Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
01-16-2012, 09:31 AM   #1
New Member




Join Date: Jul 2011
Location: Cornwall
Posts: 22
Sensor pattern on pictures?

I recently upgraded from a Samsung GX20 to a Pentax K5 (3 weeks old) and am fairly happy with it so far.

Taking some pictures during sunrise the day before yesterday, however, I encountered something I never experienced with my GX20: 4 out of about 150 frames had a distinct pattern in the picture. Visible on the LCD screen and also on the computer. Both pictures were taken against the rising sun with a Sigma 70mm f 2,8 Macro without the lens hood attached.
Both pictures have been converted to Jpeg in PS7 without any editing, apart from some sharpening.

EXIF data:
0629: AV; 1/250; f 5,0; ISO: 200; Pattern metering; RAW DNS; AWB; Highlight Correction: On; Shadow Correction: High; SR: On
0635: AV; 1/100; f 5,6; ISO: 400; Pattern metering; RAW DNS; AWB; Highlight Correction: On; Shadow Correction: High; SR: On


Firstly I didn't think much of it and so far I have been unable to take another picture in similar conditions showing the same pattern. I'm a bit lost and wanted to hear some different thoughts before I get in touch with SRS or Pentax.


Any guesses?

Thanks

Attached Images
View Picture EXIF
PENTAX K-5  Photo 
View Picture EXIF
PENTAX K-5  Photo 
01-16-2012, 09:42 AM   #2
Veteran Member
Anvh's Avatar

Join Date: Sep 2011
Posts: 4,616
That looks quite special.
I'm actually thinking this is PS7 his fault, try out a free raw converter to see if you still get the same results.
01-16-2012, 09:45 AM   #3
Ole
Administrator
Ole's Avatar

Join Date: Sep 2006
Location: Copenhagen, Denmark
Photos: Gallery | Albums
Posts: 4,407
It might be worth while to convert your RAW file to TIFF and compare the TIFF and the JPG in Photoshop. That would tell if what you see is just JPG compression artifacts.

Also try the Pentax (SilkyPix) RAW converter that came with the software that was included with the K-5. Again, convert to TIFF and JPG and then compare.
01-16-2012, 10:02 AM   #4
Senior Member
telly0050's Avatar

Join Date: May 2011
Posts: 233
cool effect!
i wish my camera can do this

01-16-2012, 10:50 AM   #5
New Member




Join Date: Jul 2011
Location: Cornwall
Posts: 22
Original Poster
Nothing to do with Jpeg compression or PS as it looks like that on the LCD on the K5 and DNG viewer in Windows. RAW or Jpeg makes no difference.
01-16-2012, 11:50 AM   #6
Ole
Administrator
Ole's Avatar

Join Date: Sep 2006
Location: Copenhagen, Denmark
Photos: Gallery | Albums
Posts: 4,407
QuoteOriginally posted by Samyo Quote
Nothing to do with Jpeg compression or PS as it looks like that on the LCD on the K5 and DNG viewer in Windows. RAW or Jpeg makes no difference.
What you see on the camera LCD is a JPG. I don't know what the DNG viewer presents - it could be the embedded JPG. But I tend to agree, that pattern must be something else.

Last edited by Ole; 01-16-2012 at 02:11 PM.
01-16-2012, 12:04 PM   #7
Pentaxian




Join Date: Oct 2011
Location: Albuquerque, NM
Posts: 6,027
My guess is something bizarre is going on with the sensor or something bizarre is going on with highlight or shadow correction algorithms which you had turned on. Definitely send those to Pentax even if you can't reproduce it. What firmware do you have? Did you get the brand new update (v1.12)?

