Taking pictures handheld; some neuromechanical aspects

[Lindolfi]

Advertisement (gone after registration)

Taking pictures handheld; some neuromechanical aspects

Regularly discussions emerge on the limits of handheld photography. So let us look into the neuromechanical aspects. Neuromechanics is the field in which the interplay between the mechanical aspects of (human) motion and the neural control is studied. This includes the use of tools, like a camera for instance.

 

There are several aspects to consider

 

-----------------------------------------------------------

 

1] postural balance control

2] eye-hand control

3] inertial properties of the human body and the camera

4] shutter time

5] whether there is aiming at a static or moving subject (tracking)

6] the interplay between some electronic stabilization and human control

 

-----------------------------------------------------------

 

1] postural balance control of humans has been studied extensively. The most important result is that the sway of the body has a low frequency band content that initially looks like noise, but there is some selfsimilarity in it. It means, that the resulting movements of the head for instance while standing are irregular, but due to the smoothing out by masses and muscle properties as well as the properties of the neural control process, sometimes events of control seem to repeat themselves. The movements of the head are largest when standing (due to the long chain of elements from the ground to the head), smaller when sitting ( the chain starts at the hip or at the shoulder with back support) and smallest when lying down with the elbows on the ground.

 

2] eye-hand control is much more complex than postural control. The body configuration, specifically of the arms, is of importance. The frequency content of the motion (angular motion) of the hands covers the whole band from 0 to 10 Hz. Slow and fast components are added, making the exact direction in which a camera points impossible to predict from the last second in time. It is important to see that there is absolutely no neural relation between postural control and eye-hand control: the eye-hand control simply takes the postural sway as a given, rather than being able to include it by knowing what comes next. When aiming, slow components of the movement are due to visual feedback (about 3 Hz and slower), while the fast movements are due to mechanical feedback (through receptors in the body, up to 5 Hz), and tremor (up to 10 Hz).

The idea that you may be able to purposely release the shutter at a moment in time in which there is no movement, is an illusion for frequencies above 1 Hz. The reason for that is that although the slow movements have the largest amplitude, you still need time to extrapolate the movement to about 0.2 seconds into the future. And that excludes shutter lag! The needed extrapolation has to do with delays in muscle activation and in the visual and motor systems of the brain.

Trying to actively chase your own automatic movement control also takes away some of the attention needed for taking the picture at the right time. Moreover, often the timewindow of moving subjects is determined by them, like making a gesture, a facial expression etc., so there is no time to wait for a a calm moment in your aiming error.

All in all this implies that we as photographers are left with the quality of our aiming action and we can only choose favourable conditions (such as postures, mental states and camera settings) in which we release the shutter in order to get a sharp image, if that is what we "aim" for.

 

3] inertial properties of the human body and camera are of importance, since they shift the frequency of aiming errors down. So a child has a higher frequency content in the aiming movements than a young adult. Also a heavy camera filters out some of the high frequencies of tremor. There is a catch however: when the arms become fatigued, muscular activation becomes more irregular, making it more difficult to stabilize a heavy camera than a light one.

 

4] shutter time is of importance, since it should match the aiming movements with the required resolution of detail in the image. Obviously, when the aiming movement has some angular velocity omega and the required resolution has some angle of view alpha, the maximum shutter time is T=alpha/omega. Since omega is not constant, you may be lucky sometimes, since you may happen to have released the shutter at a low vlaue of omega, but it hardly ever becomes zero. However, if you want to have a hitrate of 50%, you will have to use your average omega as a guideline.

Alpha (in radians) can be established from focal length F and pixel pitch P:

 

alpha = P/F

 

So that means that

 

T = P/(F.omega)

 

If you use the old rule of thumb of T=1/F, you had better have an omega of P radians/sec in which P is expressed in mm.

So how do we obtain our personal omega for a particular posture and camera?

Well, the great thing about a camera is that it is a recording device for movement, so we can use it to measure omega.

 

Here's the recipe

1. Cut out of a white paper or plastic label a disk with a diameter of 1 or 2 mm. Put the self adhesive disk on a black surface, like a black background screen or carton board.

