Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
03-20-2013, 12:09 PM   #16
Moderator
Tom S.'s Avatar

Join Date: May 2008
Location: S.E. Michigan
Photos: Albums
Posts: 4,239
How does the K5 do with sharpness turned up in the camera? I believe you said it was set at -2, but what if it's set to +2 or +4?

03-20-2013, 12:35 PM   #17
Pentaxian
fs999's Avatar

Join Date: Jul 2008
Location: Luxembourg
Photos: Gallery
Posts: 3,746
Over sharpened...
03-20-2013, 04:23 PM   #18
Pentaxian




Join Date: Oct 2009
Location: North
Photos: Gallery
Posts: 3,921
QuoteOriginally posted by panoguy Quote
What this shows me is that, in practical terms, the "resolution" difference between these two camera & lens combos is getting less and less, while things like color response and dynamic range (as well as lens quality) are kicking in. Like I said, the first image is really intended to shock, since it is a "straight" comparison, while I rarely show straight images. Personally, I feel that careful post-processing is an equally important part of making the image.

Thanks for the trouble of going thru more samples.
I certainly agree with your analysis/observations on how the gap lessens once the K5 images are processed and for most needs.

The magenta/green cast is a well known issue with the Foveon sensor.
That said, without any comparison, I'd not know and can probably live with it.

I'd think the Sigma will print bigger though.
03-20-2013, 05:21 PM   #19
Pentaxian
Clicker's Avatar

Join Date: May 2008
Photos: Gallery
Posts: 3,241
DPs are best shot on a tripod according to Devlin.

03-20-2013, 06:41 PM   #20
Pentaxian
panoguy's Avatar

Join Date: Sep 2007
Location: Washington, D.C.
Photos: Gallery | Albums
Posts: 3,290
Original Poster
QuoteOriginally posted by Tom S. Quote
How does the K5 do with sharpness turned up in the camera? I believe you said it was set at -2, but what if it's set to +2 or +4?
I'm processing the raw files, so any in-camera settings regarding sharpness are ignored. Those settings only effect jpegs...

The goofy part is that, by default, Adobe Camera Raw applies a +25 amount of sharpening to raw files (perhaps to make up for the AA filter on Bayer sensor images?), and Sigma Photo Pro has a -2 to +2 slider for sharpening, but it defaults to 0, which actually does apply some sharpening! So, I've set the Sigma to -2 for all of these images, but that's not in-camera, rather in the processing software.

QuoteOriginally posted by pinholecam Quote
Thanks for the trouble of going thru more samples.
I certainly agree with your analysis/observations on how the gap lessens once the K5 images are processed and for most needs.

The magenta/green cast is a well known issue with the Foveon sensor.
That said, without any comparison, I'd not know and can probably live with it.

I'd think the Sigma will print bigger though.
You are definitely right about printing bigger! The Sigma images have clear details between pixels that the K-5 just doesn't, so the images can be upsampled (enlarged) quite a bit before you start to see the "mushiness" that comes from upsampling Bayer sensor images. I know this both from my "experiments" with these files in Photoshop, and my many years of seeing my retouching work printed out much larger than originally intended (love that crop from a 35mm frame turned into a billboard!). We used to use "Genuine Fractals" to upsize (what's it called now? Perfect Resize?), but I think there are Photoshop actions (of repetitive incremental upsampling) which can do just as well... and that the Sigma files won't need!

QuoteOriginally posted by Clicker Quote
DPs are best shot on a tripod according to Devlin.
Ah, the Devlin is in the details! No doubt these Sigmas are best on a tripod, since even ISO 400 can look pretty bad, and there is no "shake reduction" on the sensor or lens!
That's why my images are taken in bright locations with pretty obvious mid-day sun to keep shutter speeds up, and not at the more interesting sunrise or sunset times. I've heard the Sigma excels at long exposures at night as well, but I've had a tough time finding a compatible arca-plate for it from my collection...

Last edited by panoguy; 03-20-2013 at 06:47 PM.
03-20-2013, 07:26 PM   #21
Pentaxian
panoguy's Avatar

Join Date: Sep 2007
Location: Washington, D.C.
Photos: Gallery | Albums
Posts: 3,290
Original Poster
One last comparison

One last set of comparison images, mainly because I've been showing how the sharpened K-5 output is a pretty close match for the unsharpened DP2 Merrrill output. Now I let the Sigma Photo Pro slide with it's own "default" amount of sharpening. One thing to note is that I intentionally exposed the Pentax for the highlights and lifted the shadows (oh, the luscious dynamic range) and exposed the Sigma for the shadows and recovered the highlights, so they have very different shutter speeds!

