[sterretje]

The following images are taken with a refurbished K100D on a tripod and 2 sec delay, Taken in RAW, converted to JPEG using UFRAW without additional processing (except for resizing) and rotated (only the first one) in GQview; the latter did not work to well by the looks of it, so please do not comment on that issue.

The crops were converted to tiff8 using UFRAW without additional processing, cropped in GIMP and saved as JPEG.

Taken with FA31LTD
f/1.8 (distance approx. 8 meters)


f/1.8 (distance estimated 3.5 meters)


f/1.8 crop (assumed focus point) of the latter

Totally OOF (the DOF at this distance should make the flower look sharp)

f/1.8 (distance about 70 cm)

f/1.8 crop

The focus point is the indicator and there seems to be a slight focusing issue here (not as bad as the other pictures).

f/1.8 (distance unknown; taken in shop, handheld)

f/1.8 crop

Focusing is on the bottom of the 'L'; I can not say if it's 100% OK or slightly OOF as well.

And a shot and crop with the 55-300mm handheld
f/5.8 (distance approx 1.4 meters)

f/5.8 crop

Focusing looks perfect to me.

My questions:

1. With regards to the focusing in the first 2 pictures, am I expecting to much? Can't the system properly focus on this type of subjects; it looks quite good in the viewfinder though.
2. Looking at the image with the indicator, there is a slight focus problem. Will this slight focus problem result in the very bad focusing of the first set of pictures.
3. Can I draw conclusions from the fact that the 55-300mm gave a perfect focus?

Testing with the kit-lens at about 31mm and f4 gives more or less the same results as the FA31LTD. I have a K10D that seems to be worse at very close focus (40cm or so), similar for the shown images and the same result at 300mm. The K10D was in for repair, I took a test shot with the FA31LTD in the repair center that was very good, but now I battling again with it.

I'm getting so frustrated with this and no longer know what to do. Buy a new K10D, K-x, K20D or K7 or throw everything on the dump and start shooting with my Minolta X700 again.

---

Last edited by sterretje; 02-07-2010 at 09:07 PM. Reason: corrected f/1.8 for the 55-300mm to f/5.8

[Marc Sabatella]

[*]With regards to the focusing in the first 2 pictures, am I expecting to much?

Depends. If you're expecting the camera to read your mind regarding which branch to focus on, then yes - bionics technlogy has not come that far. If you're thinking DOF at f/1.8 should have been enough to put the whole tree perfectly in focus, then yes, you're expecting too much - physics doesn't allow for that.

[*]Can I draw conclusions from the fact that the 55-300mm gave a perfect focus?

Yes - you can conclude that DOF is much deeper at whatever aperture you actually shot at - definitely *not* f/1.4, since the 55-300 cannot come anywhere close to that.

[stevebrot]

Marc is absolutely correct:

• DOF at f/1.8 is VERY thin
• Even the most accurate AF systems can't read your mind as to where the focus point should be

To this I would add:

• The AF focus "point" is much bigger than the red indicator dot in the viewfinder

The bottom line is that if you require fine focus control in a complex subject, your eye is probably the better focus device. In addition, unless you are using shallow DOF as a part of your composition, adequate DOF starts at f/5.6 for subjects at moderate distance.

I would suggest spending some time with your camera and lenses doing some table-top photography. Simple subjects in a controlled environment at various apertures using both AF and manual focus.

Steve

P.S. Your X700 probably gave better results because you were using manual focus with an excellent viewfinder and had access to split-image/microprism focus aids. You also had a limited ability to pixel-peep

P.P.S. Your 55-300 is f/1.8???!!!

[dragonfly]

Another problem is that you're shooting wide open in full daylight. Lenses are typically soft wide open.

I suggest that you try your lens wide open somewhere that is darker in order to take advantage of the wide aperture. You might be more impressed with it then.

[mattdm]

Good advice above. But also: before you throw in the towel, get your lens and camera body calibrated together. Pentax service can do this for you, or, with the K20D or K-7, you can make the adjustments yourself.

[GoremanX]

+1 to all above posts.

Focusing looks fine to me in all pictures. The crops you took are from where you wish the camera had focused, not necessarily where it actually did focus.

[GerryL]

..and if you always keep on pixel peeping, you will never get satisfied as you will almost certainly find something wrong even if it is not obviously so or negligible.

