Is the "1/mm" rule universal?

[george013]

Viewing images at 100% gives you an view of the maximum sharpness (without additional sharpening).

However IMHO it does not always give a realistic assessment of whether an image sharpness is fit for purpose. If it were so, one would discard many usable images. Rather, one needs to assess whether an image is sharp enough for the required purpose: whether it's for print or for display on a monitor, the size of the image, the viewing distance and so on.

It's misleading to say the least to suggest that a 36 Mpixel D800 requires a tripod in circumstances where a 12 Mpixel D700 would not (for example). If the circumstances require a 10x8 print and the D700 produces a sharp hand-held image, then so will the D800 at the same shutter speed.

However, the 36 Mpixel camera will allow a (linear) enlargement 1.7 times larger than a 12 Mpixel image before resolution becomes an issue. If one is going to enlarge an image to the maximum size the 36 Mpixel allows, then the lower camera-shake-limited shutter speed is reduced by the same factor of 1.7.



The section in that guide on motion blur and use of a tripod is IMHO very misleading. It is true to say that you need a tripod for a 36 MPixel D800 when you wouldn't for a lower resolution camera if and only if you intend to enlarge the image more than you would have done with a lower resolution camera.

From my first post.

Keep in mind that arguments above counts for 100% view, 1 sensor pixel = 1 monitor pixel. On a print it's different again, then it print resolution will be involved.

Read above also for the printsize.

There is only one thing known, that the sensorsize and its pixeldensity. There you register the image with its blur and other faults.
The printing, if ever followed, is something else. On small prints you see less. A printresolution of 300dpi is calculated back a dot size of 0.085mm, 10 D700 pixels or 17 D800 pixels.
I remember that Ted, topicstarter, said that he never prints his images. Then he will view them only at the size of the monitor or at 100%.

Another argument used with the megapixelcameras is the possibility to crop and keep enough pixels left. Than the blur on pixelsize is getting more important too.

George

[xpatUSA]

It's not the resolution or the pixels - it's the crop factor needing taken intae account.

Which means that shooting equally shakily with a D700 would be no different than with a D800, is that correct? (same lens, of course).

Does it follow then that the captured images would be equally blurred . . . .

. . . or is necessary to introduce photographic obfuscation in the form of re-sampling, printing, post-processing, etc? Aw, Hell - might as well throw in ISO, diffraction, AA filters and Bayer vs. Foveon while we're at it

Shooting a 50mm@ 1/50th follows the rule but if yer shooting a 2x crop factor 50mm then ye have an angle of view of a 100mm lens so adjust tae 1/100th?

Yes, the "rule" does account for that by defining equivalent focal length, not what's stamped on the body. I forgot to mention that in the OP, but did 'fess up later.

Things ah normally don't say - then wish ah hadn't...crop factor - when ah always (normally) say Focal Length Multiplier. Not only is it a much more elegant phrase, it's much more accurate as well.

Well said! I too hate "crop factor" but have been known to use it, sadly. Must quit that

[rpcrowe]

The 1/focal length suggested shutter speed is simply that, A SUGGESTION. It is not carved in granite and may differ between various people and even between the same person shooting with various cameras/lenses and under differing environmental conditions...

Different people differ because:
The physical ability to hold the camera steady differs globally due to age and physical condition - including intake of caffeine...
Different people will have different ways of holding the camera/lens - some are steady and some not so steady. Using a LCD with the camera 12" or so in front of your face is NOT a steady camera hold for me. I am left eyed which means that the camera is presses against my forehead when using the eye level viewfinder...
How the person presses the shutter button. The old shooter's adage: "SQUEEZE DON'T JERK" the trigger also pertains to the camera shutter button.

Various camera/lens combinations differ because of:
the size and weight of that combination
just how ergonomic is the camera/lens
if you are shooting with a camera that has burst mode - shooting a string of three or more shots will sometimes give a sharper image with the middle shots of that burst because the impact of pressing the shutter button will not impact those shots...

Various ergonomic considerations:
the amount of wind...
how steady is your foot hold...
is the platform on which you are standing moving or vibrating?
is it so cold that you are shivering?
have you run a distance or are you breathing hard because of some other exercise?
are you excited with adrenaline pumping?

I can do a pretty good job of holding a camera steady, mostly because of training and experience. But, I still prefer to shoot at 1/2x the focal length to ensure steady shots. I am willing to boost my ISO to get a sharp image. Noise can be controlled to a large degree in PP but a fuzzy image remains a fuzzy image despite PP sharpening.