If there is one area where the Foveon excels, that is in actually distinguishing fine details with color and nuance, like branches in the distance. The Foveon sensor even picks up pixel-level color details, which is why I have to be careful not to turn on "defringe" in ACR or do excessive chroma noise removal to the converted tiffs from Sigma Photo Pro... Adobe Camera Raw is basically built for Bayer sensor cameras (as Fuji X-trans users have found out), and many of its functions which help the K-5 image will destroy the details of the Foveon image. In these backlit trees, the K-5 looks sub-par, but that isn't a fault of the Pentax, or the lens, or the processing... just the way the physical sensors work. I would be interested to try this again with a K-5IIs just to be sure it isn't entirely AA filter destruction!

Anyhow... here are the images, with slightly different framing because they were shot handheld and it was freezing outside! (And excuse the ham-fisted highlight recovery on the Sigma... I'm just not used to the supplied software yet!)


The K-5 & 31mm:



The Sigma DP2 Merrill:




And the obligatory (and eye-popping, due to Sigma's default sharpening) 100% zoom comparison, with the Sigma on the right (click to expand):



**Yes, the K-5 image has been sharpened! This isn't a "no sharpening at all" test like the first image compare. This is (roughly) 1:1 processing and what to expect for visible details.

As a coda to all of this, if anyone wants the full-res files, or even the raw files from this informal test, just send me a PM. I can zip them up and put them on Dropbox.

Last edited by panoguy; 03-20-2013 at 07:50 PM.
03-20-2013, 07:39 PM - 1 Like   #22
Veteran Member
joe.penn's Avatar

Join Date: Mar 2011
Location: Maryland (Right Outside Washington DC)
Photos: Gallery | Albums
Posts: 2,902
The sigma is hands down quite a bit sharper, our beloved K5's aren't even close in comparing. Thanks for the comparisons...
03-20-2013, 08:10 PM   #23
Pentaxian




Join Date: Jun 2009
Location: GMT +10
Photos: Albums
Posts: 10,843
By all accounts Sigma worked very hard to put some very good glass in front of the Foveon sensor, and tuned the lenses specifically for the characteristics of the DP 1/2/3 bodies. The sharpness of the images we are seeing could be due as much to the quality of those optics as the unique features of Foveon. It's hard to judge.

03-20-2013, 08:50 PM   #24
Pentaxian




Join Date: Jan 2011
Location: New York
Posts: 3,311
@panoguy, thanks for this in-depth comparison.

The DP2 can certainly take excellent photos. Whether it's purely the sensor, or lack of AA filter, or glass ... the end result is impressive. It's not as flexible as a K5 (or any DSLR), but given a choice between buying an FA31 to add to my lens collection or the DP2, I would be strongly tempted by the DP2.

Care to continue with some high ISO shots?
03-20-2013, 09:50 PM   #25
Veteran Member




Join Date: Sep 2011
Location: Melbourne
Photos: Gallery | Albums
Posts: 307
IMO.... In all these images, what I see in not just a slight increase in sharpness but a huge increase in the 'black point'. There is a lot more Black in the detail on the DP2 images than on the K5, which may mislead the eye to think it is sharper...
03-21-2013, 09:39 AM   #26
Pentaxian
panoguy's Avatar

Join Date: Sep 2007
Location: Washington, D.C.
Photos: Gallery | Albums
Posts: 3,290
Original Poster
QuoteOriginally posted by rawr Quote
. The sharpness of the images we are seeing could be due as much to the quality of those optics as the unique features of Foveon. It's hard to judge.
It is quite true that these lenses were probably designed for the Foveon (despite some reports that they are identical to the NEX versions, which are also quite good), and the 30mm is exceptional. I'm not sure it is better than the 31 Limited... the bokeh on the 31 is much better, for example.

QuoteOriginally posted by DeadJohn Quote
Care to continue with some high ISO shots?
I'd really rather not... the excessive color noise starting at 800 is not pretty at all. Not pretty unless, of course, I can convert them to monochromes...

QuoteOriginally posted by bdparker Quote
IMO.... In all these images, what I see in not just a slight increase in sharpness but a huge increase in the 'black point'. There is a lot more Black in the detail on the DP2 images than on the K5, which may mislead the eye to think it is sharper...
The actual black points are almost identical, but the large amount of darks in the details is quite obvious. In some cases it looks almost like the "black noise" found in overbaked HDRs, which is why I sometimes suspect some special sauce in the processing software. Or, it could be that the AA filter and Bayer interpolation of the K-5 are mushing out small contrasts that the Foveon sees, and the processing amplifies!

