[cbaytan]

I though you might like that, who can resist such things right?

Anyways, the answer is:

A = K200D
B = K-5
C = D3x

Wondering How'did you do that? I mean the same size? Resampling? Becauese D3x test shot is here, it's the best dSLR sample I know of, even better than 645D. Or you just put the small introductory pictures?

Digital Cameras, Nikon D3X Digital Camera Test Image

[JohnBee]

For this demonstration, all images were downsampled to 10m/px to match the K200D. Otherwise, we'd be looking at upsampling which would of resulted in a nightmare(6x worst than native for D3x) for the lowly K200 I'm afraid. However... have you noticed the stark differences in texture, structural detail and color between the sample you've linked in your post and the ones posted in my previous comment?

And to this, I'd like to introduce you to the power of RAW imaging. That is to say that each sample was produced from RAW(not JPG) which allowed me to extract maximum detail from the files over that of in camera processing as shown in the IR image links. Which I hope will help explain where the power of a camera sensor lies with regards to IQ. To which I would add, 3D, POP, color, DR and detail are all attributes of RAW processing or development. And therefore, in many cases, the only way to tap into these potentials(ie. K-5) is in RAW.

In closing, if you would like to see what the K-5 sensor can do, I would be more than happy to share the step by step used to develop those files so you can see first hand just how close the K-5 actually comes to the mighty D3x. And in some areas(DR and color) where some might even call it superior.

[cbaytan]

In closing, if you would like to see what the K-5 sensor can do, I would be more than happy to share the step by step used to develop those files so you can see first hand just how close the K-5 actually comes to the mighty D3x. And in some areas(DR and color) almost feels superior.

I'd be grateful, really.

[JohnBee]

I'd be grateful, really.

Okay, give me a few and I'll prepare the links(to the RAW files I used) and a a list of the settings and steps used in processing.
PS. Do you have Photoshop and Adobe Camera Raw(ACR)?

-

Okay good....
For the good of the community, I'll post the walk-through here in hopes of helping others along the way(who knows right).
Anyways, here's the list of files and tools that were used in this particular processing.

1. Adobe Camera Raw - r6.4.0.139
2. Photoshop CS5
3. Perfectly Clear - color calibration
4. Topaz Detail - sharpening

Now obviously one could make due with different tools(ACR isn't the best), but since I've become quite lazy over the years, I've all but hung-up my curve adjustments and PS actions in favor of one click solutions like PerfectlyClear and Topaz products. And so, I'm not saying this is THE BEST method of getting things done, though it is the way I've chosen in getting things done as quickly and as efficiently as possible.

Okay, so here's the rundown...
The first thing we need to do is download the RAW image files:

K200D RAW image link
K-5 RAW image link
D3x RAW image link

btw. in case you wondered, I chose ISO200 samples so as to gain the full potential of DR on both sides of the histogram(shadow and highlights). ex. for maximum DR latitude in highlights.

1. K200D ACR Processing: - All other settings are 0, or OFF(no sharpening etc).


2. K-5 ACR Processing: - All other settings are 0, or OFF(no sharpening etc).


3. D3x ACR Processing: - All other settings are 0, or OFF(no sharpening etc).


4. Following the ACR adjustements, the files are then ported to PS for post processing. At which point, the D3x and K-5 files were resized(constrained) w/Bicubic to 3872px to match K200D sampling respectively.

5. The next step come the curves adjustements. At which point I make use of a plugin called PerfectlyClear which is the equivalent of a tone/color and contrast/curve adjustment all in one convenient workspace.



6. The final step of the process, is(as always) the sharpening phase. And for this, used Topaz Detail, which is another multi-purpose plugin(similar to PerfectlyClear) that provides several detail tools(deblur, sharpen, microcontrast and structure etc) in one convenient workspace.


Well there you have it. And though the settings for each file may have varied slightly(see sharpening radius etc), the screen settings shown here, should be close enough to replicate the images with relative ease.

In closing, I'd like to add, that the K-5 RAW in this particular case was tailored to match the D3x respectively. And the main reason for this is where the D3x is a class leading FF sensor, and so I chose to use it as a benchmark for IQ in this particular exercise. However... I don't think the D3x is showing the most accurate colors in this particular scene(see: olive oil). And so I would admit that in doing so(matching), it was likely that the K-5 colors were pushed out of bounds in the process.

So that's about it.
Let me know how things go from this and if you have any questions.

JohnBee

---

Last edited by JohnBee; 05-10-2011 at 04:27 AM.

[john5100]

I'm reading this on my smartphone so I can't see the pictures that well but is this test fair? The K5 and D3x are at the top of their game and the K200d sensor came out 5 years ago. If I'm of base I apologize but...

