« Nikkor 28mm ƒ/1.8 Follow-Up | Main | Mystery Man? »

Wednesday, 28 November 2012

Comments

Ctein is really right about how different lenses balance differently on a camera, and affect your ability to handhold at various shutter speeds. The 28mm ƒ/1.8G lens feels much lighter and better-balanced on the D800, but it's noticeably easier to hold the camera steady with the heavier 35mm ƒ/1.4G on the camera.

Mike

I too have found that my ability to get crticial sharpness varies a lot with different lenses and camera combinations, and isn't just a straightforward rule related to focal length. When I'm doing shots that don't require precise timing I've started using the shutter delay (set to 1/8s) on my OM-D with good results.

You lost me at "between 28 and 50 megapixels."

You lost me at "between 28 and 50 megapixels."

+1 ...

>> "between 28 and 50 megapixels."

That is the full-frame resolution. APS-C cameras have the same level of pixel density (and technique requirement) for years now.

About the aperture: Sometime we stop down to get better sharpness, but many times we stop down to get more DOF. Comparing OOF areas with diffraction for sharpness, diffraction will win every time. So I am not afraid to stop down to f/11-f/16 for landscapes and f/16-f/22 for macros on my D800e.

Looking forward to the graphic novels off-topic.

Re. diffraction, I assume these numbers are for full-frame and I'd be interested in your thought about how this works with m43, with its lower resolution.

People here and there complain that on m43 diffraction "kicks in" at relatively large apertures. My experience with Oly's 45mm is that it is very uncompromisingly sharp down to at least f/6.3 on my 12Mp camera...

"It applies, in general, to any camera between 28 and 50 megapixels."

I'd think good technique should apply across the board.

Thanks / MS

Your last OT column on tea sent me down an expensive path. Just today the new Tea Source catalog came to cause me further grief. Stick to the cheaper indulgences, like lenses;-)

Re: Diffraction, I'm a low megapixel peasant (a mere 21.1 MP), and I've found that diffraction is noticable at f11, but only if you compare an f11 shot with an f8 shot taken at the same time. F8 has a bit more "sparkle", but f11 is perfectly acceptable. f16 is usually noticably soft and lower contrast, even without a side-by-side comparison, but if you can't work with an f16 image, the problem is you, not your aperture.

Smaller than that and diffraction can be quite noticable, but measurebators are often amazed at how shots I take at f/minuscule (or smaller) are sharper, crisper, and more lively than their images at whatever opto-magical aperture they've decided delivers the most per-pixel IQ.

I hope you don't mind, but I'll just link to three shots that are waaaaaaaay past any reasonable aperture for image sharpness, but still manage to look nice.

Daphnia, f48-ish, nominal f8:

http://www.dementlieu.com/images/fauna_aqinv_MG_9195_1610x1.jpg

Slime mold on grass, f32-ish, nominal f5.6:

http://www.dementlieu.com/images/_MG_8051.jpg

Moldy apricot, f32-ish, nominal f8.0:

http://www.dementlieu.com/images/fungi_MG_9053_2510x1.jpg

Dear Zafar,

No, this isn't format-sensitive. It's entirely about pixel counts. Same recommendations hold regardless of the size of the sensor. Strange but true.

But you're entirely right about depth of field. That's the reason that I mentioned that the D800's can deliver more than twice what's considered “acceptable sharpness”. Even by my standards for sharpness, which are about 75% tighter than normal, I'd have to stop down to f/19 before diffraction was as big a problem as out of focus areas.

Love your pelican picture, by the way!


pax \ Ctein
[ Please excuse any word-salad. MacSpeech in training! ]
======================================
-- Ctein's Online Gallery http://ctein.com 
-- Digital Restorations http://photo-repair.com 
======================================

Dear Cyril,

Peculiarly enough, this aperture stuff doesn't much change with format, at least over the range from quarter-scale to medium format sensors. In the majority of cases, f/5.6 will look really great, f/8 just a little less so, and f/11 definitely less so. Although there's nowhere as much of a change as diffraction-fetishists would think; it's only one component of sharpness. Throw everything into the mix and things tend to equalize.

When I originally wrote the column, I had a lot of “full frame” qualifiers in there. Reading it over, I realized that most of them were not germane, so I edited them out.

