup\$
1
  • \$\begingroup\$ The main problem with doing this is that how a picture looks is determined by a lot more than just focal length and colour, including viewing distance, media and viewing conditions. As has been mentioned in a few answers, a camera cannot replicate what is seen by the eye. \$\endgroup\$ Aug 24, 2015 at 5:57
0
\$\begingroup\$

Most images do well even if their perspective is incorrect, some images fare poorly. To view a photograph with the correct perspective:

View the camera’s image made the same size as the film / imaging chip, from a distance equal to the focal length setting. Such a viewpoint is likely not possible because modern cameras are petite thus the focal length used will be very short. Keep in mind that the closest the average person can read or view is 250mm (10 inches).

Because the modern camera makes tiny images we view them enlarged on our computer screen, TV, or as a print on paper. For these viewing methods we need to take into account the magnification applied to the final image.

Suppose an FX (full frame) 35mm size camera images the human face (portrait) taken with a 100mm lens. We enlarge this image and make an 8x10 inch print. The magnification applied will be about 8x.

The correct viewing distance is computed, 8 x 100 = 800mm = 31 ½ inches. This is about how we would view an 8x10 portrait print sitting on the mantel.

In short, an image will appear correct as to perspective if viewed from a distance about equal to the focal length multiplied by magnification used to make the displayed print.

Post Script: The only effect of focal length on perspective is to change the size of the image and the distance from which the displayed image is to be viewed.

The human perspective is simply what we see with our eye/brain combination. Standing before a glass window, you can draw, with wax pencil, the outline of objects you are observing. The perspective you observe is thus duplicated by the drawing. We are talking about the spatial relationships of the objects.

If a camera is used to image this vista, it is placed at about the same distance from the glass as your eye. The image it makes corresponds to the image drawn on the glass provided this image is viewed at zero magnification from a distance equal to the focal length of the taking lens.

If the viewed image is an enlargement, the viewing distance, to see this vista in correct perspective becomes the focal length multiplied by the degree of enlargement.

\$\endgroup\$
0
\$\begingroup\$

Photographic engineers stive to make camera and materials that produce a “faithful image”. This is an unachieved goal. We can do pretty good but no cigar. The real world abounds with light, color, and texture. If you think about it, it would be impossible to reproduce exactly such vistas.

As to focal length – you are asking about perspective. Paraphrasing a lecture, C.B. Neblette:

Facing a glass window, looking out at a vista, you can with wax pencil trace on the glass, an outline of objects you see. This trace is the “human perspective”. You can replicate this perspective with the camera by placing the camera where the eye was located and taking a picture. Matters not what size camera or focal length lens.

Next, we examine this picture, a contact image (same size as made by the camera, no enlargement), from a distance that is the same as the focal length of the taking lens. Follow these steps and you have reproduced the “human perspective”. The problem is, likely the camera chosen yields a small-scale image no suitable for viewing unless enlarged. Ok, we make an 8 x 12 inch print or slide or monitor displayed image and view. In making this enlargement you have magnified the original camera image 8x or perhaps 12x.

The fact that an enlargement was made changes the viewing distance that will allow you to perceive the “human perspective”. Suppose an APS-C camera was used with a 30mm focal length lens. To make the 8 x 12 you must enlarge 12x. Now the viewing distance becomes 30mm X 12 = 360mm = 14 inches.

In summary, it’s the viewing distance that key. Focal length times viewing distance time magnification yields, an image that reproduces the “human perspective”.

\$\endgroup\$
0
\$\begingroup\$

That question makes no sense, for example:

For example, if I am standing on the seashore and want to capture the sunrise, what should be the focal length so that I can cover the angle of view which my eyes can see, and so the size of the objects in the photo will be precisely like my eyes perceive it?

You are going to view any picture by using your eyes rather than by direct injection in your retina or optical nerve. So a picture that reflects the optics of an eye (a warped-area variable-resolution fish-eye thing with adjustable focus) will enter your vision by passing optics of an eye two times.

Things like white balance correction take a lot of pain of arranging a limited number of pigments in a manner reflecting the color impression of a human eye when exposed to pigments and lighting that actually are active over a whole continuous range of light wavelengths.

For example, a rainbow obviously is not constituted from lights of three primary colors but from a continuous range of wavelengths, but it can be emulated with a limited set of base colors (though they'll not manage the kind of saturation the actually pure colors exhibit) to the human eye.

Though a mantis shrimp would laugh at the kind of pathetic sketch human color vision can consider an acceptable substitute for the real thing.

Coming back to your question, the aim of photography is not to recreate what the human eye does, but to present to the human eye something that it considers a good replacement for the real thing.

Our media are plane surfaces, and for plane surfaces, a projection retaining straight lines straight tends to be a good fit for perception. Projection retains certain proportions of sizes for elements in different distances.

A commonly cited aesthetic is 50mm (in 135 format equivalent). That sort of gives a field of view that is somewhat related to the field of vision you have "at a glance" without particular attention or roaming looks.

It doesn't reflect the kind of detail you can see when focusing on a particular spot.

So in the end, you arrive at the situation that for various aspects of focusing and concentration, different photographic setups reflect a situation best in the manner you'd consider "what the eye does see". It is the photographer's job to render a scene accessible in a manner that feels like it matches your most captivating impression of a scene in reality.

And that is particularly important since scenes are three-dimensional, and even though our retinae aren't, what we can do with eye movement and body movement and focusing action explores a lot of the three dimensions that, in a photograph, has to be terminally flattened and represented in two dimensions.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.