Hand Held Landscape Question

[cmi]

What I'm asking ... when you hand hold your camera, if you rotate the camera ... will that rotation ... blur a very distant background, and if so, how fast would the shutter need to be to get a sharp image, you know, a rule of thumb given average possible rotation hand holding for an average hand.

Hm, not sure why I did not get a response, I'll try it once more.

To sum it up, for (nodal) rotational camera movements, the amount of movement in the FINAL IMAGE (blur lenght!) is not different for far away and near objects.  

(Sidemark: You can prove that with a cam set up for nodal rotation on a tripod. Wide angle, some very near and far objects in focus. Rotate camera to get blur and expose. Blur lenght of distant and near object will be exactly the same lenght. Note that this first seems counterintuitive since a human doesnt move around a nodal point when he turns his head, we just have a different viewing experience, with parallax all around us. Nevertheless the fact stands true and its very important to separate the different kinds of movements to avoid confusion.

But what INDEED introduce differences for near and far objects is movement along x and y axis of the camera. That gives parallax. The near moves faster than far away objects.

So you are partly correct, partly wrong. You have *indeed* speed differences introduced by camera movement, dependend on distance, you are right with this. Your misconception is that you think it is caused by rotation. Rotation has nothing to do with it, rather it comes from movement along the xy-plane (shaking hands), and THIS causes parallax.

However we are speaking of a very minor effect. The general blur caused by shaking hands will be much more visible than a small gradual difference introduced by parallax in *that* blur. So the whole issue this is only theoretical, has no practical consequences.

To sum it up, when you have camera shake, foreground shakes always a bit faster than background. Thus your concern about more blured background is unfounded and you only need to make your normal countermeasures against camera shake. You do not need to take special care for backgrounds because they move faster.


Christian


PS: I have figured this out for myself by thinking about different perspectives and movements, and it corresponds to my own observations. If Im wrong please correct me

[cmi]

Oh and by the way, an even better and faster test: Just close one eye, look at your desk lamp (or whatever) and an object behind it. Now try to rotate your head around the nodal point of your eye. (I know this sounds funny.) You of course wont be exact, but you will know when you are getting close to nodal: Then fore- and background will not move relative to another. You just canceled out parallax. Its really nothing special.

[dwdallam]

OK so the consensus is that even when rotation occurs, even though you are covering more area further away, that fact does not result in more blur for distant objects. If that is true, then whatever shutter speed creates a sharp image in the foreground will also create a sharp image in the background.

Does this mean that you can take a tripod sharp image using a fast enough shutter speed while hand holding?

[dalethorn]

OK so the consensus is that even when rotation occurs, even though you are covering more area further away, that fact does not result in more blur for distant objects. If that is true, then whatever shutter speed creates a sharp image in the foreground will also create a sharp image in the background.
Does this mean that you can take a tripod sharp image using a fast enough shutter speed while hand holding?

As long as it's digital, not analog, then the per-pixel sharpness would be the same for close or further away.  But there's more detail further away, given comparable types of objects at the two distances.  i.e., close up, one rhino may fill the frame, while you might squeeze six rhinos into the frame farther away.  So the per-rhino sharpness would degrade a lot at distance.  One could argue that there's nothing you can do about that - fewer pixels per rhino, however, small rhinos invite closer visual inspection than a single rhino close up.  So depending on how large the blur effects are compared to the pixels and digital noise, that will determine whether the distant objects look sufficiently sharp or not.

[cmi]

Dwallam, exactly like you say. Regarding tripod-sharp results handholded... well from my own experience, if Im in comfortable safe territory shutter-wise, I frequently (not too often) do get razor sharp imagery handholded. But of course I wouldnt rely on handshooting if maximum detail is mandatory.


Christian

OK so the consensus is that even when rotation occurs, even though you are covering more area further away, that fact does not result in more blur for distant objects. If that is true, then whatever shutter speed creates a sharp image in the foreground will also create a sharp image in the background.

Does this mean that you can take a tripod sharp image using a fast enough shutter speed while hand holding?