Speaking in broad terms, of course, but then this is a basics column so those are the only appropriate terms.


pax \ Ctein
[ Please excuse any word-salad. MacSpeech in training! ]
======================================
-- Ctein's Online Gallery http://ctein.com 
-- Digital Restorations http://photo-repair.com 
======================================

Dear MichaelS,

No, not really. Some things do and some things don't. For example, with 12-16 megapixel cameras (format size is irrelevant) tripods are rarely necessary. I've got an ultra lightweight carbon fiber job that I can carry with me everywhere–– weighs only 3 pounds. If I'm out doing “dedicated” photography, I usually have it with me. It's amazing how rarely I use it. I'd say that on the majority of my outings, it doesn't get used at all. I have to say that I find myself more often using a tripod for framing and composition reasons than for sharpness.

Depending on the lens I'm using, I'm good (sharpness-wise) down to anywhere between 6 and 10 stops below daylight illuminance. (The “Moon and Bay Bridge" photograph that Mike and I sold several months back as an illustration of meticulous technique was handheld.)

So, if I were writing this column for that level of resolutions, I'd be recommending ways that people can work on their techniques and on choose appropriate apertures, ISOs, and shutter speeds, rather than exhorting them to be regularly using a tripod.

Conversely, if you're well-heeled enough to be working in the 60-100 megapixel range, a tripod turns out to be necessary under the majority of circumstances (putting aside the fact that those high-megapixel cameras are a real pain to handhold).


pax \ Ctein
[ Please excuse any word-salad. MacSpeech in training! ]
======================================
-- Ctein's Online Gallery http://ctein.com 
-- Digital Restorations http://photo-repair.com 
======================================

Thank you, Ctein, for this article. I'm one of those who can't read enough of the basics in my favorite daily "magazine."

May I ask if you would you recommend Olympus Pens—in particular, the discontinued EP-3 or the new EPL-5—as a "back-up" body for adapted M-lenses? I think either Pen body (can't afford the OM-D) will "balance" well with my Zeiss wides and CV short tele. Too, I'm after the Pen's in-body stabilization. (The 2X crop factor also means I'll have a wide, normal, and telephoto MF optics ready to hand.)

I'm asking because my Ricoh Mount-A12 lens module went electronically bonkers after about 10,000 "actuations." (It is being repaired under warranty in another country.) I realized too late that the only digital bodies with native M mounts are the Leicas! I don't want to unload my M lenses and haven't given up on the GXR just yet.

Looking forward to your OT columns. Thanks!

I think a lot of photographes get hung up on "sharp as possible" rather than "sharp enough". The former is objective. The latter not, with a several factors driving the "enough" part.
The thing that always makes me chuckle on this is how, with high pixel counts, almost everyone wants to increase enlargement factors by printing to 300ppi at native pixel count, then complains about unsharp photos.

After the tripod everything else is a refinement.

Dear Sarge,

Sorry, not a clue whether adapted lenses like that work well in practice; I don't use any.

Maybe some other reader has experience with this?

pax / Ctein

@ Sarge: In principal, attaching a fine Leica wide-angle to an E-P3 or E-M5 would make for a lovely, stabilized relationship. But my own experience with this (since the E-P1) has been disappointing due mainly to two factors.

First, the Oly micro 4/3rd bodies make manual focusing with an un-coupled manual lens a chore. Unlike Sony's NEX cameras, the E-Px does not feature a focus peaking display. Nor do they automatically magnify the display when such a lens's focus ring is turned, as the do with native coupled lenses. So you must have very keen close-range eyes and fiddle with the display settings to accurately pull critical focus with, say, a Leica or Zeiss lens. Of course if you're using a wide lens at a generous distance from your subject you can just focus at infinity and call it a day.

But, secondly, even under the best conditions a lovely Leica or Zeiss M-mount lens does not necessarily produce a sharper or appreciably better-rendered image than a micro 4/3 lens on that E-P3. It seems that lens recognition and correction counts for quite a bit in these little cameras, and they know how to dress-up their own lenses well.

With a substantial assortment of wonderful M-mount lenses nobody was more eager to couple them to new, stabilized bodies than me. But the plain, bald fact has been that the results are generally not worth the effort and inconvenience on a micro 4/3 camera.

If adapting M-mounts to a fine contemporary body is something you're really enthusiastic to pursue I recommend you choose a Sony NEX body. Yes, you lose that stabilizer but you will find focusing much easier and, in my experience, get a better result slightly (not greatly) more worthwhile the effort.

You lost me at "between 28 and 50 megapixels."

