Can we discuss perspective and perception of scale?
Moderators: rjlittlefield, ChrisR, Chris S., Pau
Can we discuss perspective and perception of scale?
You know those "model village" shots, done with tilt-shift lenses or post processing. An image of a cityscape or village, usually shot from a high vantage point, is given out-of-focus foreground and background with only the mid-ground sharp and fast fall-off to the OOF parts. Our eye/brain system is innately sensitive to the shallow DoF of close viewing so this fools us into thinking the image is a close-up of a small model. The people look like ants, the cars like toys etc.
So why (oh why, oh why) doesn't an image of a tiny, focus-stacked specimen, in perfect focus throughout, look like a normal view of a big thing? I have theories and practical experiences, but will save them for (hopefully) lively discussion.
Here's an example to get the ball rolling though (an image shot yesterday during more rig testing and tuning). Could this be coiled copper tube leaning against a concrete pillar in a scrap-metal refinery? It wasn't edited to look like that (so no it wouldn't), but attempts to do so would still fail to convince anyway! Perspective with no recession gives this one away instantly, then the apparently marble-sized dust particles, and that suspicious, global ambient light - no local effects going on. If that weren't enough, it looks just like the inside of a small electric motor up close - which it is. Too many clues prevent the "big thing" illusion taking hold - assuming there actually is one - maybe.
What can/do you discern about the scale of this subject from the image? I'll tell you it was shot in crop mode (APS-C 24mmx16mm) at between 2x and 14x on sensor. Can you narrow that down to a more specific magnification? More importantly (to me) why does it look like a small thing magnified, and never just a big thing photographed normally? If the "why" of that is knowable (and I think I know of several reasons), then can we move on to discuss the "how" part too? I'm far less confident I have any rational approaches left to try.
So why (oh why, oh why) doesn't an image of a tiny, focus-stacked specimen, in perfect focus throughout, look like a normal view of a big thing? I have theories and practical experiences, but will save them for (hopefully) lively discussion.
Here's an example to get the ball rolling though (an image shot yesterday during more rig testing and tuning). Could this be coiled copper tube leaning against a concrete pillar in a scrap-metal refinery? It wasn't edited to look like that (so no it wouldn't), but attempts to do so would still fail to convince anyway! Perspective with no recession gives this one away instantly, then the apparently marble-sized dust particles, and that suspicious, global ambient light - no local effects going on. If that weren't enough, it looks just like the inside of a small electric motor up close - which it is. Too many clues prevent the "big thing" illusion taking hold - assuming there actually is one - maybe.
What can/do you discern about the scale of this subject from the image? I'll tell you it was shot in crop mode (APS-C 24mmx16mm) at between 2x and 14x on sensor. Can you narrow that down to a more specific magnification? More importantly (to me) why does it look like a small thing magnified, and never just a big thing photographed normally? If the "why" of that is knowable (and I think I know of several reasons), then can we move on to discuss the "how" part too? I'm far less confident I have any rational approaches left to try.
- MarkSturtevant
- Posts: 1947
- Joined: Sat Nov 21, 2015 6:52 pm
- Location: Michigan, U.S.A.
- Contact:
Re: Can we discuss perspective and perception of scale?
So I have given absolutely no thought about this before, but why should that stop me?
One thing is if you could put tiny but realistic human figures into this scene, our minds' eye might try to make it look like a giant structure. I have another thought, but I will leave the fun for others.
![Very Happy :D](./images/smilies/icon_biggrin.gif)
One thing is if you could put tiny but realistic human figures into this scene, our minds' eye might try to make it look like a giant structure. I have another thought, but I will leave the fun for others.
Mark Sturtevant
Dept. of Still Waters
Dept. of Still Waters
-
- Posts: 1631
- Joined: Sat Apr 14, 2018 10:26 am
Re: Can we discuss perspective and perception of scale?
The coarseness of thr surface texture is giving some information here. Of course it's too easy for us to cheat since we recognize little things like where the focus stack ends or the unusual glow/reflections you tend to see in high mag scenes--these things are less likely for a random viewer to notice I'd think.
-
- Posts: 684
- Joined: Thu Dec 16, 2010 2:49 pm
- Location: Barcelona, more or less
Re: Can we discuss perspective and perception of scale?
I think the position of the camera is decisive. If you point down, we tend to read things as small. I try to photograph my minerals (when I can) in a "horizontal" view so that they look like landscapes. I always have in mind the movie "journey to the center of the earth" (by the way, horrible version of Verne's book).
