How Deep is your Field?

Pictures to be added to this one later. Published as is for proofreading purposes.

During a conversation at work today, we were discussing how people are all seemingly buying DSLRs at the moment, whether they're actually interested in photography or not. In the ensuing chat we were talking about what are the biggest advantages of a DSLR over a digi compact, and then we moved onto more general photography related matters, including about what difference it makes where you focus in a picture when taking it - whether it makes any difference whether you focus on whatever's in the centre, your main subject, what's behind it, etc. It got me thinking that there's probably a whole collection of people out there whose photography has always been very much point and shoot with, most likely, a phone.

Now phones have come on a long way since the first camera phones came out. My first camera phone was a Nokia 6100 with a clip on camera on a wire. There was no viewfinder, the screen was tiny, the resolution abysmal (640x480 I think), but at the time I was working in an office where cameras were banned, so it worked for me as I could leave the camera part at home. Modern phone cameras are a world apart from these early beginnings. My phone, which is by no means the current state of the art, shoots pictures which have 13 megapixel resolution. It has adjustable white balance, ISO, flash control, scene modes, sweep panoramas and 1080p video. It's really pretty good, and much better on image quality than some older digital compacts, but it does have one really limiting factor in common with pretty much all phones and most low and middle end digital compacts - a tiny sensor. This makes a difference to quality in that there have to be more photosites (sensor elements) crammed into a tiny space, so they have to be smaller and therefore pick up less light and have to be amplified more, but that's not my concern with this post. My concern is the fact that with a tiny sensor you need a tiny length of lens and with a tiny short lens you cannot achieve shallow depth of field.

Depth of Field Defined

I've touched on depth of field before - one of my earliest posts on this blog (Not the length of the grass) was a mini discussion of the concept, but I re-read that today and noticed it's not exactly in-depth, if you'll excuse the pun. I'll attempt to rectify that here.

You can imagine depth of field as two points on a line extending straight out in front of you from the camera. Between those two points is the actual single "plane of focus" - you are after all focused on only one point in the image so the only completely sharp parts of the image will be things on the same subject distance plane as the subject itself. Anything in the "sharp zone" between these two points will appear "acceptably sharp" on the sensor and in any resultant image file. As you go outside of these two points the image will appear softer or more blurred, with the effect increasing as you get further away from the actual focussed point, whether that is towards the camera or away from it. There are various things you can do that will move the points on this line, and various features of the camera and lens that will limit the effect you can have. Listing the things you CAN change first:

You can change the aperture on the lens. This is the big advantage of a middle or high end compact or a DSLR over a low end compact (or pretty much any film compact other than a few specialist exceptions). Aperture has been discussed before in this blog, but is basically the diameter of the hole in the lens through which the light is able to pass. Making that hole larger by making the aperture number smaller (f/5.6 --> f/4 --> f/2.8 for example) will move the two points defining the sharp zone closer together, and reduce the amount that's sharp. Making it smaller by making the aperture number larger (also known as "Stopping Down" - f/5.8 --> f/8 --> f/11 for example) will move the two points further apart and increase what's sharp in the image.

You can change the length of the lens by zooming in or out, or by physically changing lens. A shorter lens will cause the points to move further apart, and provide more depth of field at a given subject distance. Of course it will also change the perspective of the shot so you'll then need to move and recompose to get a similar subject distance. A longer lens will move the points closer together and reduce depth of field, making less of the scene sharp in the image.

You can change your distance to the subject. The further away you stand, the further apart those sharp-zone points get, and the closer you get to the subject, the smaller the sharp zone becomes. It's actually not as simple as this when changing multiple variables simultaneously as explained here but for the purposes of this exercise it can be considered to be a truth for the moment.

The limiting factor that you can't do anything about other than change camera (unless you have a roll-film camera with different film masks available, or an XPan, or a sheet film camera - but that's one for the pedants...) is the size of the recording medium. This may be a sensor of one of myriad different sizes, or it may be film. For the purposes of this it makes no difference which it is - it's only the size that matters. It's not so much that a small sensor actually directly affects the depth of field - it's more about the lenses required for that sensor.

Short Lenses

For a given sensor size there is one particular length of lens that gives a 'normal' perspective, and wide angle or telephoto lenses are either side of that length. Going back to my own phone, according to Geekaphone it has a 4mm, f/2.4 lens and a 4.8mm x 3.6mm sensor. When you start up the camera it's set to a slight wide-angle (around 30mm equivalent on a 35mm format sensor or film) and assuming we ignore the fact that digital zooms are useless, the digital zoom can be used to get to 'standard', so we can assume probably that the "standard" lens for this sensor size would be around 6mm. We've said that a shorter lens causes larger depth of field. Now assuming we had this camera/phone with a 6mm fixed lens on it, the field of view (how much is in the shot) would be normal for this tiny sensor, but if you pointed it at a 35mm size sensor, this would be an ultra-wideangle lens. With such a short lens the depth of field would be enormous, and remains enormous even on this tiny sensor. As such, it's very difficult to achieve shallow depth of field (sharp zone points close together) on such tiny sensors, so you have to go to extremes on the other variables to overcome this - physically close in on the subject and use the widest aperture available - and you'll still be at a disadvantage over something with a larger sensor.