[JohnBee]

I'm reading this on my smartphone so I can't see the pictures that well but is this test fair? The K5 and D3x are at the top of their game and the K200d sensor came out 5 years ago. If I'm of base I apologize but...

Actually, I'd say its completely unfair.
However, the images posted were in retrospect to the fact that the K-5 is indeed worthy of its position and title as the best APS-C sensor in class.

I doubt it if K-5's CMOS sensor can give same color quality and 3D look as K200D CCD sensor, don't we all see those hazy, soft K-5 samples every day, proudly presented as perfect images? From my internet watch, on the FF and APS-C arena there is only one camera can beat the K200D's color quality and 3D look, which is Nikon D3x, even though it has a CMOS sensor, I don't know how they succeeded. I don't know how Samsung succeeded to produce such good sensor on K200D/K10 either. Don't forget K200D sensor is the same with K10 which resurrected Pentax' old reputation in the digital arena. I believe Pentax owes a lot to this sensor.

And don't get me wrong, this is not an attack on anyone or anything, however, I had a feeling that cbaytan may have based his observations on in-camera processing rather than RAW sensor output.

Wondering How'did you do that? I mean the same size? Resampling? Becauese D3x test shot is here, it's the best dSLR sample I know of, even better than 645D. Or you just put the small introductory pictures?

Digital Cameras, Nikon D3X Digital Camera Test Image

EDIT: While going over the Imaging Resource files and I noticed that the K200D ISO200 file seems to have(what looks like) motion blur in it.
And so to keep things fair, I re-processed the files using the ISO100 sample instead.

Candidate A2 - K200D ISO100

Candidate B2 - K-5 ISO200

Candidate C2 - D3x ISO200

[Marc Sabatella]

EDIT: While going over the Imaging Resource files and I noticed that the K200D ISO200 file seems to have(what looks like) motion blur in it.

I was going to point out that it looked uncharacteristically soft as well, although I wasn't seeing it specifically as motion blur. Looking closer, you're absolutely right.

One thing I have always been very aware of when using the IR samples is that for some reason, they shot from *much* further away, using a *much* longer lens, on the K200D as compared to almost every other camera. This is plain to see just from the flattened perspective of the K200D shot versus the others, which was the dead giveaway to me looking at the images you posted at screen size. Look at the line formed where the objects meet the table. Compared to the others, where there is clearly visible depth to the image (bottles further away than salt shakers), the K200D looks like someone straightened it all out onto one plane. Checking out the EXIF, you see the K-5 was shot at 70mm and the D3X also at 70mm. That means the D3X was shot from somewhat closer than the K-5 due to the difference in FOV between the cameras. But the K200D was shot at 200mm for some reason, meaning it was, well, I'm too lazy to do the math for real, but from a focal length almost 3 times longer than the K-5 and the equivalent FOV of 4 times longer than the D3X. I don't care *what* the specific lenses used were, this difference in shooting distance has got to play a role in perceived sharpness, "pop" et al. It's too bad, really, they didn't keep the lens and shooting position constant within camera brand.

So while looking at Imaging Resource samples is fine for what it is, it's not really as well-controlled a test rig as you'd like. Shooting with entirely different lenses from entirely different distances and allowing the motion blur in the K200D ISO 200 sample - these are frankly rookie mistakes that I'm surprised to see IR committing considering how thorough they are about other matters. Kind of perplexing, really.

Still, I'd be shocked if a downsized image from a 16MP or 24MP camera didn't outperform a full size 10MP image in terms of detail, so I'm not sure that repeating the test with the same 70mm lens as used on the K-5 would change much about this particular comparison. Sometimes, resolution really does matter.

[JohnBee]

Still, I'd be shocked if a downsized image from a 16MP or 24MP camera didn't outperform a full size 10MP image in terms of detail, so I'm not sure that repeating the test with the same 70mm lens as used on the K-5 would change much about this particular comparison. Sometimes, resolution really does matter.

I concur! (<- always a laugh)

I think that when we get down what we could call an accurate sensor, then we inevitably reach a point where there is little room left for improvements. ie. the K200D possessed a very good sensor, and so there is little or nothing do add to that insofar as nominal performances go.

And so it looks as though the remainder of the improvements would fall along the lines of factors such as resolution, DR and sensitivity etc. And even though we find aspects such as color depth improvements, all of these would not offer any visible advantages within the scope of nominal output, and so I think its easy for people to overlook such inherent improvements(beyond resolution) that we find with a sensor such as the K-5 over that of the K200D. However, if/when it comes down to RAW processing, then all bets are off as we find ourselves tapping into the extended ranges within the RAW data. And so I think there's a discrepancy to that effect when it comes to quantifying sensor IQ in cases such as these.