- Attachments
-
-
- Posts: 112
- Joined: Thu Oct 13, 2011 3:56 pm
Re: Can we discuss perspective and perception of scale?
Member “Sager” played with perceived scale using slag glass...
viewtopic.php?f=27&t=32874&p=202793#p202793
https://jonathansager.com/Landscape-XVIII
He uses other minerals with the same scale-bending effect.
viewtopic.php?f=27&t=32874&p=202793#p202793
https://jonathansager.com/Landscape-XVIII
He uses other minerals with the same scale-bending effect.
Re: Can we discuss perspective and perception of scale?
Beatsy,
Interesting topic and important questions.
You ask " ...why does it look like a small thing magnified, and never just a big thing photographed normally?"
I think you almost offered the answer yourself: The instructions on how to make a real city look like a toy city.
Your photo of the rotor from an electric motor almost avoid these give-aways. But there is one subtle sign: the background along the upper edge about one third from left is very blurred.
There is nothing to catch the eye but unconsciously you perceive the unnatural sharp border between the large depth of sharpness (in the foreground and the middleground) and the very blurred background, which does not appear to be very far away. The stacking technique only makes this shift more unnatural.
Apart from this detail I think it would be very difficult to determine the scale of your picture. And apart from stating that it is a stacked macro, I am unable to get closer to the real size of your object.
Interesting topic and important questions.
You ask " ...why does it look like a small thing magnified, and never just a big thing photographed normally?"
I think you almost offered the answer yourself: The instructions on how to make a real city look like a toy city.
Your photo of the rotor from an electric motor almost avoid these give-aways. But there is one subtle sign: the background along the upper edge about one third from left is very blurred.
There is nothing to catch the eye but unconsciously you perceive the unnatural sharp border between the large depth of sharpness (in the foreground and the middleground) and the very blurred background, which does not appear to be very far away. The stacking technique only makes this shift more unnatural.
Apart from this detail I think it would be very difficult to determine the scale of your picture. And apart from stating that it is a stacked macro, I am unable to get closer to the real size of your object.
Troels Holm, biologist (retired), environmentalist, amateur photographer.
Visit my Flickr albums
Visit my Flickr albums
Re: Can we discuss perspective and perception of scale?
Interesting question, and one i have been wondering about too. Few years ago i tried to photograph circuit boards like cityscapes, and i didnt manage to. One of the less worse ones was this:
So i think familiarity plays a big part, and for your roll of copperwire, if you had told me it was the inside of a windturbine and two metres diameter, i would almost have believed you, mostly because i have no clue what the inside of a windturbine looks like. But most people are at least a bit familiar with insects, and they are small, everybody knows that....
Some theory i have about photos of big cityscapes looking small by adding a blurred foreground and background is that that used to be the way analog macrophotos all looked, no focusstacking possibilities, all shot on full frame, so small depth of field. If i now make a photo of something small with my phone, with its small sensor, the depth of field seems bigger. My ten year old niece didnt understand why a photo with a blurred front and back was supposed to look like something small, she just did not see it. Statistically this is n = 1 so I should test this with more children.
I look forward to more ideas, it would be cool to let small things look convincingly big.
For everyone who has ever opened a computer of videoplayer or something like that this is recognisable as something about ten centimeter wide. And to me it looks completely unconvincing, but people that had never seen something like this were wondering where in Rotterdam harbour i found this that winter ![Think :-k](./images/smilies/eusa_think.gif)
Some theory i have about photos of big cityscapes looking small by adding a blurred foreground and background is that that used to be the way analog macrophotos all looked, no focusstacking possibilities, all shot on full frame, so small depth of field. If i now make a photo of something small with my phone, with its small sensor, the depth of field seems bigger. My ten year old niece didnt understand why a photo with a blurred front and back was supposed to look like something small, she just did not see it. Statistically this is n = 1 so I should test this with more children.
I look forward to more ideas, it would be cool to let small things look convincingly big.
- enricosavazzi
- Posts: 1475
- Joined: Sat Nov 21, 2009 2:41 pm
- Location: Västerås, Sweden
- Contact:
Re: Can we discuss perspective and perception of scale?
