"Pain of focus?" - Pentax k-x with smc DA 18-55 mm / 3,5~5,6 AL WR

Very new to this forum (but reading dpreview since months) I'll just start with a beginners question:

Why does the image look that blured especially on the left side (right side is not that bad at all)? But I thought I did focus more in the middle of the picutre. Is this the best I can get and expect with this cam and this lens here?

Pentax k-x with smc DA 18-55 mm / 3,5~5,6 AL WR:





For comaprison I shot the same scene with a compact cam. My fujifilm f50fd delivers:





Of course colors are much better with k-x but my concern is focus, focus accuracy, sharpnes and resolution. f50 as a compact cam has no problem with DOF since focal lengt 8mm is very short and almost everything in the scene will lie in the field of focus.

DOF certainly is shorter for the APS-C cam at 24mm but I would not expect the image to be generally more blured on the left side. Blured by shake? Or am I just fooled by the scene and everything is pretty ok in the picture?

Any hints welcome

P.S.: I've read Tim's thread ("Lens choice for K7 makes me rethink the body" http://forums.dpreview.com/forums/read.asp?forum=1036&message=35353945 ) with much interest but I'm still not sure if another lens would help much or was needed - maybe there's only some problem with my copy? (Or maybe I simply didn't manage to get all out of cam and lens ...)

roland66 said:

Why does the image look that blured especially on the left side (right side is not that bad at all)? But I thought I did focus more in the middle of the picutre. Is this the best I can get and expect with this cam and this lens here?

Click to expand...

normally i would think you're just looking at DOF, but the building behind the trees and the building on the left side is nice and sharp. and it doesn't do it on the right side (or the top/bottom) so it's probably not just the way the lens renders images. so, something could be misaligned somewhere.

find a nice flat surface with some detail, like a brick wall, and shoot a straight shot of it. that should help make apparent if there's a problem.

normally, i'm not for the brick wall test, but it's the best way to see if you have this particular problem.

roland66 said:

Very new to this forum (but reading dpreview since months) I'll just start with a beginners question:

Why does the image look that blured especially on the left side (right side is not that bad at all)? But I thought I did focus more in the middle of the picutre. Is this the best I can get and expect with this cam and this lens here?

Pentax k-x with smc DA 18-55 mm / 3,5~5,6 AL WR:





For comaprison I shot the same scene with a compact cam. My fujifilm f50fd delivers:





Of course colors are much better with k-x but my concern is focus, focus accuracy, sharpnes and resolution. f50 as a compact cam has no problem with DOF since focal lengt 8mm is very short and almost everything in the scene will lie in the field of focus.

DOF certainly is shorter for the APS-C cam at 24mm but I would not expect the image to be generally more blured on the left side. Blured by shake? Or am I just fooled by the scene and everything is pretty ok in the picture?

Any hints welcome

P.S.: I've read Tim's thread ("Lens choice for K7 makes me rethink the body" http://forums.dpreview.com/forums/read.asp?forum=1036&message=35353945 ) with much interest but I'm still not sure if another lens would help much or was needed - maybe there's only some problem with my copy? (Or maybe I simply didn't manage to get all out of cam and lens ...)

Click to expand...

Looks fine to me.

I was all set to say that in good light there should not be THAT much difference and the dslr will pull away when you get away from that.

Just yesterday I paid $15 Australian (about $12 US) for a 5mp Kodak easyshare c340 from a pawn shop, never a top line camera even when new ....photos from it are not great for the most part (compared to my K-x or Nikon D50) but given time to focus ...maybe a couple of times...and with flash it has a very sharp little lens and I will take it for times when I do not need a dslr and if I break it ....so what.....ahs some useful scene modes as well.

Back to your shots....the K-x one actually to me looks MUCH better than the fuji....I am actually dissapointed with that, I thought it would do better.

Look at the building in the centre...no comparison regarding sharpness (is this where you say oops i made a mistake and the fuji is the top one?).

There can be lenses that are sharper on one side than others....that is a faulty lens but I do not see that in those shots.

The Pentax kit lens is highly regarded but yes, I would look to better lenses.
Disclaimer I do not have the kit lens.

