Pixel Shifting on the new FF - The benefits

mikeodial · 04-23-2015, 08:03 AM

Now the K3ii is announced, it is pretty obvious the Pixel Shifting/Super Resolution mode in this camera will be in the FF. What is interesting to note if this feature executes quickly enough, it could provide medium format quality for certain types of landscape or "non moving" targets. While some are complaining about having to use to a tripod, the benefits could be massive from this technology in terms of effective resolution and detail. I would certainly try an use it in outdoor photography of all sorts. I also wonder what the impact of this would be in creating richer color ala a more realistic HDR from a single exposure.

Also if you could use this feature with bracketing could open up a whole world of better HDR imagery.

Very interesting.

**jatrax** · 04-23-2015, 08:15 AM

Cannot wait to try it, lots of interesting applications come to mind. The big question I have is whether it will be practical outdoors with wind causing leaf motion.

The studio applications are really going to be interesting.

mikeodial · 04-23-2015, 08:23 AM

Originally posted by jatrax

Cannot wait to try it, lots of interesting applications come to mind. The big question I have is whether it will be practical outdoors with wind causing leaf motion.

The studio applications are really going to be interesting.

Yes, the leaf motion may be a problem, but if there is better dynamic range and improved picture quality, it may be possible to overcome some of that with higher shutter speeds. My understanding this will work in camera only, but combined with a nodal rail and panorama head, could permit some really high res/high quality shots previously only possible on much higher resolution platforms/film.

vonBaloney · 04-23-2015, 08:41 AM

My big question is what sort of files will it produce and what software will be able to deal with these files? Are we talking jpg-only after in-camera synthesis? (There was a confusing sentence about this on the Ricoh page about it.) Or a RAW file, but not a Bayer RAW file? Will we able to use the results with the usual programs (Lightroom, etc) that work with RAW files and be able to have all the usual adjustments available to us? I'd love this technology, but at what cost in post-processing?

derekkite · 04-23-2015, 08:53 AM

Originally posted by vonBaloney

My big question is what sort of files will it produce and what software will be able to deal with these files? Are we talking jpg-only after in-camera synthesis? (There was a confusing sentence about this on the Ricoh page about it.) Or a RAW file, but not a Bayer RAW file? Will we able to use the results with the usual programs (Lightroom, etc) that work with RAW files and be able to have all the usual adjustments available to us? I'd love this technology, but at what cost in post-processing?

Maybe the stacked output shooting modes is an indication. The raw files contain each shot and the finished one. To access them you need to use the pentax software, other raw converters only see the finished version.

JohnBee · 04-23-2015, 08:58 AM

Originally posted by mikeodial

Now the K3ii is announced, it is pretty obvious the Pixel Shifting/Super Resolution mode in this camera will be in the FF. What is interesting to note if this feature executes quickly enough, it could provide medium format quality for certain types of landscape or "non moving" targets. While some are complaining about having to use to a tripod, the benefits could be massive from this technology in terms of effective resolution and detail. I would certainly try an use it in outdoor photography of all sorts. I also wonder what the impact of this would be in creating richer color ala a more realistic HDR from a single exposure.

Also if you could use this feature with bracketing could open up a whole world of better HDR imagery.

Very interesting.

Based on what I've seen with the K3II, I'd say I'll likely buy the new Pentax FF purely based on this function if it delivers. Though I'd need RAW in order to make full use of it. Otherwise, I'd be better served using a third party software such as Photo Acute which isn't camera dependant.

quant2325 · 04-23-2015, 09:14 AM

1) The files shouldn't be much bigger, if at all. This is a completely different way of doing this than Olympus. You will get 48MP (or whatever) effective resolution without massive files.
2) The K-3II obviously becomes the sharpest APS-C camera on the market, assuming you are a landscape photographer (use a tripod). Your photos of the next Yosemite vacation (or macro of a still bug) will have enough sharpness to put large posters on your wall.
3) This technology will be massive for the pros once it hits the FF and medium format Pentax cameras. They use a tripod anyway, so using this technology for landscape photography is no big deal.

LeGuillotine · 04-23-2015, 10:09 AM

Could it be possible to use focus correction system to make DOF deeper, in macro shooting for example?

**Rorschach** · 04-23-2015, 10:58 AM

Originally posted by LeGuillotine

Could it be possible to use focus correction system to make DOF deeper, in macro shooting for example?

Focus stacking, or Z-stacking as it is known, should be relatively easy to automatize with AF macro or other lenses. Obviously it would be tripod only, mirror up and probably electronic shutter required, plus you'd maybe need a fraction of a second of wait time between each shot to let the vibration caused by the focusing motor and movement to die. Also, unless the processing pipeline is REALLY beefed up, I doubt you could combine the image stack in camera at least if it is composed of more than say three images. Or maybe you could but processing time might be excruciatingly long.

At my work, I have done quite a bit of this Z-stacking on a Zeiss StereoDiscoveryV12 stereomicroscope, it's Zeiss camera and PC-based software. A stack of 20-30 images takes from 30s to a minute or two depending on how fast a PC we are talking about. Vibrations and movement obviously are to be avoided at all costs which is why the system sits on a freaking steady and heavy stone table. But the end result is worth it

In microscopy magnifications the DOF is, well, microscopic (

), so it really is a game changer. You get images that are in focus throughout the specimen: insect, moss, mineral fragment, your pubic hair, whatever

Edit: forgot to mention that those Zeiss microscope cams are truly puny in the megapixel department, from 3 to 5 depending on model, if my memory serves. So a stack of even just 5-10 full resolution Pentax FF or even K-3 images would be a real acid test for any in-camera imaging processor, I'd imagine.