You lost me at 'pixels'!!

Technical expertise coupled with practical experience and clear thinking often leads to good common sense advice.

Posts like this may be "basic", but they're also a breath of fresh air. It's nice to be reminded on occasion that sound technique works just as well with today's wonder-cameras as it always has. Thanks, Ctein.

+1 for the tripod, haven't taken a picture for money in 30 years that wasn't on a tripod, and probably most of the pictures I've taken not for money...

Good to note here that there are a lot of terrible tripods out there as well: too flimsy, not high enough, or too spindly when all the way up, and poor lock downs. It's pretty common knowledge among pros that if your tripod is a drag to carry, then it's the right one! Any large format user will also tell you, you might need two for 8X10, or at least an additional light stand you can jam under the front or back, and that making the camera on a tripod slightly "off-balance" will settle the camera down faster in a breeze or if the tripod is kicked.

Camera balance is also a main contention, and why I hate, hate, hate zooms the size of coffee cans, and why I keep asking for primes in the f/2.8 sizing. Shot a whole job on a Nikon APS-C and their 35mm f/1.8, and thought: "Hey, this is the way it's supposed to be!". Again I say: "where are my APS-C sized f/2.8 primes in 16mm and 24mm?".

Sarge,
My experience with adapted lenses on a Pentax k10 was that when you tell the camera what focal length the lens is, it is a good idea to test several different focal length settings for the in body stabilization for a particular lens. It seems that the camera makers need to make some assumptions about the overall weight of the lens which affects the frequency of vibration, the center of mass for the camera lens combo, the entrance pupil of the lens among other things. IE a 200mm non telephoto lens behaves quite differently than a 200mm telephoto. How you hold the camera seems to be an issue with the in camera stabilization, holding the camera with one hand supporting the front of the lens seems to decrease sharpness compared to just holding the camera body. On the other hand maybe the in camera stabilization simply wasn't working to spec in that camera.

I finally came to the conclusion that my photos were sharper with the stabilization turned off, but I'd like to try a live-view camera with sensor stabilization to get a better idea what is going on.

My three tricks for sharper hand held slow shutter photos are:
increase the polar moment of inertia of the camera by attaching a collapsed monopod or tripod to the camera or just bolt a 2x4 to it.

If you are using an eye level camera , try to have the camera make firm contact with your head. Holding it upside down with the body pressed against your forehead works. I used to have a small block of wood with a accessory shoe attached that I would put in the hotshoe of the camera so it would make firm contact with my head. You are effectively adding 5 pounds to the mass of the camera. Of course this works even better if you are a chicken ( I'm not kidding, google camera on chicken )

Waist level viewing with a taut neck strap beats all the above.

The best advice for hand held marginal shutter speeds is to shoot three exposure bursts then pick the sharpest one later. Usually it turns out to be the second one but not always.

Tripods work best of course*, the heavier the better. Once upon a time the hot setup was to fill the center column of a tiltall with lead. If you aren't using a heavy tripod usually exposure times longer than 1/2 second are sharper 1/4.

*unless you are on a bridge or a dance floor where handheld with knees bent works better.

3 things to add:

1) the weight of the camera/lens affects how easy it is to hold steady and therefore the resulting camera shake and sharpness

2) surely ones ability to hold a camera still is very dependant on format due to the vastly different amount of magnification needed to produce the final image - so a medium format negative needs less magnification to produce the viewable image and the shake is therefore magnified less too, whereas a digital compact image is magnified a lot to produce the same sized final viewable image, thus magnifying the shake much more

3) regarding diffraction - diffraction is caused by physically small aperture openings. So f22 on MF has an aperture opening much larger than that on smaller format models. So on MF diffraction is not really a problem. Please tell me if I'm wrong

I'm enough of an amateur that sharpness and resolution are quite important to me. In my non-rigorous experiments of photographing a bunch of subjects over the years and then looking at the results, I can summarize four principles that I find important:
(1) A tripod is an immense help with any shutter speed that isn't "fast".
(2) Getting the focus right is more important than DoF, if you have a single important point for the viewer's attention.
(3) Getting the DoF right is more important than diffraction effects, unless you want to go down the dark and arcane pathways of focus-stacking.
(4) If your subject is moving, principles (1) through (3) and any guidelines you've developed for hand-holding with or without image stabilization will likely go out the window.

Of course, you may not always want a sharp picture of a moving object, or be able to get it even if you want it...

