46

Having just spent some time reading questions about auto-focus, I came up with a question of my own: How did cameras even work before auto-focus technology was invented?

Presumably everybody used manual focus. But here's the thing: I've tried manually focusing my DSLR. It's absurdly hard. Given how utterly tiny the image in the viewfinder is, I have no idea how you'd ever get an image in immaculate focus.

Or maybe people didn't? Maybe back before 24 megapixel images enlarged to fit on the side of a bus, focus wasn't quite so critical? Certainly if you print something out the size of a postcard, focus errors are going to be a heck of a lot less noticeable.

Also, my very first camera was a Fisher-Price "toy" camera. (Film, obviously.) I'm pretty sure it didn't have any focusing controls at all. (And this is way too long ago for auto-focus to have existed.) How does that work? Is the lens just permanently focused at infinity or something?

  • 14
    "back before 24 megapixel images" go ahead and properly scan some sheet film. I'm not sure if digital wins over film when it comes to maximum possible printing size. Digital sensors are just too expensive in large sizes. – null Sep 3 '16 at 14:35
  • 1
    Good drum scan of a negative yields 100MPx files. But that's 100M of real, 3 color pixels, so a Bayer filtered sensor would have to have 300MPx to match 35mm film. Not even mentioning larger formats. BTW, postcard at arms length is exactly same as a billboard across the street. Usually billboards are pretty lowres compared to your postcards. – Agent_L Sep 5 '16 at 14:00
  • 4
    That's not how it works Agent_L. 1 100MP drum scan would be the same number of pixels as any other 100MP image. And you aren't getting past about 2400dpi on 35mm film unless you want grain the size of golfballs on your prints. – Matthew Whited Sep 5 '16 at 20:12
  • 1
    It's really incredible how the less-informed keep upvoting the generic crap answers. I also have noticed that one of the things that none of the answers below address is that back in the day, manufacturers concentrated on putting out wider-aperture lenses, such as the Noctilux, f/1.0 , or even the Nikkor 50mm f/1.4, lenses which provide a – Cascabel Sep 6 '16 at 5:17
  • 4
    @Agent_L if I have a microscope and some patience I can pull a gigapixel off of the same negative, but that doesn't mean that there's any more genuine information there than there would be at 30MP. – hobbs Sep 6 '16 at 20:40
84

Through-the-lens focusing cameras had focusing screens — usually ground glass or fresnel lens (related: What is a focusing screen?). View cameras (the old-style large cameras with bellows) projected the image onto the focus screen. The photographer directly inspected the image on the focusing screen (perhaps using a loupe to magnify areas of the image), often under a blackout hood. When it came time to capture the image, the focusing screen was replaced by the film holder, at which time the film could be exposed.

image on focus screen of Sinar F 4×5
Image on focus screen of Sinar F 4×5. Image © Guillaume Piolle, CC-BY-SA-3.0

Split-prism focus screens were the most common on SLRs. The prism pattern in the center of the screen purposely "bent" the out-of-focus rays in opposite directions, magnifying the visibility of out-of-focus areas. Most split-prism screens also had a micro-prism ring or collar surrounding the split-prism circle in the center of the screen. The micro-prisms especially helped for the fine focus adjustment near in-focus. This 5 minute YouTube video demonstrates the use and effect of both the split prism and the micro-prism ring.

Below is an example of the split-prism focusing. The ring of micro-prism "stippling" surrounding the split-prism is also visible in the unfocused image:

SLR split prism, unfocused SLP split prism, focused
SLR split prism focusing screen images, unfocused (left) and focused (right). Images © Dave Fischer, CC-BY-SA-3.0.

When the objects are no longer misaligned, then the lens is perfectly focused on the objects.

The corollary to being in focus is being able to determine the distance to the subject. This same technique was also used for optical range finding (not rangefinder cameras). Split-prism coincidence rangefinders were used in militaries to determine distance to target, in order to set artillery coordinates and propellant charges:

View from a coincidence rangefinder.
View from a coincidence rangefinder. Public domain, from Wikimedia Commons

Rangefinder cameras used a separate optical path for focusing, the range-finding focusing mechanism. This showed two overlaid images. When the images were perfectly overlaid, the subject was in focus. This example from Wikipedia illustrates the concept:

Rangefinder camera window, unfocused (left) and focused (right)
Rangefinder camera window, unfocused (left) and focused (right). Image © Alexander Koslov, CC-BY-SA-3.0