The above, of course, is pertinent for shooting without any type of shake control in camera or lens. I use the Image Stabilization of my Canon lenses whenever possible. It is a Godsend to an old man like me!

Finally, if you frame the images the same, an image from a crop camera needs to be sharper than an image from a full frame camera for the same size print. The image from the crop camera will need to be enlarged to a greater degree.

[mikesan]

It's not the resolution or the pixels - it's the crop factor needing taken intae account. Shooting a 50mm@ 1/50th follows the rule but if yer shooting a 2x crop factor 50mm then ye have an angle of view of a 100mm lens so adjust tae 1/100th?

Things ah normally don't say - then wish ah hadn't...crop factor - when ah always (normally) say Focal Length Multiplier. Not only is it a much more elegant phrase, it's much more accurate as well.

I respectively beg to differ on both counts:

George 13 has shown that the sensor size (crop factor) is not a significant factor, assuming the same degree of enlargement of the image. A given amount of motion will move the image on the sensor by a fixed amount, regardless of the sensor size.

I'll take as an example a picture with motion blur. I do that for that's something that can be measured. The motion of the camera results in a motion of the image of x mm. I say mm. First the comparison between the FF and the crop-camera. There is no difference between those concerning motion blur. A 100mm lens on a FF will give the same motion blur as on a crop camera. X mm will be x mm.

I too don't care much for the phrase "crop factor", but the phrase "Focal Length Multiplier" is even more objectionable. It implies that a change in sensor size effects the focal length of the lens which, we all know, is not the case. How about "Field of View Factor". Awkward, but it expresses the idea more exactly.

[Glenn NK]

I ask because I now have four cameras all with different pixel sizes. Suppose I shoot a far-off target with each camera having a non-IS 50mm lens. The rule says 1/Fmm = 1/50 sec or speedier gives a good chance of a blur-free shot, eh?

Should there be an additional factor in "the rule" to account for this?

Ted:

This became very apparent with the advent of the Nikon 800 series; many 'togs had less than really sharp images, and of course with the high MP, they felt that they should be getting very sharp images - they weren't.

The lesson learned (seemed a bit counter-intuitive) was that smaller pixels DO reveal minor movement of the camera more than do larger pixels.

With the introduction of the Canon 7D Mark II, this problem cropped up again.

I recall when it was widely accepted that 8 MP gave at least equivalent resolution to film cameras with colour film. Now we're well into the 40's for MP. If 1/mm was adequate for film under most circumstances, then it surely won't be adequate for the newest high resolution sensors.

I'm very leery of "rules" because so many factors are involved; technique probably being one of the important ones. As our friend from Europe demonstrated (can't recall his name), propping the end of a 200 mm lens improved the resolution beyond that achieved with a "standard" tripod mount. I have found that hanging a small cloth bag with a handful of rocks in it over the lens hood of my 70-200 (increased mass = less vibration) has helped considerably when using a tripod.

So, yes, I think there are additional factors to be considered.

Glenn

[dubaiphil]

I think that, like many photography 'rules', this one is a solid guide for full frame 35mm shooters. VR/IS can pretty much screw up the rule if the shooter so wishes.

More important is gaining a full understanding of your gear and your subjects, and making mental notes of what to apply to what situation. Sure, you may be able to shoot down to 1/8th second at 200mm hand held but is that to the detriment of the final image?

Even peeping at your LCD doesn't give a complete picture to final sharpness of frozen motion. So if shooting for example a motorsports event I'd always give the advice of shooting faster that you'd expect first before slowing shutter speeds, just to get some shots in the can rather than a whole set which are slightly out once reviewed after the event.

Shooting street photography I have different scenarios logged in my head and depending on the focal length, distance of subject, direction of their and my travel etc I will dial in a min shutter speed to auto ISO - sometimes up to 1/640th at 50mm. If shooting at very slow speeds I'll also know what I can get away with and shoot a burst to be able to pick a best sharpness of backgrounds from the set.

[george013]

The lesson learned (seemed a bit counter-intuitive) was that smaller pixels DO reveal minor movement of the camera more than do larger pixels.

Glenn

Compared to the D700 any movement is registered with 1.7x as many pixels. It's not only the pro models of Nikon and Canon. The D7000 has the same pixelsize as the D800, and a smaller sensor. So to gain the same results you will have to shorten any shutterspeed with a factor of 1.7.

I want to say something more about the difference between analogue and digital. In analogue the print was the first possibility to judge the image. Analogue there are practical 2 ways to judge: screensize or full size. And both on screen resolution. I wouldn't know how to simulate a printed A4 image on the screen.

