Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
10-21-2008, 05:48 AM   #16
Pentaxian
audiobomber's Avatar

Join Date: May 2008
Location: Sudbury, Ontario
Photos: Albums
Posts: 6,631
Original Poster
QuoteOriginally posted by Class A Quote
perhaps the K100D has a BF focusing problem that also extends to manual focusing
If you can tell me the mechanism by which my K100DS arrives at the exact same misfocus in MF or AF, I'll look into it. I confirmed that MF and AF were the same during the test, and again later. I've taken several thousand photos with the K100DS, more than a few at F1.4. I think I would know if it has a focussing problem.

QuoteOriginally posted by Class A Quote
JPEG conversion (although the parameters are all at default values, I'm sure the RAW->JPEG conversion is vastly different between the two cameras), and Picasa's behaviour when downscaling from two different input resolutions.
I think you're clutching at straws. The posted versions were downsized by Picasa, but I have the full size photos, and they tell the same story about relative resolution.

QuoteOriginally posted by Class A Quote
I cannot believe either that the 50mm on the K100D cannot produce a sharper picture. My 18-250 zoom produces sharper pictures than this one on a K100D.
Remember that sharpness wasn't boosted in the camera or in post-processing. Does anyone not sharpen? Remember also that this is extreme pixel peeping. These kinds of differences show up in the small print on the box. I tried to do the test with artificial flowers, but it would have shown absolutely nothing. The flowers always looked great. You can't see resolution differences in the Cheerios in the bowl either. It's only the fine print that shows a difference. Pretty irrelevant to most photography. It would become relevant if you were archiving printed pages, or printing very large prints, or for severe cropping. Those are the areas where resolution matters. As far as taking beautiful photos, both cameras are more than up to the task.

According to my tests, a K100DS has about the same resolution with an 18-55mm as a 50mm prime, because the limits of the camera are reached. I also did some testing with an 18-250mm. It has slightly better resolution than the kit lens, less than the 16-45 and less again than the 50mm (tested on the K20 of course). You have an 18-250mm. I assume you have a K100. Do you have a 50mm prime? Why not do a test? Get out your tripod and some fine print, and let's see how the prime outresolves the zoom on your K100.

PS Here are some photos taken with the 18-250 on my K100DS. Let me know if you spot a problem with my camera. http://picasaweb.google.com/bonhommed/Pentax18250mm#slideshow/5223089912796759026

Here are some 16-45mm photos taken with the K100DS and K20D. Let me know if you can spot which were taken with which camera. http://picasaweb.google.com/bonhommed/1645mm#slideshow


Last edited by audiobomber; 10-21-2008 at 06:32 AM.
10-21-2008, 09:10 PM   #17
Pentaxian
Class A's Avatar

Join Date: Aug 2008
Location: Wellington, New Zealand
Posts: 8,934
QuoteOriginally posted by audiobomber Quote
If you can tell me the mechanism by which my K100DS arrives at the exact same misfocus in MF or AF, I'll look into it.
On possibility is that the sensor is misaligned. Then AF and MF would give the same result but the focus on the sensor would be slightly off.

It is also possible that the sensor is perfectly aligned but both the focusing screen and the AF alignment are a bit off.

It is also possible that all is well and you are reporting on a factual inability of the K100D to exploit prime sharpness. I don't exclude that possibility, I just find it hard to believe.

QuoteOriginally posted by audiobomber Quote
I've taken several thousand photos with the K100DS, more than a few at F1.4. I think I would know if it has a focussing problem.
Why should you if you could attribute any imperfections to the limited sensor resolution?

QuoteOriginally posted by audiobomber Quote
I think you're clutching at straws.
Why should I?
My K100D will not take better/worse pictures whether I'm wrong or right. I'm happy with them. I just found your result a bit implausible (and I'm not the only one).

I merely pointed out some variables which you might no t have had under control in your experiment. Comparing RAW files, i.e., using the same external JPEG processor, would eliminate different in-camera JPEG processors, for example.

However, you could still be right. I cannot be sure that any of the reasons I mentioned made a difference. I'd be inclined to think so because I believe the K100D will show the difference between a kit lens and a prime.