  • 12
    Interesting. Is that why when TV does a fake "I'm looking through a camera" shot, it has that circle bit in the middle? I always wondered what that was meant to be... – MathematicalOrchid Sep 3 '16 at 15:16
  • 19
    Indeed. Much like the floppy disk is used as the "Save" button icon in applications long after the floppy disk and floppy drive ceased being available in computers, some things are just, well, iconic. – scottbb Sep 3 '16 at 15:27
  • 1
    Same “effect” in that each element shows light from one side of the lens or another. It is literally the same thing as the center elements, only smaller. – JDługosz Sep 3 '16 at 22:35
  • 3
    It's possible to retrofit split prism screens to some DSLRs. – Chris H Sep 4 '16 at 8:30
  • 1
    Not only is it possible to retrofit, there are actually companies still making them. The company I bought mine from is no longer accepting orders, but there are other suppliers out there that sell similar products. About the only requirement is that the camera body allows focusing screen replacement, and ideally that you have reasonably bright (f/4 or so, or brighter) glass. – a CVn Sep 5 '16 at 11:19
25

How did photography work before auto-focus was invented?

Pretty well for those willing to learn how to do it with the tools we had at at the time. The same is true now. The only difference is that now we must learn how to tell an AF system to focus on the part of the frame we want it to bring into focus.

Presumably everybody used manual focus. But here's the thing: I've tried manually focusing my DSLR. It's absurdly hard. Given how utterly tiny the image in the viewfinder is, I have no idea how you'd ever get an image in immaculate focus.

Due to the ubiquity of AF in most modern cameras, the focusing aids that were once included in what a photographer saw through the viewfinder are usually no longer present. Split prisms and/or prism collar micro screens were common in SLRs before AF came along. Some cameras had one or the other. Many cameras had both. Other types of cameras often incorporated a parallax rangefinder type of focusing aid.

Viewfinders were generally larger and brighter as well, even on consumer grade cameras. Now only the top pro models tend to have large, bright viewfinders that were more commonplace in the pre-AF era.

Lenses were also designed to allow finer gradations of focus adjustment. Focus rings on lenses had to be rotated much further to get the same change in focus position that now results from a very small movement with current lenses.

Or maybe people didn't? Maybe back before 24 megapixel images enlarged to fit on the side of a bus, focus wasn't quite so critical? Certainly if you print something out the size of a postcard, focus errors are going to be a heck of a lot less noticeable.

There is some truth to that for casual photographs, which is what the vast majority of photographs are. But there were (and still are) also large and medium format photographers that went to great pains to produce manually focused images suitable for display at very large sizes.

Some of the limits of sharpest focus were imposed by the recording medium. Part of the problem with roll film is that it doesn't like to sit flat in the camera. It's similar to the problem you'll have if you ever try to perfectly focus a projector onto a flexible portable screen flapping in the breeze. There were a few advanced cameras that actually used a type of vacuum to pull the film flatter against the back plate.

Focus with color film was also limited by the varying depth of the three color layers in film. If you were perfectly focused for one color, the other two layers were ever so slightly out of focus.

Digital sensors, in contrast, are so near perfectly flat that now we have to coat the back surfaces of lens elements to prevent unwanted reflections from bouncing off the layers of the sensor stack. The theoretical limits of best focus are now much smaller. Even with newer and much sharper lenses available today the limiting factor in the highest resolution systems is fast becoming the resolving power of the lens rather than the resolving power and flatness of the recording medium.

Also, my very first camera was a Fisher-Price "toy" camera. (Film, obviously.) I'm pretty sure it didn't have any focusing controls at all. (And this is way too long ago for autofocus to have existed.) How does that work? Is the lens just permanently focused at infinity or something?

"Fixed focus" cameras employ relatively narrow apertures combined with relatively short focal lengths. This gives a large depth of field. The focus is set at the system's hyperfocal distance so that anything from half that distance to infinity appear to be in focus. Cameras with this type of design can still be found. Among them are some (but far from all) webcams, cell phone cameras, trail cameras and other surveillance cameras (although they don't make up near as many of the cell phone cameras as they did a few years ago).