Last edited by panoguy; 03-21-2013 at 09:48 AM.
03-21-2013, 09:51 AM   #27
Forum Member




Join Date: Nov 2010
Location: CA
Posts: 70
Do you have access to a K5 IIs? It'll be interesting to see the differences of no-AA vs no-AA.
03-21-2013, 10:28 AM   #28
Pentaxian




Join Date: Jun 2009
Location: GMT +10
Photos: Albums
Posts: 10,843
FWIW, the Imaging Resource Comparometer has the SD1 Merrill, as well as the K5 IIs, D800E etc, allowing side-by-side comparisons.

Imaging Resource "Comparometer"

At say 200 ISO, even with just the 19mm Sigma lens on the SD1, the results are very nice, even compared to the K-5 IIs or D800E. It is said that the SD2 + 30mm is an even better performer, although alas that combo isn't on the Comparometer.

I still think the specially tuned Sigma glass is making a significant contribution to how nice the SD Merrill images look.
03-21-2013, 11:01 AM   #29
Forum Member




Join Date: Apr 2012
Posts: 78
I am sure that foveon-like aa-filter-less sensors will become the standard eventually.
But, the difference still seems a bit too much. For a real comparison only between sensors we would have to compare an SD1 against a K-5 with the same sigma lens (let's say the 35mm 1.4 for example).

P.S. It might be noise, JPG compression or just pure sharpness, but the image to the right seems sharpened. Especially the reflections in the windows and reflections on smooth surfaces (black cars).

Last edited by TrueFocus; 03-21-2013 at 11:10 AM.
03-21-2013, 11:16 AM   #30
Pentaxian
panoguy's Avatar

Join Date: Sep 2007
Location: Washington, D.C.
Photos: Gallery | Albums
Posts: 3,290
Original Poster
QuoteOriginally posted by TrueFocus Quote
For a real comparison only between sensors we would have to compare an SD1 against a K-5 with the same sigma lens (let's say the 35mm 1.4 for example).
Well, that's the problem... there is no version of that 35mm lens for Pentax (and frankly, most of the rest of Sigma's line aren't up to the level of the Limited primes, so everything would be degraded). As I said above, there are NEX mount lenses that are similar in spec to the DP2 and DP1 lenses, but different in # of elements and type of glass, etc. So, we have to take the best we've got!

Personally, I'd love to have an SD1 Merrill converted to Pentax K-mount, but that doesn't seem easily possible due to flange distance, despite the fact that Sigma's mount is physically very similar to Pentax. I think converting SA mounts to Canon EF is a better bet, due to them sharing the same electrical aperture and AF contacts! Some folks have already done this with earlier SD cameras... but $2k is a little rich for me to be mucking about modding the lensmount on a camera body!

QuoteOriginally posted by TrueFocus Quote
P.S. It might be noise, JPG compression or just pure sharpness, but the image to the right seems sharpened. Especially the reflections in the windows and reflections on smooth surfaces (black cars).
That's the Foveon difference. Even without any sharpening applied (which, in the screenshot below, is shown as "-2" on their scale), the images look sharper because they have much more detail per-pixel. It really takes certain visual situations to bring it out, but it is definitely not a sharpening routine.


Last edited by panoguy; 03-21-2013 at 11:59 AM.
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!
camera, color, comparison, dslr, k-5, k-5 ii, k-5 iis, k5, lens, light, pentax k-5, sigma
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
Pentax 645D colours and the Sigma Merrill larkis Pentax Medium Format 8 11-13-2012 10:45 AM
Are Tv and Av obsolete on the K-5? Yeti Pentax K-5 9 03-31-2012 04:32 PM
New Sigma DP1 and DP2 Merrill with the same 46MP Foveon X3 sensor from the Sigma SD1 UdonUdon Non-Pentax Cameras: Canon, Nikon, etc. 29 02-20-2012 12:30 PM
Sharpness and metering problems with Bellows on the K-5 Matchete Pentax K-5 25 02-14-2012 12:31 PM
Is the Autofocus up to speed on the K-5 or K-7? GLXLR Pentax K-5 30 06-03-2011 10:38 AM



All times are GMT -7. The time now is 06:29 PM. | See also: NikonForums.com, CanonForums.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