Recently, a hilarious episode has inflamed a friend’s family whatsapp group.
The story has its origin in a family wedding where the boy’s sister (x) – the apple of all eyes, and the bride’s cousin brother (y) – a borderline good-for-nothing individual as per his own family’s admission, seemed to have hit it off rather well during the celebrations.
While forced politeness deemed their banter as sibling-natured, especially since the lady was older to the gent by 6 years, their obvious chemistry threatened this thin garnish at every step. The only hope held on to by the families was the short natured festivities and the busy life of Bombay thereafter, when everyone is expected to drift their own ways.
And everyone seemingly did.
Except for when technology laid bare the budding romance in a speedy turn of tides. X and Y were on their way back from a date, X driving the scooty, Y riding pillion, helmet free all, over-speeding with the euphoria of fresh romance – perhaps sweetened even more by thrill of its clandestine nature – singing aloud on an uninterrupted enthralled scooty ride.
The literal speedbreaker to this enchanted state presented itself in the form of two pot bellied traffic police-wahlas. With an integrity which could fie on the powers of generational wealth, not to mention at the meagre bribe expressed as a token apology by the two lovebirds, these two protectors of law flourished a speeding ticket, and what was much worse, took a picture which went straight to the Mumbai traffic app.
Things would still have been prevented from going south, if the scooty wasn’t registered under the girl’s father’s name. What papa X received was a very clear photograph featuring a couple beyond doubt – smiling by the force of habit or hormones – two people who had no business as per him to even be in contact.
Among the ensuing family drama was our friend, now very content with the elevated feeling of being a wise-ass, professed with a few well timed deliveries of “I told you so” on the group chats.
Everyone had their doubts about X and Y’s mutual inclination towards each other, but it fell upon our Mumbai traffic police to provide the grainy, but well focused picture and soothe suspicions .
Was the depth of field deep because surveillance cameras are small with tiny sensors, satisfying the often parroted relation between sensor size and depth of field? Or is there something else hidden in plain sight at play in the matter..
The Circle of Confusion
Let’s begin with a topic so self aware that it admits to confusion in its very name. The circle of Confusion is also known as bokeh, blur circle, disk of confusion, blur spot, blur of indistinctness. But definitely as a concept much simpler to understand that its rather imposing names.
Like travellers whose journeys might join for a moment, but who have an obligation to keep moving on their own paths, the rays of Light too are relentlessly in motion. Once converged, they urgently start the process of diverging.
If we catch these constant travellers upon their meeting point, we have an in-focus point image. If we catch them a little earlier, or later, they are a bit more than a blur.
The lens is an unprejudiced lovely being. It has a particular focal length, which is defined by how long it takes for it to converge light rays coming from infinity, or collimated light rays – to a point.
The plane of focus can only be singular, which means only those point sources of light (emanating from objects) which lie on that plane will be perfectly focused onto the sensor plane. The light rays from objects placed before or after the focal plane, too, will be converged perfectly by the lens to form a point, but the point will not land on the sensor plane – but rather before or after it as the object is placed.
Circles of confusion are the result of those incomplete convergences, or the divergences post convergence, all confluences of a 3 dimensional world mapped onto the 2 dimensional sensor plane. They are the points of travelling light waves which were caught too early or too late from their point of meeting, and hence appear enlarged – circles instead of points.
And they are the building blocks of Depth of Field.
What’s a field? The field of view, everything visible from the lens and being recorded on the sensor is the field of our image. And how deep we can see into this field defines its depth. To borrow a good definition: the distance between the nearest and furthest element of a scene which is in acceptable focus is called Depth of Field (DoF).
Acceptable Focus sounds like a very subjective term, and as we’ll shortly see, so it is.
Depth of field depends on a few factors, all clearly understandable with the help of our friends: the circles of confusion.
Distance:
You’re probably reading this blog on your mobile phone. Do continue, but without shifting eye focus, try and take note of the entire visual surrounding. As long as your focus remains on the screen, the rest of the scene seems pretty soft focused. Now look at a distant object in the same space, say something 10-15 feet away, or more. Aren’t the surrounding areas, the foreground and background of your object of focus also seemingly sharp enough to be observed well?
So what’s happening here? With this experiment, we can all intuitively say that with distance, the depth of field increases. Why is that?
Light rays coming from far approach collimation. It means they come together slowly, but it also means that they diverge slowly. The circles of confusion remain small enough to appear more like point objects on the sensor since the light rays haven’t sufficiently diverged to form the larger circles (of confusion) yet.
Light rays coming from closer have steeper angular relations. Meaning they converge and diverge within a short distance. This makes bigger circles of confusions, meaning instead of point sources, they are registered as circular blurs on the sensor.