EDIT Having said that, due to the small size posted (when clicked on)l size both pics seem too pixelated to tell for me.
Maybe ask in the Pentax forum.

neil
http://www.flickr.com/photos/26884588@N00/

neil holmes said:

roland66 said:

Very new to this forum (but reading dpreview since months) I'll just start with a beginners question:

Why does the image look that blured especially on the left side (right side is not that bad at all)? But I thought I did focus more in the middle of the picutre. Is this the best I can get and expect with this cam and this lens here?

Pentax k-x with smc DA 18-55 mm / 3,5~5,6 AL WR:





For comaprison I shot the same scene with a compact cam. My fujifilm f50fd delivers:





Of course colors are much better with k-x but my concern is focus, focus accuracy, sharpnes and resolution. f50 as a compact cam has no problem with DOF since focal lengt 8mm is very short and almost everything in the scene will lie in the field of focus.

DOF certainly is shorter for the APS-C cam at 24mm but I would not expect the image to be generally more blured on the left side. Blured by shake? Or am I just fooled by the scene and everything is pretty ok in the picture?

Any hints welcome

P.S.: I've read Tim's thread ("Lens choice for K7 makes me rethink the body" http://forums.dpreview.com/forums/read.asp?forum=1036&message=35353945 ) with much interest but I'm still not sure if another lens would help much or was needed - maybe there's only some problem with my copy? (Or maybe I simply didn't manage to get all out of cam and lens ...)

Click to expand...

Looks fine to me.

I was all set to say that in good light there should not be THAT much difference and the dslr will pull away when you get away from that.

Just yesterday I paid $15 Australian (about $12 US) for a 5mp Kodak easyshare c340 from a pawn shop, never a top line camera even when new ....photos from it are not great for the most part (compared to my K-x or Nikon D50) but given time to focus ...maybe a couple of times...and with flash it has a very sharp little lens and I will take it for times when I do not need a dslr and if I break it ....so what.....ahs some useful scene modes as well.

Back to your shots....the K-x one actually to me looks MUCH better than the fuji....I am actually dissapointed with that, I thought it would do better.

Look at the building in the centre...no comparison regarding sharpness (is this where you say oops i made a mistake and the fuji is the top one?).

There can be lenses that are sharper on one side than others....that is a faulty lens but I do not see that in those shots.

The Pentax kit lens is highly regarded but yes, I would look to better lenses.
Disclaimer I do not have the kit lens.

EDIT Having said that, due to the small size posted (when clicked on)l size both pics seem too pixelated to tell for me.
Maybe ask in the Pentax forum.

neil
http://www.flickr.com/photos/26884588@N00/

Click to expand...

Just clicked on them again and this time they are fine ....must have been a glitch on my pc.

Edit (again) With regards to the sharpness on the edge....while not super sharp compared to the centre it is not bad and seems better than the fuji in the same place as well....the other side you can not see as far back on the edge to compare but the top of the building on that side at the back seems about the same.

Most wide angle lenses will be a little less sharp at the edges.
neil
http://www.flickr.com/photos/26884588@N00/

Thanks for your ideas. Misalignment might be.

arachnophilia wrote:
...

arachnophilia said:

find a nice flat surface with some detail, like a brick wall, and shoot a straight shot of it. that should help make apparent if there's a problem.

normally, i'm not for the brick wall test, but it's the best way to see if you have this particular problem.

Click to expand...

Yes, i guess this is what i'll have to do. Find a more simple scene where it's more clear what i should expect and compare it with what I get.

neil holmes said:

...

Back to your shots....the K-x one actually to me looks MUCH better than the fuji....I am actually dissapointed with that, I thought it would do better.

Look at the building in the centre...no comparison regarding sharpness (is this where you say oops i made a mistake and the fuji is the top one?).
...
Maybe ask in the Pentax forum.
...

Click to expand...

Slight disappointment with the fuji I guess is mainly because of dynamic range and exposure. The scene has high contrast (sky included) and k-x tends to over expose what the f50fd does not. And of course f50fd has comparatively very limited dynamic range. So the dark parts in the scene look so much worse for f50fd in this comparison.

The part where the f50fd is better regarding sharpnes is the the "Hotel" (roof an building) and all that left side of the picture. So if the scene was brighter there I'd expect better colors from f50fd there but k-x still would be blured.