QuoteOriginally posted by audiobomber Quote
The posted versions were downsized by Picasa
Yes, and depending on whether you start from a K20D resolution or a K100D resolution, things could look different. Judging sharpness with downscaled/upscaled images is always problematic.

QuoteOriginally posted by audiobomber Quote
Remember that sharpness wasn't boosted in the camera or in post-processing.
You didn't boost it but every camera has its own level of default sharpening. This level could be higher for the K20D than for the K100D.

QuoteOriginally posted by audiobomber Quote
According to my tests, a K100DS has about the same resolution with an 18-55mm as a 50mm prime, because the limits of the camera are reached.
This is not in accordance to an article on sensor resolution I've read. Only a few lenses are diffraction limited. The others exhibit aberrations which are higher than the sensor resolution. IIRC, The article says that for an APS-C sensor most lenses don't need more than 7MP.

QuoteOriginally posted by audiobomber Quote
You have an 18-250mm. I assume you have a K100.
Yes.
QuoteOriginally posted by audiobomber Quote
Do you have a 50mm prime?
Not yet. I'm trying to get a 50mm f1.4. If anyone got one for sale, I'd be interested. I cannot get it in New Zealand and the German prices (could get it from there) are not as good as the US ones.

QuoteOriginally posted by audiobomber Quote
PS Here are some photos taken with the 18-250 on my K100DS. Let me know if you spot a problem with my camera.
Thanks, I'll have a look soon.
10-21-2008, 10:19 PM   #18
Pentaxian
Class A's Avatar

Join Date: Aug 2008
Location: Wellington, New Zealand
Posts: 8,934
QuoteOriginally posted by audiobomber Quote
Here are some 16-45mm photos taken with the K100DS and K20D. Let me know if you can spot which were taken with which camera. Picasa Web Albums - Dan - 16-45mm
I had a quick look and of course this isn't easy as a slightly misfocused K20D image could appear to be a K100D shot.

Also, the perceived sharpness differed when looking at the full resolution(?) / slideshow versions versus the versions where I could look at the image #. This just confirms what I said about downscaled/upscaled images.

To me images #4, 5, 8, 9, 11, 14 appear to be sharper than the others. I don't expect, though, that all these are from the K20D whereas all the others are from the K100D because of the difficulty to judge sharpness from downscaled images and/or the potential for soft K20d shots.
10-21-2008, 10:28 PM   #19
Pentaxian
Class A's Avatar

Join Date: Aug 2008
Location: Wellington, New Zealand
Posts: 8,934
Just noticed this again:
QuoteOriginally posted by audiobomber Quote
- All images were cropped in Windows (2000mp wide for K20D, 1300 wide for K100) and then resized to 1200 mp wide using Picasa.
Resizing from 1300 to 1200 (in case of the K100D) is practically bound to result in a slightly fuzzy image. This fractional scaling factor of 0.923 means that effectively blurring is introduced. This will make it more difficult to spot any difference between the two K100D images, if it actually exists.

10-23-2008, 06:49 PM   #20
Pentaxian
audiobomber's Avatar

Join Date: May 2008
Location: Sudbury, Ontario
Photos: Albums
Posts: 6,631
Original Poster
QuoteOriginally posted by Class A Quote
Just noticed this again:

Resizing from 1300 to 1200 (in case of the K100D) is practically bound to result in a slightly fuzzy image. This fractional scaling factor of 0.923 means that effectively blurring is introduced. This will make it more difficult to spot any difference between the two K100D images, if it actually exists.
I'm looking at non-downsized images, and I can see that the downsized images are an exact representation, only smaller.
10-23-2008, 07:13 PM   #21
Pentaxian
Class A's Avatar

Join Date: Aug 2008
Location: Wellington, New Zealand
Posts: 8,934
QuoteOriginally posted by audiobomber Quote
I'm looking at non-downsized images, and I can see that the downsized images are an exact representation, only smaller.
Exact representation?
Do you not recognise a difference in sharpness when comparing them side by side on the screen? That completely baffles me. That only makes sense when the original image is not sharp to begin with (which it should when taken with a prime in focus).