George

[arith]

All good stuff from Richard; I would say every image is sharper with a steady hand and I think it best to keep your arms and camera clear of your chest, breathe out, squeeze the button and hold down, do not release until you hear the shutter close.

[dem]

Finally, if you frame the images the same, an image from a crop camera needs to be sharper than an image from a full frame camera for the same size print.

Why? If by "frame the images the same" you mean taking both shots from the same vaunting point but using a shorter focal length on a crop sensor camera to have the field of view identical to that of a full frame camera, then there is no need for the crop camera image to be sharper (if we assume the same megapixel count on both sensors).

The image from the crop camera will need to be enlarged to a greater degree.

The "enlargement" has already been done in camera by the lens! The lens on a crop camera needs to be sharper than the one on a full frame camera. That is true. But what is the connection to the camera shake?

[dem]

Analogue there are practical 2 ways to judge: screensize or full size. And both on screen resolution. I wouldn't know how to simulate a printed A4 image on the screen.

George, you mean "Digital" here, not "Analogue".

There seems to be two camps:
1) people who assess sharpness by looking at a printed A4 (or smaller) image from a foot away
2) people who assess sharpness by looking at 100% pixel-to-pixel reproduction on their LCDs (which is similar to looking at a 2 meter poster from a foot away)

The 1/mm rule is for the fist camp only.

[george013]

George, you mean "Digital" here, not "Analogue".

There seems to be two camps:
1) people who assess sharpness by looking at a printed A4 (or smaller) image from a foot away
2) people who assess sharpness by looking at 100% pixel-to-pixel reproduction on their LCDs (which is similar to looking at a 2 meter poster from a foot away)

The 1/mm rule is for the fist camp only.

Yes, a miswriting(?).
And I agree with you completely.
I just wonder how many people are taking pictures with in mind: I need a A4 or A0 print. Not counting people with an assignment.

George

[xpatUSA]

Ted:

This became very apparent with the advent of the Nikon 800 series; many 'togs had less than really sharp images, and of course with the high MP, they felt that they should be getting very sharp images - they weren't.

The lesson learned (seemed a bit counter-intuitive) was that smaller pixels DO reveal minor movement of the camera more than do larger pixels.

Yes, that was my thought too. I suppose, to be seriously pedantic, one should consider not only the amount of camera shake but it's type, of which there can be several. But then words like "translational", "rotational", "pitch", "yaw", etc. would appear - and people's eyes would glaze over pretty quick, eh?

Still, from an angular POV, a movement of 3.9um in the image plane (irrespective of sensor size) in one camera is most certainly less than a movement of 8.45um in another - given the same effective** focal length. Inversely, it can be said that, for an equal camera movement, more of the smaller pixels would be covered by that movement; therefore, there would be more blur when that higher-resolution image is viewed at 100% before post-processing. Indeed, how could the higher-resolution image not be more blurred under such conditions? Anybody?

"effective" is used in this context to cover focusing at less that infinity.

Although post-processing such as deconvolution or re-sampling downward can "restore" sharpness lost due to the shakes, surely our preference would be for the best possible capture at the moment of truth?

[NorthernFocus]

This became very apparent with the advent of the Nikon 800 series; many 'togs had less than really sharp images, and of course with the high MP, they felt that they should be getting very sharp images - they weren't....

This has long been a mystery to me as there were sensors with higher pixel density on the market previously and no one noted this phenomenon. Apparently any discussion went unnoticed until E.F.Hutton spoke on experience with the D800

[george013]

This has long been a mystery to me as there were sensors with higher pixel density on the market previously and no one noted this phenomenon. Apparently any discussion went unnoticed until E.F.Hutton spoke on experience with the D800

Nikon warned for that with the introduction, See the Technical Guide.

George

[NorthernFocus]

...it can be said that, for an equal camera movement, more of the smaller pixels would be covered by that movement; therefore, there would be more blur when that higher-resolution image is viewed at 100% before post-processing. Indeed, how could the higher-resolution image not be more blurred under such conditions? Anybody?..

While I do enjoy a good technical discussion, from time to time logic suggests that a pragmatic approach may be more productive.

I currently have the benefit of direct comparison between three different bodies each with different pixel pitch/density. In order of increasing density(decreasing pitch), 16MP full frame, 36MP full frame, and 24MP APS-C(1.5x). If one can extract one's self from the minutia and consider things from a statistical point of view, consider this. Same person shooting with same technique, same degree of shaky hands, same lenses, same tripod(when used), same subject matter, from the same stationary and/or moving platforms over a comparative period of two years and tens of thousands of images. On average the sharpness of images shot in conditions where noise is not a factor correlates directly with pixel density. It's as simple as that. While it may be intellectually stimulating to explain it, I'm fairly certain said explanation won't change the resultant imagery.