[sterretje]

Hi people, thanks for the replies.

First of all, I was convinced that I did change the aperture to 5.8 as I noticed the error before posting, but obviously I didn't. Marc's comment on the 55-300mm did initially not make sense till I saw Steve's comment. I have corrected it now (f/5.8).

Further I understand the effects of focal length, aperture and subject distance on DOF. I know that at f/1.8 the DOF can be very thin, but that applies more to close-ups than to longer subject distances. After the initial disappointment I decided to check the DOF with http://www.dofmaster.com/dofjs.html; for the lemon tree, it is -182cm/+335cm and for the bush it is -40cm/+52cm with the FA31LTD@f/1.8 and for the yellow bolt it is -2mm/+2mm with 300mm@f/5.8. Assuming that these calculations are correct and from your comments the following:

The lemon tree is probably just lacking some expected detail. The bush in my opinion is OOF; the root cause being me not understanding what the camera does with regards to focusing. Two posts indicate that I can not really rely on the red sqare *which I indeed do. First question is how big the focus area is (is it the size of the 'circle' or smaller/bigger and does it apply to all focus points) so I can take that into account for future photos. The second one is how I can predict where the camera will decide to focus; the answer is probably that I can't and that might imply that I need a split screen (easier) or need to learn to use the viewfinder (never been able to do that reliably with the X700) and use manual focus without relying on the focus indicators for photos similar to the first two shots.
Note: As I could remember where the red focus square was for the second shot (the flower in the middle) I decided to crop that one.

I'm indeed guilty of pixel peeping here. This simply happened because I was not happy with the results at the moment that they showed on the monitor and decided to have a 'closer' look. I will also pixel peep while trying to judge if an image is good enough to be printed on A4. I will not attempt to print the bush one on that size. I however have close-up shots (table-top, simple subjects with 31mm@f/1.8, 40cm distance) that are good enough to print on A4.

The slik board and the yellow bolt were added as references. As I commented that I can not judge if the former is OK or slightly OOF, I'm happy with that result. The bolt for me is excellent with regards to focus.

Regards WimS

[Marc Sabatella]

The bush in my opinion is OOF; the root cause being me not understanding what the camera does with regards to focusing. Two posts indicate that I can not really rely on the red sqare *which I indeed do. First question is how big the focus area is (is it the size of the 'circle' or smaller/bigger and does it apply to all focus points) so I can take that into account for future photos.

The center point is "about" the size of the etched inner brackets on the focus screen, You can test for yourseff by putting a black dot on a piece of white paper and seeing if the camera can focus on it from a variety of different positions. If it can, then the dot must be within the range of the sensor; if not, then it isn't.

The second one is how I can predict where the camera will decide to focus; the answer is probably that I can't and that might imply that I need a split screen (easier) or need to learn to use the viewfinder (never been able to do that reliably with the X700) and use manual focus without relying on the focus indicators for photos similar to the first two shots.

Pretty much true, yes. Of course, the real answer here is, don't shoot wide open when there is no need to. But when there is such need, know DOF will be extremely thin - thinner, in fact, than the viewfinder will be able to show. meaning it is normal for some things to appear in focus in the viewfinder but not turn out to be in the picture. It takes practice to learn to compensate - to see the whole one of apparently-acceptable-focus in the viewfinder and to place your subject within it so that it is in the part that is *actually* in focus. It can be done, but it's definitely a bit of an art.

BTW, note that DOF calculations are based on assumptions as to how big you will print and how from how far away you will view the image. When pixel peeping at 100%, they tend to be overly optimistic.

[sterretje]

Thanks Marc.

You're talking 'inner brackets', do you mean the 'circle'? I can not imagine that you mean the square brackets as the already contain the numerous focus indicators.

With regards to your comment on DOF calculations. If one has a huge painting (think the 'nightwatch') people tend to get close to look at the details. The same applies to bigger than normal photo prints (A4/A3). I have a 45x30 (or slightly bigger) photo hanging on the wall (over 10 years old now) and still get close at occasion to look at the detail. I'm still wondering how I ever took that shot (partly knowledge, partly luck I guess).

[Marc Sabatella]

Well, it's not a circle on my camera - a pair of rounded parens with a gap between them is more like it. But it does suggest a circle, yes.