2. Place the target at a distance of a few meter, say 4 meter.

3. Use the focal length F that you use often in critical circumstances.

4. Select a shutter time of ten times your 1/F, (for instance for a 50mm lens, use t =1/4 sec.)

5. Take several pictures with exact focus and look at the length of the trail of the center of the image of the target in terms of number of pixels N covered. You can use the average from the pictures you took, to increase reliability.

6. An estimate of omega is now omega=N.P/(F.t)

7. Your critical shutter time is now T =t/N

 

Here are 3 typical (not best) examples of an exposure at 1/4 sec with a 50mm focal length, while I was standing: (the images are at 1200%)

 

N=10

 

T= (1/4)/25 = 1/40 sec.

 

So a very direct approach is to simply count the number of pixels covered by the center of the target and divide the shutter time used by this number N to get the critical time needed T for the focal length tested. If that T is roughly equal to 1/F, you can use the rule of thumb, else you will have to use shorter exposure times if you want to use all of the resolution of the sensor available. In my case, I end up with the 1/F rule, but that is just me. You may be steadier.

 

This is being very strict! In real life, things are not that strict and you can use a Fuzz Factor Z. This factor depends on the quality of the optics and other factors that may be more limiting than the sensor resolution. In my real world a Fuzz Factor of 2 or 3 is often quite acceptable. So multiply you critical exposure time with the Fuzz Factor and use it as your personal rule of thumb for that particular posture.

 

It also shows that a camera support (tripod, beanbag, ground, table etc.) can be really helpful to increase sharpness if the situation makes you run out of exposure time.

 

In stead of P as the pixel pitch of your sensor in mm, you can also use the diameter of the circle of confusion that is used for DOF calculations, it was 0.03 mm for 24x36mm film, but now 0.02 mm may be more appropriate for the Leica M9 for instance. 0.02 mm is three times more tolerant than the pixelpitch of the M9.

 

5] whether you aim at a static or moving target makes a difference. All of the above is dealing with a static subject. Tracking a moving subject makes things worse: you are adding the error from aiming to the error of tracking. Again this tracking error is personal and this time it is not easy to measure without a software driven target of which the trajectory can be programmed.

 

6] some camera manufacturers have introduced image stabilizing hardware and electronics. They include sensors in the lens or camera body and based on their signals, hardware is moved (rotated) to counteract aiming movements in a certain frequency band. Some of this band overlaps with the frequency band in which the human system is using feedback. This in itself can produce problems, since now two control systems, which are not connected to each other, try to solve the same problem in different ways. It is like trying to drive a car with two people using one wheel. However, the manufacturers have chosen the frequency band and gains of the system quite carefully, so that is still very usefull, as long as you relax and not clasp the camera too tightly.

 

Hopefully that this has been useful. Most likely you already had some feel for what is possible with your camera, body and brains. But if you keep running into surprises and miss shots due to aiming error, you may want to rethink your workflow and perhaps even run tests as outlined in item 4] above.

 

Being able to use long exposure times handheld is not a "man thing", but just neuromechanics filled in with some reality checks. This is a personal thing and not a one-size-fits-all limit.

 

Perhaps I've used jargon in some places. My apologies. I'm always willing to clarify if needed.

Edited March 11, 2012 by Lindolfi

[Guest Ornello]

Very interesting. I have had some success hand-holding at lower speeds than are commonly considered 'doable'. The equipment does make a difference too.

Edited March 11, 2012 by andybarton

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[Lindolfi]

Very interesting. I have had some success hand-holding at lower speeds than are commonly considered 'doable'. The equipment does make a difference too.

 

You are not alone and perhaps my little essay explains why. If you read it, you will see that I do not make any assumptions about what is 'doable'. There are personal physiological and neural properties, which everyone has to find out themselves. It is not a contest with a commonly considered limit or with anyone elses longest shutter times! It is the craftmanship of knowing yourself.

Edited March 11, 2012 by Lindolfi

[jaapv]

Given the individual physiological limits, to what extent is this trainable?