I can't see a problem with the first pic. If there was depth-of-field or focus issue - i.e. you had focussed on something close and the trees further away on the left were out of focus - then the building in the background (hotel) would be even more out of focus, but it isn't. (The lettering 'h-o-t-e-l' is quite clear).

Was there any wind? Could the trees have been moving slightly in the breeze causing a trace of blur?

As another poster said - if you are worried, set the camera on a tripod square-on to a brick wall (or a sheet of newspaper) and shoot that to see if you have an obvious difference between left and right sides of the frame.

Best wishes
--
Mike

roland66 said:

Very new to this forum (but reading dpreview since months) I'll just start with a beginners question:

Why does the image look that blured especially on the left side (right side is not that bad at all)? But I thought I did focus more in the middle of the picutre. Is this the best I can get and expect with this cam and this lens here?

Click to expand...

Having been told we can be clueless, I would venture the following gueses this time.

• There is a shutter speed difference between the two - how about wind / breeze shaklng the leaves? The building in the Pentax is sharp, the Fuji unsure

• User test not valid? Not enough sample size (i.e. not enough testing), test subjects not a good target, AF issues, not on tripod, effect of subject movement etc.. etc...

• Are you using a UV filter on the Pentax?

Aside from lens problem.

--



Ananda
http://anandasim.blogspot.com

'There are a whole range of greys and colours - from
the photographer who shoots everything in iA / green
AUTO to the one who shoots Manual Everything. There
is no right or wrong - there are just instances of
individuality and individual choice.'

To me something is wrong. The flowers on the trees left look shaky, kind of bad bokeh, but that is not the case on the right. In general, the left is looking bad yes.

You might try another lens and focus on the trees on the right closer by. If result is the same, sensor is out of alignment and you need to send the camera in.
--
------------------------------------------------
http://s259.photobucket.com/albums/hh315/alex_837/
Just trying to get better.....

No really breaking news.

But at least one thing seems to be one factor (maybe not that much in my above sample): blure by shake - and shake reduction. I was wondering about some slight blure in many (almost all?) pictures across the whole frame. It seems like under good conditions it's better so switch off shake reduction - even when shot hand held!. In some examples (with f24mm, 1/320s, f1/7,1 ISO100) the result was a nice tick better.

So while trying and practicing I think I'll learn more about what cam, lens and me can do about focusing. And hopefully I'll get some nice shots by the way and anyway. When I find time and a wall (or something similar sufficiant) I may do tests with that.

After what you said, it's clear your camera isn't working right. You have to send it back.
------------------------------------------------
http://s259.photobucket.com/albums/hh315/alex_837/
Just trying to get better.....

First of all I'd like to thank anybody for taking your time and for avdvice so far. And of course I don't want to be boring to anybody, so I'm sure only the interrested will hold on. ;-)

Somehow it just looks like debruyne ist right and something's wrong with the camera. At least it doesn't come up to my expectations regarding focus and especially auto focus. And as arachnophilia wrote, something could be misaligned.

I still didn't find an ideal target but at least I found a wall for some test - with house number (No 14 - but streetname is too small ...). Always tried to focus on No. 14 (or slightly right there on the wall). Of course that's far away from an ideal test but it's a quite realistic 'real life' test.

My findings: Auto focus with shake reduction in P-mode gave rather blured result (IMGP9930). Auto focus without shake reduction and shutter priority could do a little beter (IMGP9941). Best result I could get was by focusing manually in live view (with shake reduction off and shutter priority - IMGP9943). I'd expect to get at least somthing like that with auto focus (and maybe with SR on). But still I'm not sure that the 'plane of focus' in the scene is where I'd expect it to be.

Here the link to my (hopefully soon obsolete and then deleted ...) album:
http://forums.dpreview.com/galleries/3490625903/albums/pain-of-focus-ii

For comparison there's a simple-as-that point and click shot with f50fd (DSCF9643).

So I guess I have to ask my dealer soon for some help .

These pics aren't suited for a test. You need a wall with detail and a tripod. Different distances, is on and off.
--
------------------------------------------------
http://s259.photobucket.com/albums/hh315/alex_837/
Just trying to get better.....