Dear Michael,

1) It's complicated. The more a camera weighs, the more its inertial mass damps out vibrations and tremors. But the more a camera weighs, the harder it is to hold. Every individual has a sweet spot that represents the optimum trade-off between damping and what they can comfortably carry. There is no one answer. Also, physical size matters. If the camera is too large or too small to comfortably fit in your hands, your muscles work harder to hold onto it. Again, no universally correct answer.

2) This column is about digital cameras, not film cameras. Sensor-to-print magnification is not relevant in digital cameras; it's entirely about the number of pixels you are mapping to the final print. Many folks get this wrong. But that's why I wrote, “It applies, in general, to any camera between 28 and 50 megapixels."

3) That's not really the way diffraction works for general photography. f/22 produces the same number of diffraction-limited line pair per millimeter (about 75) regardless of the absolute size of the aperture. So, it connects to pixel size, but not format size. To learn more about how diffraction really works in both theoretical and practical terms, read these two columns:

http://theonlinephotographer.typepad.com/the_online_photographer/2008/11/diffractionquan.html
http://theonlinephotographer.typepad.com/the_online_photographer/2008/11/diffraction-in.html

The comparison illustrations there, by the way, were made with a 12 megapixel quarter-scale sensor camera which, according to the diffraction fanatics, ought to look like total crap at f/8. Not so much! As I said in this column, diffraction in the real world does not produce the same effects on image quality that it does in theory.


pax \ Ctein
[ Please excuse any word-salad. MacSpeech in training! ]
======================================
-- Ctein's Online Gallery http://ctein.com 
-- Digital Restorations http://photo-repair.com 
======================================

Michael Kay, you are wrong. If the physical size of the aperture was what mattered, you could shoot a 400mm lens at f22 (18mm aperture) and get a picture as sharp as a 100mm lens at f5.6 (17mm aperture), but a 24mm lens at f8 (3mm aperture) would be well past a hope of sharpness.

Discussions of diffraction are all based on assumptions about final print magnification. The reason you can stop a MF lens down more is because the 645 (just to pick an example) negative is going to be less enlarged in an 8x10 print (to pick another example) than a 35mm one.

Hugh Crawford -- there are a number of techniques for cameras using tension in the strap to help stabilize things. (Also some standard techniques for the same thing in rifles; the requirements are the same, holding it pointed exactly where you want while gently squeezing off a "shot", without of course jerking.) A standard for TLRs of course.

I find something similar works great for small cameras with an LCD on the back. Grip the camera firmly at the sides, pull your elbows in against your tummy, and push the camera out against the strap by bending at your elbows. Works much better for me than trying to hold a flat camera up to my round head. (Works better for small cameras, P&S size).

"Sensor-to-print magnification is not relevant in digital cameras; it's entirely about the number of pixels you are mapping to the final print." Ctein

"Discussions of diffraction are all based on assumptions about final print magnification." James Sinks

What we have here is a failure to agree. I await resolution.

Dear D.,

James was talking about film. I explicitly said digital operates by different rules.

Always read in context. Sentences are embedded in paragraphs for a reason.

pax / Ctein

Interesting article, I'm still trying to find out exactly how or why two consecutive snaps come out where one is sharp and the other isn't. Sometimes it's obviously a mis-focus, other times just doesn't seem to make sense.
Thanks for the writeup!

Dear Ctein,

Thank you for your good offices. The short course on "basics" has been transformed into a seminar on applied theory and advanced technique by the discussion. Just what I needed.

Kenneth,

Thanks for the tip re the Pens. I didn't realize they have no focus peaking. I'll get the GRD IV instead as my back-up and AF camera. I'm already familiar with it's UI which I'm liking a lot on the GXR. If not fixed for good, I'll switch to the Nex. I hope it will be trouble-free until its upgrade or the Nex 8 comes along (when the Nex 7/6 can be had at a discount).


Hugh,

Thanks for your "tricks" on how to get sharp, handheld photos. Your DIY stabilization workarounds also inspired me to get the HoodLoupe (for 3" LCD screens) for eye-level shooting with my camera which has no integrated viewfinder. The GXR has no Ibis; it's the one with the ADJ button-toggle. You called it "adjective button" when Mike reviewed it sometime ago.

The chicken steadycam link is a hoot. I'm rehearsing my antero-posterior "VOR" head moves. Handy for when your lens focus-shifts.

The comments to this entry are closed.