[ajohnw]

. . or is it hangover from the glorious days of film, when "everybody" shot 35mm cameras with 100 ASA film and rules were needed because folks were perhaps less technical than today, what with the web and that?

I ask because I now have four cameras all with different pixel sizes. Suppose I shoot a far-off target with each camera having a non-IS 50mm lens. The rule says 1/Fmm = 1/50 sec or speedier gives a good chance of a blur-free shot, eh?

Now my pixel sizes range from 18.24um (binned low resolution) down to 4.33um. So, if my hand shakes by some amount, say the equivalent of 9um at the sensor, my feeling is the high-res camera would show some shakes but the low-res camera would not. (Trying to keep it simple here; mentions of Airy, Abbe, Rayleigh, Nyquist, et al are strongly discouraged).

Should there be an additional factor in "the rule" to account for this?

OTOH, is the rule itself dead? Does anybody actually use it?

No is the short answer.

I can't believe it's the real Ted asking this either.

It could be argued that min speeds should go the wrong way against crop factor - pixel densities are generally higher as the format comes down in size and the miss guided thing called reach goes up suggesting an increase in focal length. So a 4/3 for instance rule would be 1/(2*focal length) Obviously rubbish especially now cameras have IS of one sort or another and it's doubtful if IS is better on say canon crop than on canon ff.

It's always been a curious rule anyway and like many something for people to worry about rather than relaxing and taking the shot. An associated rule was the very popular 135mm lenses of old - reckoned to be the longest focal length that could be hand held. I decided to give a miss and bought 100mm and 200mm and never regretted it.

A similar subject is 7x binoculars - chosen because this is reckoned to be the maximum magnification that an average person can comfortably use follow moderate exertion - say walking up a hill. I'd guess most people would cope with 10x maybe more.

Anyway certain other factors other than just pixel sizes that relate to optics and software have effects that can not be divorced from this area.

John
-

[Mike Buckley]

It's always been a curious rule anyway and like many something for people to worry about rather than relaxing and taking the shot.

It has always helped me, so I don't understand your point that I'm unnecessarily worrying about something rather than relaxing. Ironically, I can relax because the guideline has always been so helpful to me.

[ajohnw]

It has always helped me, so I don't understand your point that I'm unnecessarily worrying about something rather than relaxing. Ironically, I can relax because the guideline has always been so helpful to me.

In an odd sort of way maybe I was suggesting people should simply try and see what their own limitations are rather than just applying a rule that wont fit all. Also at the same time that if some one is aware of the rule and tries to go further they are likely to think this is difficult and try to take extra care - not a good idea in my opinion. It important to be relaxed.

Really though given IS the rule doesn't apply now at all. At least on the gear I have used and the implications of format size and reach most definitely don't apply.

John
-

[dem]

In order of increasing density(decreasing pitch), 16MP full frame, 36MP full frame, and 24MP APS-C(1.5x)... On average the sharpness of images shot in conditions where noise is not a factor correlates directly with pixel density. It's as simple as that.

Does this mean that higher pixel density leads to sharper images, and the APS-C camera is the sharpest out of the three cameras you've got? If we take this argument further, would you expect a 1/2.3" 16MP point and shoot camera to produce sharper images than any of your cameras as it has a higher pixel density?

Possibly the correlation is negative. In any case, how do you compare sharpness between different formats? At the same ISO? the same shutter speed? looking at prints of the same size (36Mp ff should be the best in good light)? by viewing them at 100% on the screen (16Mp ff should be the best for camera shake and subject movement)?

Does the 16Mp ff appear sharper simply because it is less noisy at higher ISOs and is more often used at faster shutter speeds than the other two cameras?

just saying that pixel density might not be the answer to everything.

[NorthernFocus]

Does this mean that higher pixel density leads to sharper images, and the APS-C camera is the sharpest out of the three cameras you've got?...

That's what the data suggests.

... If we take this argument further...

Don't. For one thing it's not an argument. Re-read my post. I simply attempted to supply some raw data. Same guy, same glass, same subject matter, three cameras to compare, those are the results. Not making any claims about absolutes or situations other than mine. Anyone interested can look at my work and decide for themselves whether the data is meaningful or if it's rubbish.

After years of painstaking technical analysis of the alternatives, I've been reborn. I think, finally, as a photographer. Now I'm just trying to offer a glimmer to others...