Does hand-held camera shake scale with print size?

[will_frost]

<p>Hi all,<br>

Given that you should use 1/focal length for hand-held shutter speed, should you double it when you double your print size?<br>

In other words: if I'm using 1/50th of a second on a 50mm lens, and printing at 4x5, should I be using 1/100th of a second at 8x10, and 1/200th at 16x20? Or once you get beyond 1/100th with good grip technique, does the blur from being hand-held get swamped by all the other factors that introduce blur? (Depth of field/focus errors/grain or noise, etc.)<br>

Thanks everybody,<br>

Will</p>

[dave_s]

<p>Yes, double it for every doubling of magnification.</p>

<p>People who are fussy think the 1/focal length rule is optimistic, by the way. I usually try hard to get to 1/250 with a normal lens, but that's me. YMMV.</p>

[will_frost]

<p>Thanks Dave,<br>

Ah, then if I'm doubling it for print magnification, then if I double the number of megapixels, that would count too, right? Or if I double the area of the sensor?<br>

E.g.: 10 megapixels - > 20 megapixels<br>

or:<br>

4/3rds format - > full frame<br>

Both of those cases are like doubling the magnification, right?<br>

Will</p>

[johnw63]

<p>The higher the shutter the sharper everything will be. It really doesn't matter how BIG you print it, because smaller photos get looked at closer up and big shots make people stand back further. I think I read somewhere that the distance viewed is related to the size so all things remain about the same. This means it will still look sharp no matter how big it is, as long as it was sharp with you pressed the shutter.</p>

<p>That rule of thumb is a MINIMUM goal, not the best shutter speed. If you can get away with a faster shutter and get the result you want, do it. If that changes the aperture too much and messes up your desired DoF, then your stuck with the minimum.</p>

<p> </p>

[starvy]

<p>Will, if you are really worried, buy yourself a very cheap and light tripod. That is a better option than hand-held although ultimately, getting a camera with some sort of stabilisation would also go a long way.</p>

[will_frost]

<p>John,<br>

Right, 8x10 is a kind of worst case scenario, because the "natural" viewing distance where the image fills your field of view is the same as the "close" viewing distance of about 15-18". However, I've noticed that no matter how big you print, after people have gotten a good look at a normal distance, they always walk up to the distance that resolves the finest detail.<br>

The main thing I'm interested in is getting a handle on how blur effects work - how they add together, and how they can limit what I can get done.</p>

[david_littleboy__tokyo__ja]

<p>My experience is that images are either sharp or grossly fuzzed; there may be a continuum in there somewhere, but in actual real life, the higher the shutter speed, the higher the percentage of sharp images.</p>

[SCL]

<p>Erwin Puts, the guy who for years has evaluated Leica lenses, stated in one of his publications a couple of years ago, that if you are shooting (any focal length) at less than 1/500, sharpness can be improved with a tripod. As I get older, I'm starting to believe him.</p>

[jim_j2]

<p>Ah, . . . and thank goodness for IS!</p>

[mike_hitchen]

<blockquote>

<p>Ah, then if I'm doubling it for print magnification, then if I double the number of megapixels, that would count too, right?</p>

</blockquote>

<p>Assuming 2 sensors of the same size but different pixel density, then no. If you shake by 0.05mm, the movement on the sensor is 0.05mm whether that covers 10 sensors or 20 sensors (I'm inventing numbers here, but the sense is the same). And in turn that will cover the same distance on the printed picture.</p>

 

<blockquote>

<p>Or if I double the area of the sensor?</p>

</blockquote>

<p>Yes: if you take the same picture on a Canon 5D and a Panasonic G2, then when you print both to the same size (for instance, letter) then you are magnifying the roughly G2 twice as much because of the differetn sensor size. Which comes back to your original post. I am ignoring here the differing linear ratios but again the logic holds.</p>

 

[will_frost]

<p>Mike,<br>

Hmm. I'll have to spend some time thinking that over. So, it's the size of the movement relative to the size of the format that matters? A 0.05mm movement recorded on a 24x36mm sensor would look half as small (half as blurry) as a 0.05mm movement on a 12x18 sensor?<br>