[Lindolfi]

Favourable postures and relaxed concentration ('flow') can improve, but the 'engines' of tremor and postural plus aiming control can not be trained. Those engines can be negatively influenced by things like alcohol, however.

[jaapv]

Doesn't that depend on the amount of alcohol? I am told that in competetion shooting alcohol is regarded as doping, as a limited amount will stabilize your hand.

[Lindolfi]

Advertisement (gone after registration)

Absolutely, that is why some beer combines well with playing darts, within limited quantities. It helps people to get into 'flow' if they can not find it without alcohol.

 

Another remark: it is really very productive to try different postures and mental states. Most people have a good intuition for favourable postures. A favourable mental state can also be discovered by trial and error.

Edited March 12, 2012 by Lindolfi

[jaapv]

Interesting. Makes me wonder though. I am quite good at handholding , gun markmanship and in my job, but hopeless at darting and ball games.

[lars_bergquist]

This is interesting. Two points occur to me:

 

• As a rifleman and pistolero (well, ex-) I was taught breathing control. We all felt this made a difference. Well, the difference could be measured in target points, the difficulty was to sort out the casuses and the effects. What is the opinion on that?

 

• Also as a rifleman, I learnt to use sling support and adapted that technique to photography. I have found it helpful. What it does is essentially fusing the shoulders and the camera into one unit; I presume it does shorten the postural chain; am I right in that?

 

Another lesson was that conscious attention to the tool (gun, camera) and on posture and the trigger finger invariably lead to bad results, usually pulling the shot. All experienced shots do agree with the Zen archers: You must be one with the target. This of course brings in the mental aspect.

 

Shooting and alcohol: Some shooters can benefit from a beer. But this just diminishes the nervousness that plays havoc with their trigger control, it does never lead to good scores, and it is bad for your coordination. The problem has to be attacked at its root.

 

LB

[Lindolfi]

Interesting. Makes me wonder though. I am quite good at handholding , gun markmanship and in my job, but hopeless at darting and ball games.

 

Darting and ball games requires something extra: feedforward. That involves, based on sensory input, to prepare a sequence of muscle activation before the effect is visible (due to the short time of the action and the flight phase in which nothing can be controlled anymore).

[Lindolfi]

Lars, good points.

 

Breathing cycle is of importance. Either consciously or just intuitively applied can be effective.

 

Indeed limiting degrees of freedom, by closing chains of (body) segments is effective and part of finding the right posture.

 

I really can not agree more with your point on attention: any attention to the detail of the action pulls the photographer out of the 'flow' and is counter productive. The target is the only thing of importance for the conscious awareness. The rest is solved by other brain parts that don't like interference from the frontal lobes of the brain. The most sensitive part for interference is the pre-motor cortex that has several functions that are in competition with each other.

 

I agree with your point on alcohol. Also judgement of performance changes both with the performance itself that may give the illusion you get better.

Edited March 12, 2012 by Lindolfi

[pgk]

In gallery target shooting (which requires accuracy within extremely limited timescales and is undertaken from a standing position with no slings) a shooter can/will improve with practice - so there must be a training and exprience factor which kicks in too? As this is so, I assume from what you have said that there will probably be a limit on how much this experience can help improvement as this will be finally constrained by both the physiology and mental capacity (within this context) of the shooter? So practicing with a camera and low shutter speeds should also help increase the 'hit' rate of sharp images at low speeds?

[Lindolfi]

Absolutely Paul, agreed, including the distinction between the processes taking place, some of which can improve, others not. The learning process does help, but there is a limit. That limit varies a bit between people, but when it is reached, a tripod or other camera support can help us.

[kdriceman]

Absolutely Paul, agreed, including the distinction between the processes taking place, some of which can improve, others not. The learning process does help, but there is a limit. That limit varies a bit between people, but when it is reached, a tripod or other camera support can help us.

 

I believe my limit has changed over time and the rate of change is increasing. Sigh.

[250swb]

Assuming you aren't doing theoretical calculations based on taking photographs in the vacuum of space how is a gusty wind factored into the equation, or indeed how standing on various surfaces, like springy grass, or on the side of a hill, factored in? And I would have thought the time factor available to hold your breath and get comfortable and relaxed would be a significant variable. If you need to press the shutter to capture the moment before you are comfortable this affects the expectations of what is possible. How would that be factored in? Or are we simply taking about a perfect world?