Some cameras had no way to see through the lens prior to taking the shot. A cursory viewfinder was attached to the side of the camera or, if you had a deluxe model, to the front standard that held the lens. No accommodation was made with such viewfinders for lenses of varying focal length. The photographer just had to know how wide the angle of view for the lens being used was. Focus was set by estimating or measuring the subject distance and lining up a mark on the lens to a scale with that distance. Aperture and shutter speed were also set manually with no metering built into the camera. Some of those cameras used roll film with anywhere from 6-12 shots per roll. Others used sheet film that had to be changed after every exposure. The latest versions of such press cameras were still being used by local press photographers in my hometown as late as the early 1970s.

WeeGee

  • 1
    How does the thickness of different color layers in film compare to the achromatic (doublet) lenses used in most cameras still having some chromatic abberation? Wouldn’t the layers be stacked in order to work with the difference in focus of different wavelengths, thus giving better chromatic abberation thanna single sensor plane? – JDługosz Sep 3 '16 at 22:45
  • 2
    There were some lenses that attempted to do that, but no one ever figured it out well enough to make a difference. CA tends to be spherical rather than flat field and the state of the technology at the time with regard to achromatic corrective elements was nowhere near what has been done since. Simulating design changes within hours using supercomputers instead of having to wait months to do it with physical prototypes has revolutionized what is possible with lens design and manufacture. The problem with the flexibility of film coming off a tightly wound roll didn't help any, either. – Michael C Sep 4 '16 at 5:01
  • Does this mean that if I bought a more expensive camera body, it would have a larger viewfinder? It's sometimes really hard to tell which bit of the picture the AF system has actually focused on... – MathematicalOrchid Sep 7 '16 at 9:25
  • In general the higher priced DSLRs have larger viewfinders. But AF systems are another subject entirely. If you are depending on the information you see in the viewfinder without understanding the actual map of your camera's AF system you are doomed to fail re:AF. photo.stackexchange.com/a/61597/15871 – Michael C Sep 7 '16 at 11:00
  • Ugh! I can smell the stench of that cigar just by looking at that picture. – Olin Lathrop Jun 24 '18 at 12:10
11

There are diferent topics here.

1) A toy camera, and some new cameras, for example survilance cameras, some phone cameras do not need to focus because its focus range is very extense. Normally this is due two elements combined. A wide angle lens, and a small aperture. So there is no need to focus at their designed range.

Try to have in focus a very close range on a toy camera and you can not.

2) The size of a image on a bus is not 24 Mpx. It can be more or less. That is a different topic, but before "digital" photography you had commercial film photography that could use large format film that captured quite high resolution images. For this were not 35 mm slides, but 6×6 cm 4×5″ etc. https://www.google.com/search?q=film+photography+formats

Inclusive now you can "squeeze" a lot of information from a 35 mm film on scan.

Scanning a big color slide with a drum scanner has a lot more information than 24 Mpx. https://www.google.com/search?q=drum+scanner

3) The cameras had a focusing screen. In comercial photography you used a magnyfing glass.

4) And they had another feature, the focusing ring extended for a loooooot more angular turn. This means that on modern lenses on manual focus, a little turn on the lens moves a lot the focusing distance. On old lenses you need to move a lot more the ring giving you more prescition on the focus.

5) People were experienced doing that, finding and feeling the focus moving left and right feeling the center of the focusing "ballet".

6) You could still use manual focus this days on a DSLR camera, for example when you shoot video. As I comented on point 4, if your phisical angular turn of the focusing ring is bigger you have more control. So in video people use some rigs that adapts to the focusing ring. https://www.google.com/search?q=focus+rig+dslr.

  • 2
    We also were younger, and had better vision! :D I wouldn't want to try to manually focus now. – DocPixel Sep 4 '16 at 0:59
  • "Try to have in focus a very close range on a toy camera and you can not." -- indeed, I recently tried photographing some documents with the fixed-focus camera on my Samsung Galaxy Young. It was impossible to get them both in-focus and with the text at a legible size. – Periata Breatta Sep 5 '16 at 0:32
  • "some phone cameras do not need to focus because its focus range is very extense" Surely you mean that their depth of field is large. The focus range would, at least to me, be the possible difference in positioning of the focal plane of the image. All cameras need a focusing mechanism, but not all cameras necessarily need (for their use case) for that focusing mechanism to be adjustable. – a CVn Sep 5 '16 at 11:23
  • Yeap. That is what I ment with "designed range". The range is normally for landscapes (Let us say "infinite" to a "selfy"). – Rafael Sep 5 '16 at 13:23
  • Re: #6. It depends on what camera you are using. The introduction of Dual Pixel AF and STM lenses combined with touch screens used to designate the desired point of focus has given a lot of DSLR video shooters the ability to let the DSLR do the focusing now. – Michael C Sep 6 '16 at 5:45
8

The current answers explore manual through-the-lens focussing and fixed focus cameras, and do a good job of explaining them, but they miss another approach - distance estimation.