Long Lenses

We said that long lenses reduce depth of field, so again let's consider the phone camera. A telephoto lens might be, say, four times the length of a "standard" lens. On a 35mm format camera a standard is 50mm and a 200mm is very much a telephoto. On the phone, if 6mm is standard, then 24mm would be telephoto. This is a wide-angle lens on 35mm format and will still have very limited depth of field. A 200mm lens attached to the same sensor would be the equivalent of a 33x zoom, giving us an equivalent focal length of over 1500mm. This is exceptionally long, and would have the same depth of field as this relatively modest telephoto would on 35mm.

It should be added at this point that all a digital zoom on a camera or phone does is crop in on the centre of the image and interpolate enough pixels to make the image up to full size. As such, it does not affect sharpness / depth of field in any true optical way, but as the image is magnified, any loss of sharpness becomes more obvious. I've not tested it side by side with a real optical zoom or long lens but I suspect the effect will be similar.

Other Sensor Sizes

Working upwards from the phone sensor, many digital compact camera sensors will also be very small so we'd need to step into the mid-range to get a worthwhile increase. There are numerous sensor size charts on the net - for example this one. The phone was a 1/3" sensor. Canon Powershot G cameras (other than the G1X), Nikon 1, etc go slightly larger, then you get into the micro four-thirds popular in Olympus and Panasonic cameras, then to the APS-C found in most DSLRs and many compact system cameras, up to "full-frame" 35mm equivalent, and on above that into medium and large format film and digital backs. The same lens used on any sensor size with the same subject distance, cropped to give the same angle of view will always give the same depth of field, but if you physically move the camera to compensate then depth of field will be affected. This rather excellent depth of field calculator lets you play with the various parameters.

Take our standard 50mm lens on a full frame 35mm SLR. At f/2 (pretty wide open) and the lens focused at 40 metres we have apparent sharpness from 20metres to infinity. If we take that lens and put it on a 6x7cm camera such as the Mamiya RB67 which has around four times the sensor size, we end up with a wide-angle lens. In order to take the same photo, the photographer needs to move much closer to the subject (one quarter of the distance), which will reduce the depth of field available. My maths might be wrong here (it's late at night) but assuming the photographer moves to 10m from the subject to make it the same size in the frame, and keeps all other variables the same, the depth of field reduces to only 15 metres - between 6.5 and 22 metres, so anything more than 12 metres behind the subject will appear blurred. This means that it's much much easier to achieve shallow depth of field as the sensor gets larger. This could also be taken as being harder to maintain a large depth of field, which is why lenses for large format cameras often go to smaller apertures than those on smaller formats, with f/64 being a commonly available small aperture on 5x4 inch cameras.

Back to the point

So we have our advantage for larger sensor cameras - depth of field is more controllable for a given length of lens - as well as a somewhat circuitous explanation of why it matters where you focus. There's one more point to consider that aspiring serious photographers should be aware of and that's the concept of hyperfocal distances. When you focus on a given point and choose your aperture to set the depth of field, one third of that "acceptably sharp" section is in front of the focused point (towards the camera) and two thirds are behind it. This allows you to maximise your depth of field by choosing where in the frame to focus. If you focus on infinity when shooting a landscape, you throw away two thirds of your depth of field beyond infinity (Buzz Lightyear kept labouring this point). If you can't see infinity and you're shooting at an actual plane (the front of a building) then focusing on the building itself puts two thirds of the depth of field inside or behind the building where you can't use it. If instead you focus in front of infinity or in front of the building, you should be able to keep more of the foreground in focus while not losing the sharpness of the infinity / building point. Some lenses (mostly older ones) have depth of field marks on the focus scale which allow you to pre-calculate this and pre-focus the lens. Without this you're basically reliant on guesswork/experimentation, on clever Canon-specific modes (DEP and A-DEP both use hyperfocal focussing) or on using the Depth of Field Preview function on your camera.

Depth of Field Preview

Found on most SLRs, this is usually a button but sometimes a lever, usually on or close to the lens mount. When you press it the lens is stopped down to the taking aperture and the image will (assuming you're not set wide open) darken. In this darker image it's possible to see the extent of the image that will be acceptably sharp. With enough dexterity it's possible to adjust the aperture and the focus distance while holding this button down, and watch the changes as they happen. Try it next time you have a camera and a few spare minutes. Manually focus on something distant, set f/8 as the aperture, and hold down the depth of field preview button. Now adjust the focus towards the camera. The item in the distance shouldn't start to go blurry until the focus point is quite a long way back towards the camera, but things in the foreground should start to get sharper as soon as you move the focus. It's a win-win for shots where you have the time.

DEP and A-DEP

Your remaining depth-of-field control options if you have a Canon DSLR are these two modes. DEP allows you to individually focus on two scene elements. The camera will work out the optimal aperture and focus distance to get both elements sharp, allow you to recompose, and take the shot. A-DEP is a simpler beast where you have to put both of your subject elements beneath focus points at the same time, but otherwise the logic is the same. In both modes, the camera will choose aperture and shutter speed to give you AT LEAST enough depth of field, if possible. It may, depending on lighting conditions, give you more. I don't think it's designed to allow you to deliberately get a shallow depth of field in bright conditions.