 

Steve

[Jeff S]

I might have missed it, but didn't see reference to "pressing" or "squeezing" techniques to depress the shutter release and the related use of tools or aids to do so, e.g., soft release setting (M9), soft release buttons, thumb accessories for holding, etc. I presume these issues fall under several categories, including inertial properties, and perhaps another extension of the stabilization concept (in addition to Lars' sling).

 

Jeff

[Guest Ornello]

Given the individual physiological limits, to what extent is this trainable?

 

The main thing is not to 'jam' the shutter but squeeze it. Most motion is not "camera shake" but instantaneous movement caused by pushing the shutter too violently. The speed of the movement is the issue, not so much the amplitude.

 

Some people hold their cameras too tightly, causing their muscles to quiver at a high frequency. It's the speed of the movement, not the amplitude, as I said.

Edited March 12, 2012 by Ornello

[jaapv]

Hmmm...it has been said before....

 

 

http://www.l-camera-forum.com/leica-forum/leica-m9-forum/130720-m9-faqs-frequently-asked-questions-answers.html#post1548925

[Lindolfi]

250swb, varying conditions are all part of the aiming task. The values you get from a steady even floor indoors are just a starting benchmark. Nobody will assume that you will be as steady in a cross gale. Turbulence of wind is hard to predict or counteract. On a steady but uneven floor you have a better chance, since you can adjust your posture and even impedance of your legs to the surface. It has been measured that people do this without knowing it.

 

You understand that the first steps I wrote down in my short essay will not produce definitive results for the wildly varying conditions we may want to take photographs in. It is up to you and the 1.5 liter of living jelly inside your skull to make the best of it.

 

Jeff S. Indeed the shutter release action is important, but only beginners tend to not only press the button but also the whole camera down. The control of your finger and the sensory interplay with the steel and springiness of the button is remarkable. The test whether you do it right is just attach a small laserpointer to your camera and see wether the wanderings of the dot does anything special at the release of the button. If you do it right, you see nothing happening.

 

Ornello, I disagree: I've done the experiment with the laser pointer and extended it with an experiment with a 3D acceleration transducer and it appears you don't see the pressing of the shutter in the midst of the complex motion that is due to all the processes I wrote about in my first posting. Test was done with three experienced photographers and an M9.

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[Guest Ornello]

250swb, varying conditions are all part of the aiming task. The values you get from a steady even floor indoors are just a starting benchmark. Nobody will assume that you will be as steady in a cross gale. Turbulence of wind is hard to predict or counteract. On a steady but uneven floor you have a better chance, since you can adjust your posture and even impedance of your legs to the surface. It has been measured that people do this without knowing it.

 

You understand that the first steps I wrote down in my short essay will not produce definitive results for the wildly varying conditions we may want to take photographs in. It is up to you and the 1.5 liter of living jelly inside your skull to make the best of it.

 

Jeff S. Indeed the shutter release action is important, but only beginners tend to not only press the button but also the whole camera down. The control of your finger and the sensory interplay with the steel and springiness of the button is remarkable. The test whether you do it right is just attach a small laserpointer to your camera and see wether the wanderings of the dot does anything special at the release of the button. If you do it right, you see nothing happening.

 

Ornello, I disagree: I've done the experiment with the laser pointer and extended it with an experiment with a 3D acceleration transducer and it appears you don't see the pressing of the shutter in the midst of the complex motion that is due to all the processes I wrote about in my first posting. Test was done with three experienced photographers and an M9.

 

I am not following you. I said the main causes of blurry photos are:

 

1) Too much of a jab on the shutter release, a high velocity movement.

2) Gripping the camera too tightly, which causes tremors, which are high in velocity too.

 

It's the (angular) velocity of the movement that matters most. Do your measurements take that into account?

 

"No sudden moves, see...!"

 

http://thesocietypages.org/socimages/files/2008/07/gun1.jpg

Edited March 12, 2012 by Ornello