As for lookng at large photos up close - sure, people do that occasionally, but normal viewing distance is from farther away. Anyhow, my point wasn't that you shouldn't look close - just that DOF calculations don't account for that. They assume "normal" viewing distances.

[sterretje]

Thanks again.

'circle' was between single quotes to indicate that it was not a real (read full) circle. I will take this info into account when focusing.

PS the yellow bolt did fill the 'circle' completely which does explain the excellent focus there.

[stevebrot]

Further I understand the effects of focal length, aperture and subject distance on DOF. I know that at f/1.8 the DOF can be very thin, but that applies more to close-ups than to longer subject distances. After the initial disappointment I decided to check the DOF with http://www.dofmaster.com/dofjs.html; for the lemon tree, it is -182cm/+335cm and for the bush it is -40cm/+52cm with the FA31LTD@f/1.8...

The calculations are correct, but the calculator page left out one small qualifier that may be found along with a similar calculator on Bob Atkins' Web site. In regards to determining an appropriate value for circle of confusion, "...this is based on the appearance of a standard sized print at a standard viewing distance. While the "standard" is rarely explicitly defined, it seems to be close to an 8x10 print viewed from about 1 foot."

The size of and viewing distance to the final image is a factor in determining DOF. When you enlarge the image (pixel peeping), you essentially shrink the DOF to a fraction of what the calculator says it should be. Now you know why no one is happy with their lens performance. Everyone is evaluating sharpness on out-of-focus images.

Back in the day when we were taking a 35mm negative and generally going no larger than 11x16 inches (sorry for the English units), we never had a problem with this sort of thing. It took a combination of excellent technique, great taking lens, fine grain/high acuity film, quality enlarger, and great enlarging lens to get a respectfully sharp print at that size and we were thrilled when we got it. Today, we can "enlarge" a much smaller image to the equivalent of wall size and examine it with a magnifying glass. DOF...Ha! Ha!

That being said...I agree that your first two images are unreasonably soft. To be honest, they look more appropriate for a P&S as opposed to a top-shelf SLR with a world class optic. Something is not right. Colors and contrast are off as well. Are we sure that something awful did not happen in PP or RAW conversion?

Steve

[ma318]

--------from dofmaster faq------

What is the meaning of circle of confusion?

Circle of confusion is difficult to explain in non-technical terms. You shouldn't worry about it if you're just beginning to learn about depth of field. The circles of confusion listed on this page are reasonable values that you can use to get reasonable estimates of depth of field.

Note that the circle of confusion is subjective. Thus, some people use 0.025 mm as the circle of confusion for 35mm format, while others will use 0.030 mm. There isn't a "correct" value. The values of circle of confusion on this website are calculated using, approximately, the circle of confusion that Canon uses for calculating the depth of field tables for their lenses (Canon Camera Museum - Len Details).

See Anthony Collins' article "Aperture and Depth of Field - how it works" for a somewhat technical explanation of circle of confusion.


Depth of Field FAQ

Circles of Confusion for Digital Cameras

[stevebrot]

--------from dofmaster faq------

...Note that the circle of confusion is subjective...

Probably the most accurate thing that can be said about the subject of DOF! It is the range of adequate sharpness for a person with average visual acuity at normal viewing distance. The whole concept is extremely subjective.

I have found this paragraph from the Wikipedia article particularly helpful (emphasis mine):

Precise focus is possible at only one distance; at that distance, a point object will produce a point image.[1] At any other distance, a point object is defocused, and will produce a blur spot shaped like the aperture, which for the purpose of analysis is usually assumed to be circular. When this circular spot is sufficiently small, it is indistinguishable from a point, and appears to be in focus; it is rendered as “acceptably sharp”. The diameter of the circle increases with distance from the point of focus; the largest circle that is indistinguishable from a point is known as the acceptable circle of confusion, or informally, simply as the circle of confusion. The acceptable circle of confusion is influenced by visual acuity, viewing conditions, and the amount by which the image is enlarged (Ray 2000, 52–53). The increase of the circle diameter with defocus is gradual, so the limits of depth of field are not hard boundaries between sharp and unsharp.

In any case, it is enough to say that calculated DOF is meaningless when you pixel peep because you are viewing at other than the standard amount of enlargement.

Steve