I am sure the following is already mentioned, or implicitly assumed, in several of the above posts, but I think the main reason why actual landscapes shot to resemble miniatures are startling is because they contain easily recognizable large objects (cars, trains, buildings) while providing the typical optical rendering and perspective of macro and close-up shots. Enhanced-DOF macro shots typically contain no recognizable cues that suggests the subject is on the scale of a landscape. This might change if you add miniature vehicles, people and buildings in a close-up subject, either in post-processing or as credible physical miniatures. The latter may be difficult/expensive to find or build, and cheap commercial miniatures display many giveaways that spoil the effect, including plastic mold lines, thick paint layers, lack of sufficient detail, and obviously artificial texture and translucency of the used materials.
Years ago I did http://www.photomacrography.net/forum/v ... 60#p118760 . One takeaway is that it is easier to get fine detail in natural subjects than artificial ones.
Years ago I did http://www.photomacrography.net/forum/v ... 60#p118760 . One takeaway is that it is easier to get fine detail in natural subjects than artificial ones.
--ES
Re: Can we discuss perspective and perception of scale?
Thanks for your thoughts and ideas to date folks. All interesting. I think they already show (collectively) that there are lots of factors that contribute, and a continuum of "effectiveness" with each aspect. That's what I've found to date too.
The way the "model village" effect grabs you is super-effective though. It *makes* you see ant-sized people and model cars. At the other end of the scale, the one I'm pursuing, there doesn't seem to be a mechanism to do that. The best I've found so far can only make the sense of scale disappear (as do the suggestions and examples on this thread - so far). There's nothing to *make* you see people-sized ants, or model cars as real ones in any macro-scale shot.
I think *the* main confounding factor is perspective and related perspective distortion. Imagine photographing a real-world model of an ant that stood 20-feet high to the shoulder. The brief is close-up, detailed shot with the ant looming over the (human-sized) viewer. So you take a 12mm UWA lens, walk right up to one of the front legs, tilt the camera up (portrait orientation) so the head and whole body are in the frame too. Click. You now have a composition with a foot filling the bottom third of the frame, a big fat shin in front of you, a head in the top third (apparently smaller than the foot due to perspective) and the rest of the body receding into the distance (perspective again). The *key* thing is your perspective of the individual parts. In that single image, you are looking down on the foot, straight at the leg and up under the chin. You'll never get that from a single, straight-on, macro-scale shot - or stack. That's why I was asking about stack and stitch by rotating around no-parallax points recently - in an attempt to get that. It has too many practical difficulties to pursue further at this stage. And I still don't know if it would work anyway.
I'll leave it at that for now, but thanks again for commenting and I hope we can continue...
Cheers
The way the "model village" effect grabs you is super-effective though. It *makes* you see ant-sized people and model cars. At the other end of the scale, the one I'm pursuing, there doesn't seem to be a mechanism to do that. The best I've found so far can only make the sense of scale disappear (as do the suggestions and examples on this thread - so far). There's nothing to *make* you see people-sized ants, or model cars as real ones in any macro-scale shot.
I think *the* main confounding factor is perspective and related perspective distortion. Imagine photographing a real-world model of an ant that stood 20-feet high to the shoulder. The brief is close-up, detailed shot with the ant looming over the (human-sized) viewer. So you take a 12mm UWA lens, walk right up to one of the front legs, tilt the camera up (portrait orientation) so the head and whole body are in the frame too. Click. You now have a composition with a foot filling the bottom third of the frame, a big fat shin in front of you, a head in the top third (apparently smaller than the foot due to perspective) and the rest of the body receding into the distance (perspective again). The *key* thing is your perspective of the individual parts. In that single image, you are looking down on the foot, straight at the leg and up under the chin. You'll never get that from a single, straight-on, macro-scale shot - or stack. That's why I was asking about stack and stitch by rotating around no-parallax points recently - in an attempt to get that. It has too many practical difficulties to pursue further at this stage. And I still don't know if it would work anyway.
I'll leave it at that for now, but thanks again for commenting and I hope we can continue...
Cheers
- rjlittlefield
- Site Admin
- Posts: 23603
- Joined: Tue Aug 01, 2006 8:34 am
- Location: Richland, Washington State, USA
- Contact:
Re: Can we discuss perspective and perception of scale?
The first time I saw one, it certainly did that. But lately, I notice that one of my first thoughts is "Oh, somebody tipped the lens".
To be honest, it's a little disappointing to lose the illusion. Sort of like understanding how magic acts actually work.