Still... when talking micro-contrast, detail, 3D look etc etc. all of these seem to fall under the influence of resolution and resolving power. And so in this way I guess we could say the K-5 improved in many of the areas cbaytan mentioned in his initial comment.

PS. I've interacted with the owner of Imaging Resources on a number of occasions about testing methods and/or inconsistencies in many of the samples provided in there Comparometer. And the K-5 was no exception to this as the initial tests posed problems that made it difficult to draw accurate conclusions from them. At which point, he agreed to re-shoot them. However... he would not re-shoot the D7000 images which I think received a bum wrap in terms of sensitivity performance.

On a side note, I wish I had not given my K200D away to my sister. It would have been great to perform some head to head testing between the K200D, K-5 and K20D sensors in RAW.

[JohnBee]

Still, I'd be shocked if a downsized image from a 16MP or 24MP camera didn't outperform a full size 10MP image in terms of detail, so I'm not sure that repeating the test with the same 70mm lens as used on the K-5 would change much about this particular comparison. Sometimes, resolution really does matter.

Marc I wanted to ask you(and forgot), what your thoughts were on how the 24Mpx sensor looks compared to the 16Mpx K-5?
I was personally surprised at the results as I sharpened both files using the same profiles and if memory serves me right, I tweaked the radius settings on the 3Dx to extrapolate finer detail than with the 16MPx file.

Another notable aspect of this is where the 3Dx files were shot with a prime lens whereas the K-5 is shot with a 3'rd party zoom and the K200D is shot with the kit lens(<-bummer). Which raises the question as to how things would look if the bar was even?

Okay now I'm really thinking of ordering a refurbished K200D for a face-off! Unfortunately, I don't and will likely never own a D3x

[Marc Sabatella]

Marc I wanted to ask you(and forgot), what your thoughts were on how the 24Mpx sensor looks compared to the 16Mpx K-5?
I was personally surprised at the results as I sharpened both files using the same profiles and if memory serves me right, I tweaked the radius settings on the 3Dx to extrapolate finer detail than with the 16MPx file.

I didn't see a difference worth commenting on. Not surprising to me, since we're downsizing both more than enough to hide any rough edges, and these were not shot under challenging lighting conditions that might expose differences in noise.

the K200D is shot with the kit lens(<-bummer).

Really? You mean the 50-200? I always assumed they were at least using something like the DA*60-250 or DA*200.

[JohnBee]

Really? You mean the 50-200? I always assumed they were at least using something like the DA*60-250 or DA*200.

Ah sorry about that, I was actually thinking of the outdoor scene:



Which I thought did a better job at pushing DR and resolution. However, the problem here was where D3x uses a 50mm prime, the K-5 got a 17-70 and the K200D got all of the 18-55. Which all works to make it rather hard to draw any sound conclusions from.

However, the studio shots seem, to make use of * lenses. Though, in this case, the K-5 gets a Sigma zoom lens whereas the D3x gets another prime(yet again) and so again, making it very difficult to draw any meaningful conclusions from.

To which I'd add, all in all, as good as the Comparometer truly is in function, it is equally frustrating in function when it comes down to critical evaluations.

[john5100]

Still works great for shows...
Jimmy Eat World shot with K200D and Sigma 24-60mm F2.8

[TourDeForce]

This has been a really interesting thread. It also confirms my observations about the K200D - it is in a class by itself. Entry level camera with weather sealing, simplified shooting modes available, and image quality that STILL can potentially compare to much newer and more expensive cameras.

My co-worker went through several (newer model) Nikon and Canon cameras in an effort to obtain an equivalent 'WOW' factor with his photography. He even borrowed my K200D to determine if his skills were up to task because he could not re-produce the 3D effects and vibrance that seemed to come naturally from the Pentax .jpg files. At one point he even owned a KM - but to no avail.

I've considered upgrading, but the image characteristics I'm looking for just don't seem to be there in the K-X, K-7, and K-R. The K-5 would be my choice, but it's too expensive for a poor man like me.

[mamonmon]

I am a proud owner of pentax k200d. it's my first dslr and did lots of shots with it. i still own the k200d but i already upgraded to k7 for faster AF and higher iso.

it is an entry level camera when it was released but it doesn't feel like an entry level camera. top lcd, weather sealing and a big body.

[planedriver]

it's good to read this thread..and maybe I'm repeating myself..the k200d and it's sensor is just something special. First I bought the K-x..and the high iso was impressive..but not the low range...now with the K-5 both are excellent..however there's no way I'm going to sell or replace my k200..even the K-5's images look (too) digital compared to those taken with the k200...

Andras