Have you tried using an LCD monitor with another than its native resolution? Do you notice how much softer the image becomes because of the rescaling? I couldn't stand looking at a rescaled image like that. I couldn't buy an HDTV set which didn't have an option to turn off "overscan". Many people don't bother not getting a 1:1 pixel mapping. To me the difference is huge.

I'll try to obtain a prime and will compare it to my super zoom. I'll post the images here, but it will probably take a while before I can do it.

BTW, do you plan to reveal the solutions to the "16-45mm photos taken with the K100DS and K20D" puzzle?
10-23-2008, 07:47 PM   #22
Pentaxian
audiobomber's Avatar

Join Date: May 2008
Location: Sudbury, Ontario
Photos: Albums
Posts: 6,631
Original Poster
Geez, responding to this post this is like a homework assignment. That's why I've been avoiding it.

QuoteOriginally posted by Class A Quote
One possibility is that the sensor is misaligned. Then AF and MF would give the same result but the focus on the sensor would be slightly off.

It is also possible that the sensor is perfectly aligned but both the focusing screen and the AF alignment are a bit off.

It is also possible that all is well and you are reporting on a factual inability of the K100D to exploit prime sharpness. I don't exclude that possibility, I just find it hard to believe.
A misaligned sensor I expect would give a pretty distorted image from one side of the frame to the other. I still think if my camera's focus was off, I'd have noticed. But if you know of some way to test for the above, I'll give it a go.

QuoteOriginally posted by Class A Quote
My K100D will not take better/worse pictures whether I'm wrong or right. I'm happy with them. I just found your result a bit implausible (and I'm not the only one).
I love the images from my K100DS. It takes beautiful photos. The kind of resolution I was trying to measure here is only significant for limited uses. My main use is for cropping wildlife photos. If you don't think there's a major difference between the K100 and K20 for heavy crops, well you are deluding yourself.

It's significant to me that the two people in this thread who've moved from a 6mp camera to the k20D agree with my results.

QuoteOriginally posted by Class A Quote
I merely pointed out some variables which you might not have had under control in your experiment. Comparing RAW files, i.e., using the same external JPEG processor, would eliminate different in-camera JPEG processors, for example.
Point taken. I agree that there were uncontrolled variables in my test, but I don't think they were significant. Pentax Photo Lab converts raw files pretty much exactly the way the in-camera processing does from anything I've read or seen. To display photos, you have to do some sort of processing. You can't look at raw data. There's an extra step in Picasa so that I can show you what I saw, but as I said, the end result was not changed by downsizing.

QuoteOriginally posted by Class A Quote
However, you could still be right. I cannot be sure that any of the reasons I mentioned made a difference. I'd be inclined to think so because I believe the K100D will show the difference between a kit lens and a prime.
I have a couple of raw images of a similar scene, the K100DS with kit lens and 50mm. They do seem to show more difference than the photos I posted above. Tell me how you want me to process them and I'll do it and post them here.

I still firmly believe that a top notch prime lens on a K100 will not deliver the kind of resolution you get from any decent lens on a K20. It may well deliver a nicer looking photo due to contrast, colour rendition, bokeh etc, but not the same resolution.

QuoteOriginally posted by Class A Quote
Yes, and depending on whether you start from a K20D resolution or a K100D resolution, things could look different. Judging sharpness with downscaled/upscaled images is always problematic.
Maybe, but I haven't seen it as long as both are processed in the same way.

QuoteOriginally posted by Class A Quote
You didn't boost it but every camera has its own level of default sharpening. This level could be higher for the K20D than for the K100D.
Yes, it definitely could be. I decided to use the camera defaults only because it avoided me having to make decisions about saturation, contrast and sharpness. I showed what Pentax thought was optimum. At least it took my manipulations out of the equation. If I had shot raw and processed myself, someone would have objected to my processing.