01-16-2012, 12:19 PM   #8
Senior Member
JenniferLeigh's Avatar

Join Date: Feb 2011
Location: Southern Alberta
Posts: 281
Could it be a filter? My Hoya ND8 leaves a faint cross hatch type pattern on my images.
01-16-2012, 12:34 PM - 1 Like   #9
Veteran Member
v5planet's Avatar

Join Date: Jan 2010
Location: Seattle
Photos: Albums
Posts: 1,915
I doubt this has anything to do with something physical, like a filter. Because of the way Bayer camera sensors work they are sometimes prone to creating detail artificially where there was none. I am struggling to find a link to the article I read months ago, but I believe the thrust of it went something like this: the pixels in your raw file do not each map to a single physical pixel on the sensor surface. Instead, they are dynamically generated from information gathered by an array of adjacent "pixels" sensitive individually to Green, Blue and Red light. The camera's software has to make educated guesses as to how to use this information to create a single pixel, and sometimes -- especially in areas with lots of busy detail, like a field of grass -- it can goof and create lines and other patterns that simply weren't there.

This is what it looks like to me. However, as to why it was triggered, I'm not sure I can hazard a guess. You don't seem to have any fine detail in these images that would confuse the camera's software, especially in the first image where we would expect smooth, continuous tones. I do notice, however, that the pattern fades towards the right edge of the first image. Since both examples you gave us are clearly crops from a larger image, I would suggest going back to the uncropped originals and seeing if the area of the frame the the first image corresponds to shows a similar fading of the pattern in the second image. This would give you a hint as to whether it's an issue that varies across the sensor surface in a consistent fashion (in which case I'd think something was wrong with the sensor itself). If you don't see the same fading in the second image in that part of the frame, something else will be confusing the software, I would think.

[edit] Found an article that discusses this somewhat:

http://www.cambridgeincolour.com/tutorials/camera-sensors.htm

Scroll down to the 'Demosaicing Artifacts' section. The second row of (smaller) images, shows a pattern VERY similar to yours, second from the right. Now what I don't understand is what would be triggering your camera's sensor/sensor software into thinking you had fine repeating detail in either of these images.

Last edited by v5planet; 01-16-2012 at 12:42 PM.
01-16-2012, 04:34 PM   #10
Veteran Member
v5planet's Avatar

Join Date: Jan 2010
Location: Seattle
Photos: Albums
Posts: 1,915
To follow up, I found this thread regarding a similar problem with an Olympus camera:

Adobe Forums: Weird Demosaicing pattern on RAW files

The post I've linked directly to above mentions something interesting:

QuoteQuote:
This appears to be an example of a rare but well-known maze-pattern phenomena that occurs when the two green channels (there are twice as many green pixels as there are red and blue pixels so the green pixels are split into two distinct sets) don’t have the same amplification coming off the sensor but ACR/LR treats them as if they do, or when the compensation for how much they are off isn’t enough.
This is something that might be worth looking at. In particular, I would encourage you to go back to the offending images and see if the problem appears in the same part of the frame in each image. As I mentioned in the last post, I noticed your first image does not demonstrate the pattern towards the right edge. Locate what part of the uncropped frame that belongs to and look at the same spot in your other offending photos to see if the maze pattern is absent there also. In context of the thread I linked to, this would make me wonder if there is a problem in balance between the green photosites on your sensor, in which case you may want to talk to Pentax about a replacement sensor.
01-17-2012, 01:52 AM   #11
Veteran Member




Join Date: Jan 2009
Location: Budapest
Posts: 821
That's the famous maze artifact of demosaicing. Happens mostly if not the proper demosaicing algorithm is used in raw conversion, sometimes it can be forced to appear under special conditions. You shouldn't see that on the camera LCD though - try to reset your cam. Any chance you can upload/share the DNG straight from the camera?
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!
awb, camera, correction, dslr, gx20, k-5, k-5 ii, k-5 iis, k5, pattern, pentax, pentax k-5, shadow
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
Landscape Rock pattern RobG Post Your Photos! 2 06-26-2011 06:15 AM
Pentax Model Number Pattern yusuf Pentax DSLR Discussion 11 05-04-2011 09:08 PM
DA* 50-135 SDM failure pattern? yusuf Pentax SLR Lens Discussion 7 10-15-2010 10:00 AM
WINNERS Project 52-2-44 Subject: Pattern Iris Weekly Photo Challenges 13 05-20-2010 01:28 PM
Project 52-2-44 Subject: Pattern mithrandir Weekly Photo Challenges 22 05-17-2010 08:45 AM



All times are GMT -7. The time now is 03:14 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