What do you think of this softness issue

Some of you may remember posts where I have had troubles with finding a lens which is sharp enough for ‘everything in focus’ scenic shots, particularly edges. My Tamron 17-50mm has been away several times for testing. Also the Shop I bought my K30 from gave me a second body as it seemed to perform a small amount better with the Tamron than the first body.

I’ve now discovered that the problem may indeed have been the body all along. Because the battery on my K30 was dead I used my old K100Ds and found all my lenses perform much better on this camera. Here is an example shot with some 100% crops. They are default (factory settings) jpegs and RAW images show the same issues. The shake reduction is off. I know the K30 is 16mp compared to the 6mp of the K100Ds but I found reducing the K30 image size to that of the K100Ds makes no difference. I am going to contact Pentax and see what they say. If any of you have a K30 and an older model maybe you could try this for yourself. 28mm M prime used.




Reduced photo




K30 edge


K100Ds edge




K30 roof middle


K100Ds roof middle




K30 centre


K100Ds centre

I forgot to mention the test was at f8. Another issue is why can I see CA’s on the K30 and very little on the K100Ds using the same lens. It makes you wonder how valid lens tests are, as this 28mm lens would get a different review depending which body it is used on.

Ian Leach said:

Ian Leach wrote:

I forgot to mention the test was at f8. Another issue is why can I see CA’s on the K30 and very little on the K100Ds using the same lens. It makes you wonder how valid lens tests are, as this 28mm lens would get a different review depending which body it is used on.

Click to expand...

Your K-30 is more than twice the resolution as the K100ds, so the softness is likely due to that, add a little sharpening. This will also amplify any CA.

Secondly, focus could be a little different causing a change in CA. Also, could be a difference in processing. My K-5 was always a bit magenta in the shadows on Lightroom, so I had to adjust that and save it as default for the camera. The newer Lightroom profiles might be better.

Finally, my K-5 always had noticeably more purple fringing than my K-7, not to the point of being a problem, but definitely noticeable side by side with the K-7. However, I don't think this is the issue here because the older Sony CCD sensors had even worse purple fringing than the K-5.

edit: I see you tried resizing, but if the focus is a bit different, or I have seen where the higher resolution sensors just don't perform as well in the corners (but the 16 MP seems to be one of the better ones in the corners). If the lens has a bit of field curvature it might be a bit worse in the corners, try focusing out 2/3 of the way towards the corner too. Live view focus on the K-30 of course to get best results.

Eric

--

I never saw an ugly thing in my life: for let the form of an object
be what it may - light, shade, and perspective will always make it
beautiful. - John Constable (quote)
See my Blog at: http://www.erphotoreview.com/ (bi-weekly)
Flickr Photostream: http://www.flickr.com/photos/28177041@N03/ (updated daily)

Hi,

I have neither the K30 nor the Tamron lens, so of course I'm going to try to help.

1. The test photo shows the tree in the middle of the frame. If you use the center point for focusing, won't the house be a bit out of focus? I normally focus/recompose, so you may have done this as well. Even so, AF tests should be done on a tripod, and focus/recomposing will adjust the focal plane slightly - but at the distance of your test it may not matter so much.


2. I used to have a Tamron 28-75 2.8 for Canon (very similar optical formula to your lens) which is why I'm commenting. The exif doesn't show it, but I think you took the photo at 17mm, is this correct? The Tamron 28-75 and the 17-50 are notorious (at least on the Canon forum) for bad focusing at the wide end, especially for subjects that are far away -if your lens focuses much better close up, then this is likely your problem. The 28-75 is parfocal, so your lens should be as well. Try zooming in to your subject, lock focus, then zoom out and take the photo. If the photos are sharper, then the problem is the poor AF with the lens at wide angle, not the camera.

Anyway, just some thoughts. Hope it helps.

Your K100D shots are tagged as being taken in May 2007 but the car number plate is an 09 so that is not right. There is a significant difference in exposure between the two cameras which would make me question whether the same aperture was used on both. One says 1/180sec and the other 1/1000 sec which is a difference of about two and a half stops. Are you sure your K30 was set to f8 and stopping down OK?

Thanks for the comments. The reason I bought higher mega pixel cameras is to print larger, so if the size of the sensor is the issue what's the point of upgrading. The level of softness is to much to correct with sharpening. Also my samsung NX10 is 14 mega pixels and has no such issues with the 30mm lens.


Thanks again Ian

Thanks for the comments but I took these photos with a 28mm manual lens, the Tamron is still in germany being fixed for a different issue. At f8 everything should be in focus anyway, I have tried changing the focus point but at these distances it shouldn't be relevant.

Thanks again Ian

Yes definitely try again, also take a good look at the lens to see if the aperture blades are working as they should... look at the blades while off the camera and during 1+ second exposures while on both cameras to see if there is a flaw. I look forward to seeing the even more controlled test. BTW you can set the focus ring to a point, remember that exact point when you swap the lens, half a mm either way really shouldn't make a difference at that focal length and aperture.

Others are right that higher resolution sensors show greater aberrations and some sensors are more prone to certain types of aberrations.

Ian Leach said:

Ian Leach wrote:

I forgot to mention the test was at f8. Another issue is why can I see CA’s on the K30 and very little on the K100Ds using the same lens. It makes you wonder how valid lens tests are, as this 28mm lens would get a different review depending which body it is used on.

Click to expand...

I recommend you test the K-30 again using LV mode. That should eliminate any chance of back or front focus as a possible source of the image softness.

Ian Leach said:

Ian Leach wrote:

I know the K30 is 16mp compared to the 6mp of the K100Ds but I found reducing the K30 image size to that of the K100Ds makes no difference.

Click to expand...

How did you do that? As you've only posted 100% crops from the full size files it's pretty well impossible to get a true impression but in at least some of them things look as bad from the K100 as from the K-30 - especially the CAs on the window, which look stronger on the K100 version. I'm not saying they are as bad - just that some of them look that way in what you've posted.

I don't know what raw converter you use: ACR gives a range of choices for MP count, one of which should be 6.3MP for your K-30, which is so close to the 6.1MP of your K100 that comparison should be easier. It's possible to get an exact match in some editing software.

It's obvious from the crops that your viewpoint was different between the shots; not by much, but it's important that they are the same. Use a tripod if you have one; if not, put a bean bag on a chair or box and sit each camera on it. If possible use a spirit level to make sure the corners are all the same.

Hi,

I have the tamron and it's a very sharp lens even to the edges at 17mm @ F4 even on the demanding 16Mp of the K01 and K5,the extreme corners are however a little softer,but one thing to remember is at that focal length there is a strong field of curvature and it can be misleading to assume that just because the center is sharp that focusing is correct;as the edges are behind the center plane of focus. Try AF compensation to attain sharp focus across the frame at infinity (landscape view). It is normal to have moderate CAs,easily correctable in RAW. If this isn't successful it could be a lens issue of which unfortunately there are many complainers.

I Think your K30 suffers from backfocus.

If you have the K30 capture mode set to 'M' and are using ISO AUTO then it actually uses TAv mode (see page 67 in manual). As the K30 cannot identify the aperture for a manual lens, it shoots wide open. The K100D will not let you use ISO AUTO if the capture mode is 'M'

There is an easy solution - never use ISO AUTO when using a manual lens.

Andrew