Shutter speed.
 

Over on the Bird forum I put some advice on a thread about minimising camera shake saying to use 1/focal length as a general guide, not forgetting to put in the crop factor so for a Canon with a 400mm lens that would mean 1/640 sec. Postcardcv came back and said you did not have to allow for the crop factor so 1/400 would be sufficient even though the lens is effectually a 640. Who is right? Has anyone got any authoritative references? :confused:

Mike,

Can't offer any authoritative references, but I always tell people to use the fastest they can get away with - it doesn't (or shouldn't!) take long to work out what you can successfully hand hold at. Sometimes it is not possible to get the shutter speed you want, so you just do your best.

For example take my current picture of the Jay in the gallery - I pushed the ISO up to 1600 but still only got 1/125 at 550mm, so I made sure I was rock steady and got away with it... guides should only be guides etc.

The short answer is that it depends.
Let's forget digital for the moment and just think about film as pixel densities just confuse the issue. To start with consider Peter's case. A given lens at a set subject distance will produce a given image height on the film regardless if the film was full frame 35mm (24mm height) or APC (15mm height). Assume the image height is 5mm and we want the image on the print to be 5cm the amount of enlargement will be x10. In this case the crop factor plays no part in the equation.

On the otherhand if the image was to fill the height of a full frame 35mm (24mm height) camera it would over fill the height of the APS-C (15mm height) frame because of the reduced field of view. Therefore using the same lens for a APS-C camera the same subject has to move further away to fit into the smaller field of view. This where the effective field of view of a 400mm lens becomes 640mm on a 1.6 crop factor camera. At this point Peter's statement is still true.

The problem starts when we come to printing the images.
The amount of enlargement required to produce an A4 print from the full frame 24mm high image is x8.75 compared to x14 for the APS-C 15mm high image. This is 60% extra enlargement which will show up camera shake problems more.

This extra enlargement also has an effect of Depth of Field Calculations and Bokeh. A reason why some prefer a full frame sensor to a APS-C sensor.

I use shutter speed this way, If my focal length is say 45mm then I use a shutter speed of 1/100s or as near as I can get it. If the aperture does not give you enough light then alter your ISO until it looks right. You need twice the shutter speed for the focal length you are using. A word of caution. Even though you can sometimes get away with it, if you use a slow shutter speed then use a tripod.

I am not sure that the focal length rule works anyway on long lenses. I can handhold my 400/5.6 at, say, 1/500 and get sharp shots most times but if I add a 1.4tc bringing it up to 560mm I struggle to get sharp shots at 1/1000 or even faster (when using a tc I always use a tripod which gives me good results).

I have a theory that the focal length=shutter speed rule works on short focal lengths (say 100mm or less) but as you go up the focal length it becomes less effective. Of course IS comes into it as well.

I do not know if you should add the crop factor or not but to be on the safe side it is worth while to add it.

I just posted my thinking up on BF but thought I'd put it here too -

If I take a shot with a 1DsmkII it's full frame at 16.7mp, if I then crop the image to 1.3x then I get left with ~9.8mp (mich is roughly what a 1DmkIII gives) but if cropped to 1.6x I'd be left with ~6.5mp (basically what the old 300D gave). If I took the same shot with the same lens with each of these three bodies then I would need the same shutter speed to eliminate camera shake. The central portion of the three images should be the same, the larger sensors purely record more data around the edges. So by my thinkng crop factor cannot have any effect on the shutter speed you need to eliminate camera shake. (I'm not sure that I made my thinkng any clearer with that).

The focal length=shutter speed rule can only ever be a guide, personally I can get away with slower shutter speeds. When I had then 400 f5.6 I was happy to handhold down to 1/200th and at times slower.

I'm in agreement with Rob, it all comes down to the printing. If you don't take into account the crop factor increase, you are doing a bigger enlargement from sensor to print, at which point the increased amount of shake will become apparent.

Another way of looking at the problem.

In this example let us assume the camera shake causes a 1mm vertical movement of image on the sensor which will be true for both full frame and APS-C sensor. In the case of full frame this movement equates to 4.1% of the frame and 6.6% for APS-C. It is only when you use different scaling to make the same size print from each system the shake becomes more apparent.

If you use the same size area of the sensor in each case and apply the same amount of scaling the shake will appear the same.

Remember 1/focal length Sec is only a general rule of thumb (guide) which works for most people.

I am of the opinion that as you get older you are less able to hold the camera still at low shutter speeds. I find the IS on my 300 F4 a boom for hand held shots but not as good as tripod.

Thanks for your clear expositions Rob, I think I now understand it. Although it is only a rule of thumb which changes from person to person and indeed, I agree, changes over one's lifetime it is a useful guide when out and about and looking at the camera settings and deciding what alterations are needed.