Anyway, I think we're all pretty much agreed that you can't make something look BIG without adding elements that are immediately recognizable, and if they're immediately recognizable as small, then overriding that impression with something like extreme perspective. "OK, that's clearly a fly, but that's his foot immediately in front of me, and that's his head way off in the distance -- dang, that thing must be HUGE!"
So, I spent some time thinking about exaggerated perspective.
My first thought was to use some existing tool to "cheat", like by shooting an ordinary stack and rendering it with background frames systematically reduced in size. But that doesn't work very well, because any scale change big enough to force perspective breaks alignment between stack frames so badly that it introduces nasty artifacts.
Then my thought was to optically force the entrance pupil to be very close to the subject, with a wide angle of view. That starts to work, but it runs into a limitation that with a small subject I need wide entrance cones to maintain sharpness, and the combination of wide entrance cones and a wide angle of view is not provided by any optics I can dream up.
Since both those ideas have not worked, my current thinking is that the best hope is through photorealistic CGI. Shoot the subject with ordinary stacking that gives a far-away perspective, use those images to build a 3D model of the subject, then render the model from a close-up viewpoint. This is not a new idea. See https://sketchfab.com/3d-models/dinomyr ... 724ec93036 for example.
--Rik
Re: Can we discuss perspective and perception of scale?
Two tricks may help
One was mentioned by Rik, having a near, wide angle perspective, so we feel "inside". That is difficult, but doable up to some point, like with a small fisheye, a pinhole lens, or a high NA microscope objective imaging in aperture.
Another big trick that fools the eye is "aerial perspective", as used by painters. Objects far away (mountains, skyscrapers) have lower contrast, because there is more fog/haze, and we unconsciously use that as hint for distance evaluation.
The effect can be simulated in macro with thick smoke, or with the object immersed in dirty water, or in post-processing. Aerial perspective haze is used routinely and heavily in CGI videogames and movies, together with the wide angle and the "large recognizable objects" tricks. In CGI there is no "real" scale size, but the designers have found ways to make us perceive the right scale.
One was mentioned by Rik, having a near, wide angle perspective, so we feel "inside". That is difficult, but doable up to some point, like with a small fisheye, a pinhole lens, or a high NA microscope objective imaging in aperture.
Another big trick that fools the eye is "aerial perspective", as used by painters. Objects far away (mountains, skyscrapers) have lower contrast, because there is more fog/haze, and we unconsciously use that as hint for distance evaluation.
The effect can be simulated in macro with thick smoke, or with the object immersed in dirty water, or in post-processing. Aerial perspective haze is used routinely and heavily in CGI videogames and movies, together with the wide angle and the "large recognizable objects" tricks. In CGI there is no "real" scale size, but the designers have found ways to make us perceive the right scale.
Re: Can we discuss perspective and perception of scale?
Just to add another data point. I find that stuffing the frame full of subject helps. Really tight framing. It doesn't make a small thing "look big" per-se, but it does make the subject "loom" (especially if shot from a low-looking vantage point like the crystal example earlier). Giving the viewer a good sense of "being in there" is perhaps a better way of putting it.
Here's a couple of test stacks, doodling with the idea. These really are the studio-macro equivalent of "snaps". My current setup makes it quick and low-effort to knock out large, deep stacks these days (so I do it even more often). All I need to do is find and mount the subject, add a background (or none as in the first pic), rack focus and find a composition that fits, lower the "standard shader" and twiddle the lights until it all looks "nice", then start the capture process. The rest (capture, import, conversion and focus stacking - no retouch) pretty much runs itself. I just tweak exposure on the outputs when I get around to looking at them. Both stacks shot at 3.4x with a 5x mitty. Second was only a partial stack - intentionally.
Edit: both of these "loom" far better when viewing larger, full-screen images. You may need to use a bit of imagination with these diddlers...
Here's a couple of test stacks, doodling with the idea. These really are the studio-macro equivalent of "snaps". My current setup makes it quick and low-effort to knock out large, deep stacks these days (so I do it even more often). All I need to do is find and mount the subject, add a background (or none as in the first pic), rack focus and find a composition that fits, lower the "standard shader" and twiddle the lights until it all looks "nice", then start the capture process. The rest (capture, import, conversion and focus stacking - no retouch) pretty much runs itself. I just tweak exposure on the outputs when I get around to looking at them. Both stacks shot at 3.4x with a 5x mitty. Second was only a partial stack - intentionally.
Edit: both of these "loom" far better when viewing larger, full-screen images. You may need to use a bit of imagination with these diddlers...