Sorry, this is not what we mean.

• A brick wall or newspaper on a brick wall should be say 6 ft from the camera. Or less. The brick wall should be unpainted and show raw bricks. Or if you stick newspaper on the wall, the newspaper must be flat.

• There should be nothing between the brick wall and you. No people walking across the wall etc... This will fool the auto focus system.

See what brick walls look like:

http://www.google.com.au/images?q=brick+wall

• You need to put the camera on a tripod, and make sure the camera lens axis is at right angles to the brick wall.

roland66 said:

I still didn't find an ideal target but at least I found a wall for some test - with house number (No 14 - but streetname is too small ...). Always tried to focus on No. 14 (or slightly right there on the wall). Of course that's far away from an ideal test but it's a quite realistic 'real life' test.

Here the link to my (hopefully soon obsolete and then deleted ...) album:
http://forums.dpreview.com/galleries/3490625903/albums/pain-of-focus-ii

For comparison there's a simple-as-that point and click shot with f50fd (DSCF9643).

Click to expand...

--



Ananda
http://anandasim.blogspot.com

'There are a whole range of greys and colours - from
the photographer who shoots everything in iA / green
AUTO to the one who shoots Manual Everything. There
is no right or wrong - there are just instances of
individuality and individual choice.'

debruyne said:

These pics aren't suited for a test. You need a wall with detail and a tripod. Different distances, is on and off.

Click to expand...

Yes, I see. If I really want to proof camera and/or lens to be bad (or not too bad) my testshots are by no means sufficiant. There's to be more and systematic testing. And of course there should be some constant light. Changing sun and clouds don't make test conditions better.

But I've just been there at that place, found the motive (wall with house number and street name) and had the two cameras. So I tried what I could get. And from a practical point of view the conditions for the two cameras were not really unfair.

f50fd has it's limitations (noise, colors, lens ...), so I thought and still believe ;-) there's cameras that can do better. (Anyway, a tripod would help f50fd too - even while it has no maual focus.) I try to learn what I could expect from k-x. From f50fd I already know. About k-x I'm still not sure. But generally I'd expect from k-x (maybe with better lens) better results than with f50fd.

AnandaSim said:

Sorry, this is not what we mean.

• A brick wall or newspaper on a brick wall should be say 6 ft from the camera. Or less. The brick wall should be unpainted and show raw bricks. Or if you stick newspaper on the wall, the newspaper must be flat.

Click to expand...

Ok. Structure on the wall is good. But different distances should also be tested (as debruyne suggested)(even more than 6 ft to infinity I think).

AnandaSim said:

• There should be nothing between the brick wall and you. No people walking across the wall etc... This will fool the auto focus system.

Click to expand...

Yes. I thought I focused with clear view but maybe I missed that.

AnandaSim said:

See what brick walls look like:

http://www.google.com.au/images?q=brick+wall

Click to expand...

Ok, ok ;-)

AnandaSim said:

• You need to put the camera on a tripod, and make sure the camera lens axis is at right angles to the brick wall.

Click to expand...

About the angles I'm not sure in my testshots. Of course I tried to have lense axis vertical to the wall. But I guess that's something one gets fooled too esay. In the end one really has to meassure distances and angles exactly to be sure.

... so I'm waiting for the great wall ...

roland66 said:

Ok. Structure on the wall is good. But different distances should also be tested (as debruyne suggested)(even more than 6 ft to infinity I think).

Click to expand...

No. Put on socks first, then put on shoe.

There is nothing to test AF at infinity. Because there is not standard test object that is flat that is miles wide and miles tall. There are only odd shaped buildings, trees with moving leaves, people moving - that's not a test of AF - because it is a moving target.

roland66 said:

About the angles I'm not sure in my testshots. Of course I tried to have lense axis vertical to the wall. But I guess that's something one gets fooled too esay. In the end one really has to meassure distances and angles exactly to be sure.

Click to expand...

Good! That's why you need a brick wall at 6 ft, not at infinity. Oh and make sure brick wall is not curved.





--



Ananda
http://anandasim.blogspot.com

'There are a whole range of greys and colours - from
the photographer who shoots everything in iA / green
AUTO to the one who shoots Manual Everything. There
is no right or wrong - there are just instances of
individuality and individual choice.'