QuoteOriginally posted by Class A Quote
This is not in accordance to an article on sensor resolution I've read. Only a few lenses are diffraction limited. The others exhibit aberrations which are higher than the sensor resolution. IIRC, The article says that for an APS-C sensor most lenses don't need more than 7MP.
I think you're misreading or misrembering the conclusion of this article. At F5.6 you could need as little as 11mp or as much as 51mp to outresolve a lens. And I assume they mean an ideal lens.
10-23-2008, 08:01 PM   #23
Pentaxian
audiobomber's Avatar

Join Date: May 2008
Location: Sudbury, Ontario
Photos: Albums
Posts: 6,631
Original Poster
QuoteOriginally posted by Class A Quote
To me images #4, 5, 8, 9, 11, 14 appear to be sharper than the others. I don't expect, though, that all these are from the K20D whereas all the others are from the K100D because of the difficulty to judge sharpness from downscaled images and/or the potential for soft K20d shots.
Photos 2, 12, 13, 14 were taken with the K20D. I can't remember which camera was used on 15 and the data is stripped off. The reason you can't tell the cameras apart is that this type of photo won't show up differences in resolution. That's why for the test I cropped in to show the small print on a cereal box. I wanted to highlight the differences. But it's not to do with the size of the image, it's the content. I can show this difference in resolution using a much smaller image than any we're talking about.

10-23-2008, 10:06 PM   #24
Pentaxian
Class A's Avatar

Join Date: Aug 2008
Location: Wellington, New Zealand
Posts: 8,934
QuoteOriginally posted by audiobomber Quote
A misaligned sensor I expect would give a pretty distorted image from one side of the frame to the other.
Why?
If it is perfectly aligned in terms of angles, but just not has the right distance to the lens then the image won't be distorted, just fuzzy. This is what I meant with "misaligned".

QuoteOriginally posted by audiobomber Quote
But if you know of some way to test for the above, I'll give it a go.
You could make a series of test chart images, each image slightly differing in focus setting. One of them should be sharp and you could check whether you gain the same level of sharpness with AF/viewfinder MF.

QuoteOriginally posted by audiobomber Quote
The kind of resolution I was trying to measure here is only significant for limited uses.
We are in agreement here. However, I still don't believe that a 6MP sensor cannot show the difference between the kit lens and a good prime.

QuoteOriginally posted by audiobomber Quote
If you don't think there's a major difference between the K100 and K20 for heavy crops, well you are deluding yourself.
I have never said this. Please don't accuse me of "deluding myself" for words you are putting in my mouth.

QuoteOriginally posted by audiobomber Quote
It's significant to me that the two people in this thread who've moved from a 6mp camera to the k20D agree with my results.
I could say that someone else in this thread, who is really experienced regarding photography, shares my scepticism regarding your inability to see a difference between the kit lens and the prime with the K100D. But in terms of the truth, it doesn't matter what others think.

QuoteOriginally posted by audiobomber Quote
Pentax Photo Lab converts raw files pretty much exactly the way the in-camera processing does from anything I've read or seen.
It will do that if you allow it to take the camera settings (the default). If you use the same settings for both K20D and K100D RAW files, you will eliminate one variable. You are free to speculate about the significance of this.

QuoteOriginally posted by audiobomber Quote
You can't look at raw data.
For the record, you cannot look at JPEG data either.
However, at least one graphics card / monitor combination will render two different JPEGs the same way. In contrast, two different RAW files will not be rendered the same way, if different sharpening, saturation, etc. levels are used.

QuoteOriginally posted by audiobomber Quote
but as I said, the end result was not changed by downsizing.
But you describe a rescaling operation in your original post?

QuoteOriginally posted by audiobomber Quote
Tell me how you want me to process them and I'll do it and post them here.
If you use one RAW processor (like Photo Laboratory) and use exactly the same manual settings for WB (not "in camera" setting), sharpness, saturation, etc. for both RAW files then you will have eliminated any RAW->JPEG conversion differences.

QuoteOriginally posted by audiobomber Quote
I still firmly believe that a top notch prime lens on a K100 will not deliver the kind of resolution you get from any decent lens on a K20.
I have never said that and I believe a really top notch prime will show the difference between these cameras. What I objected to was the conclusion that the K100D is unable to resolve the difference between the kit lens and a good prime.

QuoteOriginally posted by audiobomber Quote
I showed what Pentax thought was optimum.
I beg to differ. You showed what Pentax thought would work best for a Japanese consumer market taste. Most Europeans with a taste find the default "bright" setting to be over the top. For most holiday picture shooters very healthy skin tons and deep blue skies work well and that's why Pentax put the default values this way. They called the alternative setting "natural" for a reason.

