[Ryan Trevisol]

Santkou, thanks for that. Good to know. I took a look again and I found some interesting stuff. I even busted out the ol' ftp client, too.

I think the K-30 has more sharpening and noise reduction going on, even at low ISO. Maybe even a contrast bump. Check out these caps from the ISO 400 comparison. K-5 is on the LEFT, K-30 on RIGHT:

http://www.trevisol.com/images/p/scale.PNG
In this one, the K-30 at first glance appears to have higher contrast, but I believe it's been sharpened. Look at the areas surrounding the numbers. They have that same halo effect that you get from aggressive sharpening.
http://www.trevisol.com/images/p/card.PNG
In this one, you can see that the texture of the card was interpreted as noise. Look particularly at the light blue square (on the k-5 side, it's top row, third from left).
http://www.trevisol.com/images/p/crayon.PNG
The text on the box looks like it might have been better in focus for the K-30, or it possibly had sharpening and contrast applied to it.
http://www.trevisol.com/images/p/fabric.PNG
Once again, the K-5 shows more texture in the fabric, whereas the K-30 side looks like it had software NR run on it.

As stated below, these anomalies appear to be a result of IR's JPEG settings - note that the K-5 JPG's weighed in at 11MB's whereas the K-30's JPG's are just over 7MB.

Good point. The scientist and consumer in me is interested in compare and contrast at a fine level, however the practical just-noticeable-difference may be insignificant in practice. And for those already with K-5's I am sure this all a very mute point, however for those "upgrading" from prior mid-line dSLR's, there are some significant technical decisions to make. Cheers !

Exactly how I feel. If the K-30 is better because it captures the same or greater amount of detail than the K-5, fine. But if the K-30 applies sharpening and NR that can't be turned off, (which in the case of the scale picture above is destructive) then the K-5 is still the superior camera for someone who knows what he's doing.

---

Last edited by Ryan Trevisol; 06-28-2012 at 06:58 AM.

[Pentaxor]

So, in other words, we probably know absolutely nothing until someone with a current firmware K-5 AND a production K-30, can produce some RAW files with the SAME lens, under the SAME conditions, with the SAME settings. Not that I really thought that anything currently available out there would answer the ultimate image processing and IQ questions, but it did resolve my angst momentarily with a brief illusion of control. So, any word on an exact date for North American retail availability in July ?

Santkou, thanks for that. Good to know. I took a look again and I found some interesting stuff. I even busted out the ol' ftp client, too.

I think the K-30 has more sharpening and noise reduction going on, even at low ISO. Maybe even a contrast bump. Check out these caps from the ISO 400 comparison. K-5 is on the LEFT, K-30 on RIGHT:

http://www.trevisol.com/images/p/scale.PNG
In this one, the K-30 at first glance appears to have higher contrast, but I believe it's been sharpened. Look at the areas surrounding the numbers. They have that same halo effect that you get from aggressive sharpening.
http://www.trevisol.com/images/p/card.PNG
In this one, you can see that the texture of the card was interpreted as noise. Look particularly at the light blue square (on the k-5 side, it's top row, third from left).
http://www.trevisol.com/images/p/crayon.PNG
The text on the box looks like it might have been better in focus for the K-30, or it possibly had sharpening and contrast applied to it.
http://www.trevisol.com/images/p/fabric.PNG
Once again, the K-5 shows more texture in the fabric, whereas the K-30 side looks like it had software NR run on it.



Exactly how I feel. If the K-30 is better because it captures the same or greater amount of detail than the K-5, fine. I'll buy it instead. But if the K-30 applies sharpening and NR that can't be turned off, (which in the case of the scale picture above is destructive) then the K-5 is still the superior camera for someone who knows what he's doing.

Edit: stupid capital letter file extensions....

had you actually checked the other areas of the test image for consistency to support this claim? or even checked the other ISOs?

[Anvh]

When comparing the D7000 and K5 (same sensor) it was already noted by some that the K5 uses probably a slightly heavier filter, all in all it gives you less noise and artefacts then the D7000 and the sharpness is still decent enough.

[Ryan Trevisol]

had you actually checked the other areas of the test image for consistency to support this claim? or even checked the other ISOs?

Of course I have. I didn't have time at work to put up a bunch of screenshots, but you're free to go to the page yourself and look at the same areas of the image. I picked ISO400 because it's reasonably sensitive without typically requiring in-camera NR ("High ISO" NR starts at 800 typically).

I looked at ISO 100, where the difference was less pronounced on the K-30's side. The K-5 looked very similar from ISO100 to ISO400. The K-30 looked like it had less NR applied, and more detail came through, but the sharpening artifacts were there on the scale.

Then I looked at 1600, where the mandatory NR on the K-30 really started to get out of control. At this ISO, the scale actually looked better, the NR caught up to the sharpening, but solid colors started to look mottled, whereas, as mentioned before, the K-5's noise was tightly controlled, and finer in nature. Instead of looking blotchy and lumpy, flat solid colors looked finely textured.

Here, I couldn't resist and took one more screenshot:
http://www.trevisol.com/images/p/card1600.png
It's CLEAR that there's some NR going on. The question is, is this a setting that IR didn't disable before taking the test shots, is it a difference in the AA filter, or does the K-30 have a certain amount of processing happening to RAW images that can't be disabled?
As stated below, these anomalies appear to be a result of IR's JPEG settings - note that the K-5 JPG's weighed in at 11MB's whereas the K-30's JPG's are just over 7MB.