Focal Length:
Wider lenses are thicker and have more power. Meaning, they have more power to refract and bend incoming light rays and bring them to a focus in a short span of distance. A 12mm lens is duty bound to gather all the light rays, spread out like unruly children in a school ground during recess, with a stern hand into a single file.
This means the rays have to be rather close to each other once they pass through the lens, their proximity means the circles of confusion will be rather small – and more likely to fall into acceptable sharpness.
Telephoto lenses are flatter and less powerful. Light rays passing through them take their own sweet distance to gently converge onto the sensor. It’s like the same kids from the playground now walking neatly in a corridor, and a dear old english teacher gently nudging them into a single column. Light rays in a telephoto lens are travelling from far, so are more collimated. But they are not as pressed together by the telephoto lens by refraction. Hence their convergences are far more spaced out and the CoC are larger, meaning shallower depth of field.
Aperture:
Squinting is our way of closing down the aperture of the eye and increasing the depth of field to get clearer information from the scene we are looking at.
Aperture is the size of the opening of the lens, adjustable through the iris, which decides the amount of light entering the camera body. A smaller aperture means a lot of the light bouncing off the object is not able to enter through the lens. Only a small section of already rather closely bound rays of light will be able to enter the camera.
Imagine a jam packed Bombay local station, the platform full of a never ending influx of people. The train arrives. And it has a comparatively small door, certainly not enough to accommodate everyone getting on the train at once. Only the people near/around the gate, that is people who are in a way converging together at the right place, will be able to enter the train.
A small aperture similarly streamlines the light rays entering the lens. Only those rays which are already somewhat together, are allowed in. The result being that their convergence and divergence happens in a narrow space leading to small circles of confusions, and a wider depth of field.
Viewing distance and device:
Everything in life is a means to an end, and so is the process of image making. Every decision we make in the process of cinematography should be done with the mind’s eye on the end result. So also the decisions which affect depth of field.
How many times, when the focus is in doubt, the AC/DoP will sit on the big monitor to take a final call, leaving the shoguns aside. It’s because on a small screen, the depth of field is automatically deeper. On a small viewing device, the photographed image appears smaller and so do the circles of confusion.
With greater distance, a circle appears as a point. On a bigger screen, the image points are spread apart, and the circle of confusions are revealed without a doubt.
Hence, the concept of acceptable focus when we describe Depth of Field. It is a subjective matter, and many times, the decision may be worked backwards from the viewing medium.
Returning to the beginning anecdote in this blog, papa X was viewing the image of the euphoric couple on his mobile phone. On such a small device, things ought to appear sharp since the circles of confusion are tiny. However, what about the mobile phone which took the shot, the shooting device? Surely it would have a tiny sensor compared to our cinema cameras. Is there any change of DoF with regards to sensor size?
To make our understanding of this subject crystal clear, let’s understand that a circle of confusion is the sole property of the lens. Each lens has its characteristic blur circle, independent of the medium recording the image. Which in short means that depth of field is a sole property of the lens. Distance, Aperture, Focal Length – all variables in defining the DoF arise from the lens. The sensor can merely vary the exposure (EI) and in some ways the colour (by debayering) of the image formed by the lens.
So when DoF is a sole function of the lens, how can sensor size ever affect it?
Why has this myth perpetrated for so long?
It’s because of field of view adjustments. A full frame sensor is about 1.5x a super35 sensor and x2 a M43 sensor. A 50mm lens is a 50mm lens for all eternity. It basically means that whenever there are parallel rays of light, this lens will take 50mm of space to converge them to a point. The lens is rightfully oblivious the the recording medium.
However, say when a 50mm Full Frame (FF) lens is applied to a a M43 (/MFT) sensor, the field of view of the M43 camera will be half of the FF camera. This is because the sensor is half the size of the FF and can only record the image as per its surface area. So, in comparison to the FF camera, the FOV will look like that of a (50×2) 100mm lens. To compensate for this, a 25mm lens might be added to the M43 lens instead of the 50mm, to give an equivalent fov.
Now as we have seen above, wider lenses form smaller circles of confusion. Which means, the depth of field in the 25mm M43 camera lens system will appear deeper than the 50mm FF one. Did the sensor have any part to play here? Not really. We wanted a particular fov and we went for a wider lens as per the crop factor calculations is all that has happened.
In short, one can make life simple by remembering that the Circle of Confusion, and hence, the Depth of Field comes from the lens. The sensor is but a noble non-reactive recording medium for the light rays falling on it.
*In a later post we will look into a possible relation between pixel size and circle of confusion, hence including sensor to the DoF mix. But for today’s advanced cameras and lenses, the concepts of this blog hold true.
Leave a Reply