For example, my old 1960 Kodak Retinette is not an SLR, and there is no through-the-lens focussing. However, it is not a fixed focus either.

Kodak Retinette 1A

Instead, you estimate how far away the subject is, and use the distance markings on the focus ring to match. (The image above has it measured in metres, my actual camera measures it in feet.)

  • 2
    I meant to touch on distance estimation in my answer, so I'm glad you did. – scottbb Sep 6 '16 at 3:40
  • I wonder how accurate your "estimate" needs to be? Judging by the spacing of the marks, it looks like if your subject is more than about 20 feet away, you don't need to worry too much... – MathematicalOrchid Sep 6 '16 at 8:04
  • @MathematicalOrchid: Well, that depends on the depth-of-field.(The narrower the depth of field, the more accurate you need to be.) The depth of field depends on the aperture. The outer ring conveniently shows that. For each f-stop, there are two marks, showing the upper and lower limit (in distance) of the nominal depth of field at that aperture for that focus point. [Note: metres, not feet.] – Oddthinking Sep 7 '16 at 0:21
8

To suppliment the answers on focusing aids in SLR cameras meant to be manually focused, let me give you a link to hyperfocal distance which is used in fixed focus cameras.

Kodak Instamatic

My parents’ (later handed down to me) looked like this, so it was an Instamatic X-15 circa 1970.

The first picure is a higher-end model that has a built-in light meter (so I guess the exposure was not fixed but based on light) and a switch for close-up or 6 feet and farther focus.

Instamatic x15

Notice that the aperature is physically very small. It is f/11, and the negative is 28mm square, so the depth of field is a little better than you get from a “full frame” and not as good as you get from a “crop sensor” today if you set it to f/11.

Given the best sharpness you could expect from the system and the assumption that you would be looking at prints 4 inches square held at reading distance, everything past 6 feet was “in focus”. But the focus isn’t as sharp as what you expect today from the same values, because the expected quality of the print was not as good for numerous reasons. 1970 color print film is not Ektar!

The point is that the range of distance considered to be “in focus” is a judgement based on the physical size of the circles of confusion you can tolerate. With coarse grained film (and print paper) you can’t get past a certain sharpness anyway; the resolution simply isn’t good enough.

4

The simplest camera you can have, something you can easily build at home, is just a cardboard box with a pinhole. You'll get a focused image right on the film (albeit inverted). This also works with human eyes - if you're near-sighted and don't have your glasses, you can get a perfectly focused image just by forming your fist in front of your eye to let just a tiny "circle" of light through.

This should make perfect sense - being out of focus means just that you get a lot more light than you want, in all the wrong places. This is basically due to the fact that in a typical scenario, light is scattered and diffused all over the place, so the same image is effectively projected from many different angles that come together "wrong". A DSLR solves this by using lens to "recombine" the dispersed light rays - making them parallel again. However, this only works for a range of distances. In our shoebox-with-a-pinhole camera, the problem is solved by eliminating all the "other" light that doesn't come directly from the object. This doesn't depend on the distance to the object (though there are limits with distance versus the radius of the pinhole, and there are other limits as well), but it receives less light overall, so you need a longer exposure. Also, you usually have a pretty hard limit on minimal distance - they have great focus anywhere from e.g. 3 meters to infinity, but you're not going to be able to photograph butterflies :)

As noted, DSLRs are a lot more complex than this - they're heavily built around being able to tweak many different settings to suit whatever fancy you have, and whatever situation you're in (in an ideal world :P). However, while it takes some practice to get manual focusing right, it certainly isn't absurdly hard. If you're routinely having trouble with manual focusing, try calibrating the view finder (there should be a "gear" somewhere nearby with a +/-) - your eyes might be the problem.

2

We made an awful lot of bad photos. Shooting candids, I figured that between bad exposure and bad focus that 60% printable shots was par for the course. Auto exposure cut the rejects in half. AF came in just in time to fix my decreasing aging vision. Now, I figure that unless I do something stupid, then rejects for focus/exposure will be only a few percent, and most of those will be weird lighting -- backlit, snowscapes, night pix... Or situations where the autofocus gets confused by some object between me and my subject. (The twig is in focus, the bird is not.)

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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