QuoteOriginally posted by audiobomber Quote
At least it took my manipulations out of the equation. If I had shot raw and processed myself, someone would have objected to my processing.
Sorry, that completely goes over my head. How can someone complain if you conduct a fair test? Unless, of course, you want to compare "out-of-the-camera-JPEG-no-changes-made" picture quality, of course. However, I thought the latter was not your objective.

QuoteOriginally posted by audiobomber Quote
At F5.6 you could need as little as 11mp or as much as 51mp to outresolve a lens. And I assume they mean an ideal lens.
Exactly. Go find a lens that behaves ideally (i.e., no aberrations at f/5.6).

The 7MP I quoted are for f/11 for which most lenses become diffraction limited. Admittedly, most lenses will peak earlier than that (e.g., f/8) but you'd still only need 13MP to capture this resolution. Below that most lenses won't be sharper, although they could be in theory.

Just to get this straight: Obviously the K20D's resolution gives you more headroom for cropping. If you take two images with the K20D and the K100D from the same position and crop out the same subimage, and rescale both to the same size then obviously the K20 image should be sharper as the end result was created from more than twice as many pixels.

The fact that your K100D kit lens and prime show precious little difference is a bit funny though. I exclude the possibility that this means that the kit lens outresolves a 6MP sensor at f/1.4. Hence, if there is something funny with your K100D shots, the difference to the K20D shots may not be that big if the source of the problem is identified and fixed.

Perhaps there isn't a problem and everything is fine. I'll do my own experiment with a prime a.s.a.p. and will let you know. Maybe I'll come up with the same result.

Can you send me the "Cheerios" package you have used? We should avoid the variable of using two different packages.
10-23-2008, 10:09 PM   #25
Pentaxian
Class A's Avatar

Join Date: Aug 2008
Location: Wellington, New Zealand
Posts: 8,934
QuoteOriginally posted by audiobomber Quote
The reason you can't tell the cameras apart is that this type of photo won't show up differences in resolution.
Yes, and of course there is much more variance w.r.t. sharpness in these largely varying conditions.
10-24-2008, 05:41 AM   #26
Pentaxian
audiobomber's Avatar

Join Date: May 2008
Location: Sudbury, Ontario
Photos: Albums
Posts: 6,631
Original Poster
When I have some time I'll do some tests of my K100DS' focal accuracy. That will be more productive than continuing the point-by-point debate.
10-24-2008, 10:33 AM   #27
Pentaxian
Marc Sabatella's Avatar

Join Date: Sep 2006
Location: Denver, CO
Photos: Gallery
Posts: 10,686
QuoteOriginally posted by Class A Quote
We are in agreement here. However, I still don't believe that a 6MP sensor cannot show the difference between the kit lens and a good prime.
Indeed, I know for a fact this isn't true in general. That is, 6MP is more than enough resolution to show the difference. I don't have any insight into what might have been going on here to obscure this fact, but on my DS, the difference between the 18-55 and the A50/1.7 was *gigantic* at f/5.6, and still quite noticeable at f/8. Looking at the images in the original post here, the 18-55 on the K100D looks about as I'd expect. The 50 doesn't look nearly as good as it should. I'll let others continue speculate as to why.
10-29-2008, 03:13 AM   #28
Pentaxian
Class A's Avatar

Join Date: Aug 2008
Location: Wellington, New Zealand
Posts: 8,934
Thanks for your input Marc. I'm looking forward to owning a FA 50mm f/1.4 and will compare it to my Tamron 18-250mm @ 50mm. I'll post again as soon as I have results.
01-11-2009, 04:12 AM   #29
Pentaxian
Class A's Avatar

Join Date: Aug 2008
Location: Wellington, New Zealand
Posts: 8,934
K100D sees difference between prime and superzoom

I promised to do a comparison myself and here are my results.

I also shot a cereal box using a tripod, 2sec dealy, ISO 200, f/5.6, and AF. I'm not claiming I used the optimal AF calibration for both lenses but did three shots for each at -60, 0, and 60 settings. For the comparison I picked the shots that came out sharpest respectively.