Does crop factor mess this up? To get the same angle of view and same size subject, you'd be using two different focal lengths, right? A normal-ish angle of view on 4/3 would be a 25mm, and 50mm on full frame, so the same amount of shake should jiggle the image by the same "relative" number of degrees? (Say you are taking a picture of a point of light, and it smears in one direction by 1 degree on the small sensor. Does it smear two degrees when you double the focal length, even if the sensor is twice as big?)<br>

Sorry, I'm not trying to make this needlessly complicated, it's just turning out that way!<br>

Will</p>

[eric merrill]

<p>Will:</p>

<p>To answer your question, it depends. Obviously, if you magnify the captured image, any faults will likewise be magnified.</p>

<p>That said, most people don't observe large prints at the same distance from which they observe small ones. They back away.</p>

<p>That means from "normal" viewing distances, you don't need to increase your shutter speed to compensate for normal unsteadiness.</p>

<p>Eric</p>

[dan_south]

Kudos to Jim Johnson. Think about buying and using IS and VR lenses. They often render a sharper handheld image.

[will_frost]

<p>Thanks for your replies, guys.</p>

[frank uhlig]

<p>And yet, disregarding all the common wisdom around here, I did recently shoot a portrait at 1/15 second, hand held, and one can see every hair and every pore crystal clear, shot from ~ 1 m, when blown up to 12 inches square and viewed from 8 inches, if one so desires.</p>

<p>Of course this was with a Rollei TLR. So the "silly rules" of 1/focal length apply only to those cameras for which they apply. And not to those without mirror shake, definitely not!</p>

[James G. Dainis]

"<I>So, it's the size of the movement relative to the size of the format that matters?</I>"<P>

 

No. Get an 8x10 view camera with a normal 300mm lens on it. Solidly mount a 35mm camera with a normal 50mm lens on top of it. The scene on the ground glass of the view camera and in the viewfinder of the 35mm camera will be the same (a little wider in the viewfinder but we won't count that.) As you fire the shutter for both cameras, tap the tripod so there is a one degree or so shake. The scene on both the ground glass and viewfinder will both have a one degree shake relative to the scene. Make an 8x10 contact print from the view camera negative and enlarge the 35mm film frame to 8x10 inches, If a window in the scene of the contact prints seems to have a 1/2 inch blur, the same window in the 8x10 enlarged print will also have that 1/2 inch blur.

[mike_hitchen]

<p>James - if you mean 'No' in the definitive sense then I am not sure I agree. My thinking is that with 8x10 camera, the lens to film distance is greatly larger than the distance lens to sensor in SLR. So the 1 degree of movement will be translated to a greater linear distance on the 8x10 plate than it is on the sensor. So when you blow them up, you may well find that in a 8x10 print from both formats that the blur is the same (I would be interested to see an example of this).<br>

But if you put one degree of movement on a 4/3 format and on a 5D, then the fact that both bodies have approximately the same lens-to-sensor distance, one degree of movement will give the same linear movement (blur) on both sensors: and that linear movement is a lesser proportion of the linear dimensions of the 5D sensor than it is on the 4/3 sensor. Which means that the blur is less on the 5D when you print them both to 8x10.</p>

[James G. Dainis]

Mike, What I said holds true for the format if one uses a lens designed for that format. If one mounts a digital camera having a 1.6x cropping factor, on the view camera, with a 50mm lens, that is the same angle of view as having an 80mm lens mounted and blur will be increased by that factor.

 

A point and shoot camera with a 1/2.5 sensor (7.75 x 4.3 mm) would have a "normal" lens of about 8.5mm focal length ( = 50mm lens on 35mm camera = 150mm lens on 4x5 Speed Graphic = 300 mm lens on 8x10 view camera). Mount them all together and the scene or angle of view would be the same on all of them. Make a one degree lateral shift and the scene would change identically on all of them.

 

Take that point and shoot camera and zoom out to 35mm actual lens focal length ( = to about 200mm on a 35mm film camera) and camera motion blur is going to be much greater than with a 35mm lens on a 35mm film or full frame digital camera.

 

Photographers from the 1800s who routinely worked with 8x10 or larger cameras would be a bit confused to see a 300mm lens referred to as a telephoto lens. One would have to explain that it is the 6x cropping factor that comes into play when that normal lens focal length is used on a 35mm film camera.

[will_frost]

<p>Once again, thank you all for your replies. Clearly I'm going to have to spend some time thinking this over. (And for digital work, I do have an image stabilized camera.)</p>