I get what you're saying but either way there's still camera shake and it will be visable on the image - if you start looking at it in this way then surely pixel pitch and thus mp count become relevant too. Surely a shutter speed that is fast enough to stop camera shake on a full frame senser will still eliminate shake when using a crop body. I have used full frame, 1.3x crop and 1.6x crop cameras and can handhold at the same shutter speeds on all three. For me any camera shake is too much, if an image isn't critically sharp then it gets deleted.

Lets all stick on 8mm fisheye lenses and shoot at 1/8sec shutter speed :)

LOL - I bet young Peter can go even slower :D

On a different note I've noticed on high magification lens as sort of "eye - brain - hand" servo loop effect. When I first look through the view finder everything is steady. But the longer I concentrate on a point I find things start to drift and then I end up over compensating for the movement. To stop this I have to look away from the viewfinder and start again.

Main reason why I lug and use a tripod 99% of the time.

I'm sorry, but I can't see what MP count has to do with image stability, as for me it's simply down to technique - which if you work on it can be improved. I watch, ask, then learn from others and there's plenty of help out on the internet to help you.

I'm willing to bet Nigel (Blake) gets a lot of his shots by good hand holding technique which again I'm sure he obtained by learning and practicing.

I keep coming back to this, both here and the thread on BF, but I just can't get my head round it... in your example blur caused by camera shake is more visable in prints from a crop camera (have I got that right)? That means that there is some camera shake.

What I am saying is that if I can shot at a given speed on a full frame camera and completely eliminate camera shake then surely this shame shutter speed would be fast enough to do the same on a crop camera.
For example, say I can shoot at 1/400th at 400mm on a full frame camera and completely eliminate camera shake then the sensor would not have moved. So the system of scaling the movement becomes irrelevant and the same shutter speed would be needed for a shake free image on a crop camera.

Yet another angle on this problem this time just considering the Pixel point of view.

We have 3 models of camera fitted with the same lens and shutter speed, a full frame, a 1.3 crop factor and a 1.6 crop factor.

A person is the subject in this case and stands at an appropriate distance from each camera so that their image just fills the frame in each case. From each camera we want to make a 10"x15" print at 300dpi so that we should have 3 identical looking prints to compare the camera shake. Again as in our previous examples we have a 1mm vertical movement for the camera shake.

The attached Table shows the figures for each model of camera to produce the print and the number of pixels the movement blur is over.

For the 1.3 crop factor the blur is 1.3 times greater than the full frame model.

For the 1.6 crop factor the blur is 1.6 times greater than the full frame model.

Therefore sensor Pixel density has no effect on the observed blur on the print.




In Peter's Post #6 he is cropping the area of the sensor and using the same scaling factor
so his statement is true in this case.


When you look at the world at the atomic level things are never completely still so in theory you will never get 100% camera shake free image. This is partially true in the field of measuring things, you are unlikely to get 2 measurements the same. So what we do is take hundreds of measurements and perform some analysis on the numbers to get a realistic value.

For much the same reason Camera sensor noise will never be zero (unless you want to operate at -273 degree C or 0K) it is a question of reducing it an acceptable level.

It really is a question of getting acceptable sharpness, i,e sharper than the eye see, when the eye can perceive blur it becomes unacceptable.

OK so you can't 100% eliminate shake, but a movement of 1mm on the sensor is huge. Surely any shot with this much movement would so far too much blur regardless of crop factor. For me accpetable sharpness would be looking sharp when viewed at 100% (then I know I can print as big as I want) and I can acheive this at the same shutter speeds regardless of crop factor.

I used a figure of 1mm just as an example. If the movement is less than a pixel then it will never register.

The limit of the human vision system is said to be 7 lines per mm.
Therefore a fair amount of shake can occur before the eye perceives it.

The bulk of the human vision system is down to the brain and not the eye. So some vision defects are due to the brain rather than the eye.

That is the council of perfection but how do you get a pixel sharp BIF even on a tripod? Also I do not routinely carry a tripod on a walk which may be lax but I am not as young and fit as I was (or dedicated enough) so the shutter speed for hand held shots is what matters to me, the old formula never was intended for tripod shots. Print size obviously does come into it as well as the eye of the beholder so as a rough guide it looks better to allow for the 'crop' factor?

Mike you have to admire those who can get perfect hand held telephoto shots. I've tried and binned more shots than I care to mention. I get the results I want with a tripod and I don't mind lugging it. On my local patch not too many opportunities for Birds in Flight shots.

I sometimes practice on the Black Headed Gulls when folk feed the wildfowl in the park.

My research and calculations basically show that camera shake is more apparent with crop cameras. A 5 pixel shake on full frame becomes an 8 pixel shake on a 1.6 crop camera. Therefore the perceived wisdom to use a higher shutter speed seems good advice.

I saw this today on the topic of this thread

http://www.bobatkins.com/photography...ter_speed.html

Interesting to come back to this thread, thanks for the link Rob - it is certainly a clear explanation.