Below you first see a shot with the Tamron 18-250. The following is by the Pentax FA 50/1.4. Note that the Pentax peaks at f/5.6 in terms of sharpness and the Tamron 18-250 is just stopped down a bit at this focal length.

I had to set the Tamron 18-250 at 55mm to get the same field of view. I should have set it even slightly higher to get exactly the same field of view. Setting it at 50mm seems like the sensible thing to do but the effective focal length of the Tamron reduces the closer the subject. To eliminate the remaining difference in field of view I cropped the same area from both shots. These are ~50% crops. I then resized to 800 pixels width for web display. To avoid any difference in exposure (test done with daylight), I applied Picasa's AutoContrast to both shots.

I think one can clearly see that in this test run the 6MP are sufficient to show the superior sharpness of the prime.

The third shot is the first shot again, with one level of Picasa sharpening applied. I've included this to show that with additional sharpening the superzoom holds its own well against one of the sharpest primes.

I'm not saying my quick test here proves anything. I cannot guarantee for optimal AF results (but it appears that I hit almost an optimum for the zoom and the prime may still have some more potential). I think a better test would avoid a flat subject but allow sharpness to be seen and compared along a tilted area. So this is not meant to be conclusive but at least it somewhat contradicts the conclusions of the OP and may be more in line what one would expect.

Last edited by Class A; 08-07-2010 at 09:21 AM.
01-11-2009, 04:38 AM   #30
Pentaxian
Class A's Avatar

Join Date: Aug 2008
Location: Wellington, New Zealand
Posts: 8,934
What my test definitely shows, I believe, is that a K100D with a FA 50/1.4 can achieve a sharper picture than the one the OP presented and used in his comparison.

Below is a 1300 wide crop that has been rescaled to 1200 width. I did this to repeat the procedure by which the OP's image was obtained.

Again, I'm not saying I've exploited the optimal sharpness of the prime, since I didn't try to find the optimal AF calibration for the lens, but I think the result is already sharper than the one by the OP.

Although I admit that it is dangerous to compare results like that, I believe it is a further piece of evidence that indicates that the OP's conclusion may not be valid.
One difference between the OP's and my test was that he used the K100D's default setting which is "bright" and I used "natural". That means my shots received less sharpening out of the camera.

Sorry, atm I cannot do better than just link to the image image in the original size (1200 pixels wide, downscaled from a 1300 wide crop). You need to click on the magnifying glass to see the full version instead of a 800 pixel wide version.

EDIT: Here's another link to another version with a better (but probably not optimal AF calibration) for the prime: prime, 1300 wide crop, downscaled to 1200.


EDIT: I noticed that the previously attached file got truncated. I therefore now link to an external version which is really 1200 pixels wide.

Last edited by Class A; 01-11-2009 at 10:57 PM.
Reply

Bookmarks
  • Submit Thread to Facebook Facebook
  • Submit Thread to Twitter Twitter
  • Submit Thread to Digg Digg
Tags - Make this thread easier to find by adding keywords to it!
18-55mm, 50mm, camera, dslr, k100ds, k20d, kit, lens, photography, resolution, shot, test, time
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
For Sale - Sold: K100DS Like New waterhouse8800 Sold Items 2 07-28-2009 08:09 PM
K100DS and Manual WB res3567 Pentax DSLR Discussion 2 05-24-2009 09:14 PM
My K100ds is broken pasipasi Pentax DSLR Discussion 5 10-06-2008 12:28 PM
K100DS Firmware aamir515 Photographic Technique 4 10-04-2008 08:00 PM
K100DS issue DanLoc78 Photographic Technique 6 05-30-2008 08:18 AM



All times are GMT -7. The time now is 03:28 PM. | See also: NikonForums.com, part of our network of photo forums!
  • Red (Default)
  • Green
  • Gray
  • Dark
  • Dark Yellow
  • Dark Blue
  • Old Red
  • Old Green
  • Old Gray
  • Dial-Up Style
Hello! It's great to see you back on the forum! Have you considered joining the community?
register
Creating a FREE ACCOUNT takes under a minute, removes ads, and lets you post! [Dismiss]
Top