AnandaSim said:

...

Good! That's why you need a brick wall at 6 ft, not at infinity. Oh and make sure brick wall is not curved.

Click to expand...

So I see: In the end it's really a kind of engineering and has to be done with enough care. And then it even might provide some hint what may be wrong because I can see where the zones of focus really are.

My first approach was just to get a rough guess if something might be wrong and then (at local store where I bought, maybe) have some experts done a real test (if they could). But in deed it's probably worth enough to do some careful test on my own anyway to see what happens and to get a feeling for that.

... waiting for (a not too great at all) wall (not too curved) ...

Of course this is no "valid test" as well but at least it shows something (picture below):

In some lucky circumstances I can have sharp results with my k-x and the lens.
BUT it seems (very) difficult and less often than to be expected.

From pictures I took so far it just looks like my k-x suffers from the ('famous') SR problem (image blure at shutter speeds rough about 1/60s to 1/160s or something like that). At least I had some good results with shutter speed 1/250s and faster. But even with 1/30s (and SR on) I can have some good results. Maybe additionally there's sometimes trouble with autofocus not focusing to what I expect to ... and of course maybe my technique holding the cam also could be improved ...

Here's my picture of the day. It's not a great picture at all but at least it seems to be as sharp as the lens can do (at shutter speed 1/640s (!), FL23.13mm, aperture 6.3). (And we see a little moiré again ...)

roland66 said:

Of course this is no "valid test" as well but at least it shows something (picture below):

Click to expand...

Yes, it is not a valid test or a valid demonstration - what do you want to demonstrate?

roland66 said:

In some lucky circumstances I can have sharp results with my k-x and the lens.
BUT it seems (very) difficult and less often than to be expected.

Click to expand...

That is because you demonstrations are not valid and they do not test parameters singly. This photo demonstrates too many things.

roland66 said:

From pictures I took so far it just looks like my k-x suffers from the ('famous') SR problem (image blure at shutter speeds rough about 1/60s to 1/160s or something like that).

Click to expand...

Then you should first put the camera on a tripod, no SR, shoot the unpainted brick wall. Reference photo A

Then hand hold the camera with SR on and shoot same wall. Sample photo B

Note: This is not a test of focus
Note: This is not a test of lens sharpness
Note: This is test of SR

roland66 said:

Maybe additionally there's sometimes trouble with autofocus not focusing to what I expect to ... and of course maybe my technique holding the cam also could be improved ...

Click to expand...

Then you are testing autofocus. To test autofocus, you need to can perform two tests - static object test (brick wall) vs moving object test. Moving object test is difficult

Then you should first put the camera on a tripod, no SR, using manual focus with Liveview and Magnification. With fast shutter speed Reference photo C

Then camera on tripod the camera on a tripod, no SR, shoot with autofocus. Reference photo D with fast shutter speed.

The test subject and method is a long ruler - there is a website/webpage DIY method :

http://focustestchart.com/chart.html

roland66 said:

Here's my picture of the day. It's not a great picture at all but at least it seems to be as sharp as the lens can do (at shutter speed 1/640s (!),

Click to expand...

why so surprise at 1/640?

roland66 said:

FL23.13mm, aperture 6.3). (And we see a little moiré again ...)

Click to expand...

I don't see moire, I see fringing on the white building - because this is an overbright white building and some lens do show colour fringing. The white paint on the building looks over exposed and the details of the walls of the white and cream buildings are too far and too small for the lens to resolve.

The leaves are larger than wall imperfections / cracks and they resolve fine because they are bigger objects.

This photo is NOT an example of

• Autofocus - because I think it is at infinity and if your camera/lens cannot AF to infinity you have big problems

• SR - because you have 1/640th sec

This is an example of

• very high dynamic range causing the picture to be not to pleasant

• shooting into bright object causing fringing ( lens or lens and sensor interaction)

• too far distance and too small object for the lens to resolve cracks/texture of painted walls.

--



Ananda
http://anandasim.blogspot.com

'There are a whole range of greys and colours - from
the photographer who shoots everything in iA / green
AUTO to the one who shoots Manual Everything. There
is no right or wrong - there are just instances of
individuality and individual choice.'