---

Last edited by Ryan Trevisol; 06-28-2012 at 06:57 AM.

[Anvh]

With the K30 you also se clearly the sharpening halos with your last sample, quite interesting.

[loco]

I downloaded the ISO 6400 RAW file from the K-30 and opened it in Lightroom. I removed all sharpening applied in LR and I see no sign of halos or excessive NR. What am I missing here? Looks like a pretty impressive image to me. Certainly much better than I can get with my K-r at such a high ISO.

[goldenarrow]

Alas, the local dealer will NOT be stocking a K-30 to display. I can only special order one with down payment to get my hands on one. I wish I could actually use a K-5 and K-30 side by side. I suppose I could rent a K-5 and buy the K-30 with a return guarantee. However, by then, all the new K-5's will be gone. The K200D is nice, and I can do some creative things with it, but I can see and feel limitations that I just don't want to live with anymore. I really dislike having to upgrade electronics every three years but that must be the price to pay for saving so much in film and processing costs. Currently, I have spent about $220/year on Pentax bodies spread out over three years. I would hope that a new K-5 or a K-30 will last me about 4 years (good luck). I wonder what kind of camera body I will be pining for in 2014, 15, 16???? (No Canikon jokes please). And more capable sensor means upgraded lenses, tripod, etc. Photography is a cruel mistress.......

[Pentaxor]

I downloaded the ISO 6400 RAW file from the K-30 and opened it in Lightroom. I removed all sharpening applied in LR and I see no sign of halos or excessive NR. What am I missing here? Looks like a pretty impressive image to me. Certainly much better than I can get with my K-r at such a high ISO.

I checked it as well and I do agree with you. however, what does baffle me is where the notion of NR application on the K30 was emphasized? I checked on the entire image including other portions that weren't emphasized by the other poster. in the other areas, it seems that the detail quality of the K30 is better even when the sharpening is reduced. now, NR would had destroyed or removed those details easily if such amount of NR processing was really applied and very noticeable at pixel level.

[Ryan Trevisol]

Unfortunately I haven't had the time to sit down with Aperture and download the DNG's to do a proper comparison.

Interestingly, the K-30 test shots for the comparometer is the one ending in "NR0" which, on the 1600ISO one, reads in the EXIF as "NR Off", but still the comparometer JPG shows the blocky NR artifacts pictured above. Downloading the NR0 DNG of that same shot, just looking at it with Preview in Windows, shows finely detailed noise just like the K-5's. Also, at least at 1600, it didn't look obviously sharpened.

So, sorry if my comments were inflammatory or misleading. I've edited them to reflect the above information.

[toukan]

Forgive me, as I am relatively new to photography (at least real photography, the kind that requires/rewards technical ability), but my question is whether the 12 bit RAW capture of the K-30 is something that could upgraded to 14-bit through a firmware update, or is it inherent in the design of the camera, and thus not subject to alteration later. As I said, i am new to this and trying to learn, but as with many things, every new piece of knowledge gained seems to only lead to a greater awareness of how much one doesn't know and opens up numerous other avenues of inquiry. Thank you all though for your highly informative discussions. I have learned just as much reading the conversation threads as any "how to" guide.

[gorski]

my question is whether the 12 bit RAW capture of the K-30 is something that could upgraded to 14-bit through a firmware update, or is it inherent in the design of the camera, and thus not subject to alteration later.

Number of bits of the A/D converter (and hence the RAW) is a hardware thing, it cannot be upgraded with a firmware update. The image sensor outputs an analog voltage, which is proportional to a number of photons that hit the pixel, and this voltage needs to be digitalized into something that is interpretable to computers. A/D converter does that. The difference between the 12bit and 14bit converter is in an extra electronic circuit of transistors, resistors, capacitors, ... and this is on the lowest level of the hardware.

As you said, it is inherent to the design of the camera, and thus cannot be a subject to alteration later.

[Pentaxor]

Forgive me, as I am relatively new to photography (at least real photography, the kind that requires/rewards technical ability), but my question is whether the 12 bit RAW capture of the K-30 is something that could upgraded to 14-bit through a firmware update, or is it inherent in the design of the camera, and thus not subject to alteration later. As I said, i am new to this and trying to learn, but as with many things, every new piece of knowledge gained seems to only lead to a greater awareness of how much one doesn't know and opens up numerous other avenues of inquiry. Thank you all though for your highly informative discussions. I have learned just as much reading the conversation threads as any "how to" guide.

as was mentioned, it is a hardware matter. anyway, do not worry yourself about 12 bit and 14 bit conversion as this won't affect your images or make your image lesser than the other. and you won't notice it at all for your practical use.

[gorski]

In practice it will probably mean a little less margin for extracting shadow details. K-5 is renowned for good retention of details in shadows, so the K-30 should nevertheless still have a noticeably better performance in that respect than other dslr of the same range.

[toukan]

Then I guess the flip side of my question is whether focus peaking or any of the other supposed improvements of the K-30 can be add to the K-5 via a firmware update? thanks again for your comments.

[Ryan Trevisol]

I would speculate that the Prime M engine probably has increased processing power, and has been speculated to be more focused on video. A quote I read said that they're delivering a full 60 FPS on the back screen in live view. That, plus the hardware h.264 encoder probably make it easier for the K-30 to display focus peaking. The K-5's processor might not be able to handle the feature, but even if it could, it would more than likely it would decrease the LV framerate, and would probably deplete the battery even faster.