[guyharrison]

Hi,

I'm not a scientist, just a dedicated landscape shooter for 25 years (have even sold a few).  For landscapes and from your question, I assume you share the landscaper's general desire for extreme sharpness, tonal range and low noise which means ISO 200 or lower.

Bottom line, don't handhold.  This does not mean you are restricted to a tripod.    Your camera back or backpack laid on the ground, tree branches, rocks, car doors and roofs, guardrails, fence posts, anything that you can add or improvise for support will be better than handholding.  If there is not a horizontal support, hold the camera correctly and lean it against a vertical surface (building wall, tree, pole, etc).  Even walking sticks can make a monopod.  When there is nothing, I sit or kneel with my feet planted firmly and use my knee as a support.  MUCH better than handholding.  I think Kirk still makes a strap with a foot loop on one end and a release clamp on the other (if you use Arca quick release style plates).  Step on the loop and pull up on the camera to put pressure on  the strap.  It fits in your pocket and, yes, still much better than handholding.

If you want "tripod sharp" while handholding, the old rule of 1/focal length just will not do it even with excellent technique (which I have). That rule gets you acceptable sharpness but not large-print sharpness. You want 1/twice (or even three times) the focal length at least, especially if you are tired or your heart is beating hard (backpacking, walking, climbing) or if it is windy, or if you camera/lens is heavy.  This, of course, restricts your aperture to fairly large ones and limits depth of field which is a different issue from sharpness (large apparent sharp depth of field from front to back is often more dramatic and emotional in landscapes, even with loss of overall lens sharpness due to diffraction, and many technically oriented shooters seeking theoretical perfection often forget that there is no substitute for a small aperture--just look at the discussions on these forums!).

This is one person's experience, but my years have showed me that any support is better than none, and there is, still, no substitute for tripod even in the age of IS, VR and other stabilizing miracles.  These types of lenses, however, make improvising an adequate support much more effective.  With a stabilized lens, good conditions and shorter focal lengths you might get away with 1/focal length if you can improvise a support.

Hope this helps on a practical level!

Guiy

Dwallam, exactly like you say. Regarding tripod-sharp results handholded... well from my own experience, if Im in comfortable safe territory shutter-wise, I frequently (not too often) do get razor sharp imagery handholded. But of course I wouldnt rely on handshooting if maximum detail is mandatory.


Christian

[BradSmith]

OK so the consensus is that even when rotation occurs, even though you are covering more area further away, that fact does not result in more blur for distant objects. If that is true, then whatever shutter speed creates a sharp image in the foreground will also create a sharp image in the background.

Does this mean that you can take a tripod sharp image using a fast enough shutter speed while hand holding?

So what is your actual experience? Isn't this the type of test that we've all done with digital because it is so easy to do and see the results (compared with how we had to do it with film and enlarging)?   If you handhold with a 1.4 or 1.8 lens on a brightly sunlit scene at 1/1000 - 1/2000  - 1/4000 and if your camera will do it, at 1/8000, do you get motion blurring at either 10 ft or at the horizon?   If you want to, put the camera on a tripod and repeat the photos exactly.   Compare.   I don't think you'll detect motion blurring in any of them.  Doesn't that answer your basic question?

Brad

[Dick Roadnight]

If you handhold with a 1.4 or 1.8 lens on a brightly sunlit scene at 1/1000 - 1/2000  - 1/4000 and if your camera will do it, at 1/8000, do you get motion blurring at either 10 ft or at the horizon?
Brad

That might answer the camera shake question, but in the real world of Landscape photography, 1/8000th does not help if you use ISO 50, f11 (for depth-of-field) and a polarizing filter: if you do not use a tripod there is always a compromise - camera shake, ISO noise, depth-of field. There has been the odd occasion when I have tried to take a marketable landscape without a tripod (or a Sinar).

[Pete Ferling]

Yes, movement should multiply over distance.  However, that's only one of too many variables that can go wrong, as the world itself makes for a bad lab to experiment, (you can turn off or isolate one variable).  Wind, camera shake, mirror slap, it all adds up for finite details.  Three things working for or against you are print size, sensor limits, and lens diffraction.  Because nothing else will overcome those at a given moment when you take the shot.  Even a gimp on an LCD monitor will not reveal issues that you'll discover only when loaded up on the PC at home.  There have been cases where I used a sharp lens, F8, 1/30 on tripod, MLU and cable release, and the wind was blowing and moving the trees in one part of a scene.  

Some scenes don't require fine detail to tell the story and make a huge print.  Others have me dragging the extra gear out and praying.  Would it be safe say that hand held landscape shots are cause for trouble?  You bet.  Does that stop me from not using hand held?  No.  I have lots of stuff that would look great on a post card or screen saver.  However, I have also returned to the same scene, dragged out my Mam 645 film on tripod, and used MLU and cable release (or my 40D and pan and stitch).  Waited sometimes for hours and took a shot that looks amazing at 40" on an epson 9800.  As if your standing there.  That was certainly work.  Then again, you have to remember that before we had DSLRs, serious landscape cameras were essentially heavy large boxes requiring tripods and cable releases anyway.

I will close with this, and that I think folks forget that just because a large sensor may have replaced the need for large boxes and film, it doesn't mean you can leave the tripod at home and hope that luck will be on your side.

[dwdallam]

Had a talk with a friend who has her BS in physics and she said if it were a laser, then yes, the further the laser point is away from it's origin, the faster it travels. No, this doesn't apply to cameras because cameras are talking a picture in a 2 dimensional universe. I don't under stand this, but she assured me that the background would not be out of focus due to an increase of speed. I've been thinking about it and one thing I can come up with is that a laser puts out light, while a camera takes it in. I can't get it further than that.

The answer to my question, however, is that no, the background will not be out of focus anymore than any part of the image as far as speed and distance are concerned.

[Jonathan Wienke]

There are a lot of variables involved in your question, but the three most important are the focal length of your lens, the pixel pitch of the sensor, and how steady you can hold the camera. Focal length determines how long the motion blur trail is on the film/sensor. If a given degree of camera shake results in a 1-pixel blur at 25mm, with a 50mm lens the blur will be 2 pixels long, a 100mm lens will give a 4-pixel blur, etc. As long as the angular range of camera shake is held constant, blur length is directly proportional to focal length.

The rule of thumb is 1/f for minimum hand-holdable shutter speed, but this is only a rough estimate. A camera with high resolution such as a 1Ds-III will be harder to hand-hold successfully than a 3MP point-and-shoot, all else equal. And the ability to hold a camera steady varies considerably from one individual to another. I can flout the 1/f rule by a factor of 2 with my 1Ds (original/classic) and still get a pretty high percentage of keepers most of the time, but that isn't always true. If it is windy or cold enough to induce shivering, the minimum hand-holdable shutter speed can increase by a factor of 4 or more.

If you live in a locality that allows private firearm ownership, target practice is a good way to hone your (photographic) shooting skills. The technique needed to group your shots closely on the target with a pistol is practically identical to the technique needed to get a sharp hand-held capture with a slow shutter speed: a firm, steady grip, and a smooth, steady release of the trigger while keeping the gun/camera on target. .22LR ammo is cheap, and you can get a decently accurate .22 pistol for less than most L lenses.

The bottom line is that there's no way to define an exact minimum shutter speed for "acceptable" sharpness; there are too many variables. But you can learn to control some of the most important ones.

[dwdallam]

The bottom line is that there's no way to define an exact minimum shutter speed for "acceptable" sharpness; there are too many variables. But you can learn to control some of the most important ones.

Isn't there a point where the shutter speed will overcome those problems, say 1/500th, or at least make them almost a non issue?

[Jonathan Wienke]

Yes, a fast enough shutter speed can always eliminate motion blur due to camera shake. But defining the exact point at which shutter speed is fast enough to prevent camera shake from degrading sharpness is like nailing jelly to a tree, for the reasons I mentioned earlier. And sometimes the minimum "fast enough" shutter speed is too fast to get a decent amount of exposure, and you have to trade off motion blur and noise to find the least undesirable combination. Sometimes there is no combination that will yield satisfactory results, and you simply must use a tripod.