Gimbal · 04-23-2015, 12:58 PM

Originally posted by quant2325

1) The files shouldn't be much bigger, if at all. This is a completely different way of doing this than Olympus. You will get 48MP (or whatever) effective resolution without massive files.

It will be three times as big as a normal raw file. (unless of course the result is a compressed jpeg).
A normal rawfile is 14bits per pixel of gray info, a full color version has 14 bits of red, green and blue for every pixel. So it's three times as large.

quant2325 · 04-23-2015, 01:28 PM

Originally posted by Gimbal

It will be three times as big as a normal raw file. (unless of course the result is a compressed jpeg).
A normal rawfile is 14bits per pixel of gray info, a full color version has 14 bits of red, green and blue for every pixel. So it's three times as large.

Ricoh is taking four pixels of each Bayer color filter array cell--twice as many green than blue or red--and aligning them (like one on top of the other). The file should have the same number of pixels. The difference is that each pixel now has full G,R and B data. This is completely different than the Olympus approach. Ricoh is not adding more cells.

Barry Pearson · 04-23-2015, 01:54 PM

Originally posted by vonBaloney

My big question is what sort of files will it produce and what software will be able to deal with these files? Are we talking jpg-only after in-camera synthesis? (There was a confusing sentence about this on the Ricoh page about it.) Or a RAW file, but not a Bayer RAW file? Will we able to use the results with the usual programs (Lightroom, etc) that work with RAW files and be able to have all the usual adjustments available to us? I'd love this technology, but at what cost in post-processing?

I would think they will be DNG files with three samples per pixel. This option has been in DNG for a long time, probably from the start. And hence long supported by Lightroom and ACR.

(For example, doing a Pano-merge in Lightroom 6 creates a DNG file with 3 samples per pixel, each one a 16-bit integer).

I expect there will be a JPEG option too, but it would be very silly not to have a raw version.

mikeodial · 04-23-2015, 07:26 PM

Just saw these links on dpreview posted as (in Japanese) official samples. Can't read Japanese, but the images are amazing.

http://www.ricoh-imaging.co.jp/japan/products/k-3-2/ex/img/bod_mainImg_01.JPG

http://www.ricoh-imaging.co.jp/japan/products/k-3-2/ex/img/bod_mainImg_02.JPG

Read more at: https://www.pentaxforums.com/forums/6-pentax-dslr-discussion/293823-pixel-shi...#ixzz3YBp7SKZu

---------- Post added 04-23-15 at 10:30 PM ----------

Sample Images?K-3 II | RICOH IMAGING

This time in English. I am very impressed!

Gimbal · 04-24-2015, 04:00 AM

Originally posted by quant2325

Ricoh is taking four pixels of each Bayer color filter array cell--twice as many green than blue or red--and aligning them (like one on top of the other). The file should have the same number of pixels. The difference is that each pixel now has full G,R and B data.

Yes, the same number of pixels, but three times as much data for every pixel. Thus the file will take three times as much space.

jonby · 04-24-2015, 02:34 PM

Originally posted by Gimbal

Yes, the same number of pixels, but three times as much data for every pixel. Thus the file will take three times as much space.

Probably true. However, I wonder if software will be able to work with the files faster than you might expect from an image with this much data in it? If demosaicing is carried out in the camera before creating the raw file then that should mean that the software doesn't have to do it. In theory this could mean that Lightroom/ACR may open files and generate previews pretty quickly. Or maybe I'm completely wrong - any thoughts? Of course we don't know whether the files will be supported yet either.

04-23-2015, 08:03 AM	#1
mikeodial Veteran Member Join Date: Jan 2014 Photos: Gallery \| Albums Posts: 370	Pixel Shifting on the new FF - The benefits Now the K3ii is announced, it is pretty obvious the Pixel Shifting/Super Resolution mode in this camera will be in the FF. What is interesting to note if this feature executes quickly enough, it could provide medium format quality for certain types of landscape or "non moving" targets. While some are complaining about having to use to a tripod, the benefits could be massive from this technology in terms of effective resolution and detail. I would certainly try an use it in outdoor photography of all sorts. I also wonder what the impact of this would be in creating richer color ala a more realistic HDR from a single exposure. Also if you could use this feature with bracketing could open up a whole world of better HDR imagery. Very interesting.

04-23-2015, 10:58 AM	#9
Rorschach Site Supporter Join Date: Nov 2007 Location: Kuusamo, Finland Photos: Gallery Posts: 652	Originally posted by LeGuillotine Could it be possible to use focus correction system to make DOF deeper, in macro shooting for example? Focus stacking, or Z-stacking as it is known, should be relatively easy to automatize with AF macro or other lenses. Obviously it would be tripod only, mirror up and probably electronic shutter required, plus you'd maybe need a fraction of a second of wait time between each shot to let the vibration caused by the focusing motor and movement to die. Also, unless the processing pipeline is REALLY beefed up, I doubt you could combine the image stack in camera at least if it is composed of more than say three images. Or maybe you could but processing time might be excruciatingly long. At my work, I have done quite a bit of this Z-stacking on a Zeiss StereoDiscoveryV12 stereomicroscope, it's Zeiss camera and PC-based software. A stack of 20-30 images takes from 30s to a minute or two depending on how fast a PC we are talking about. Vibrations and movement obviously are to be avoided at all costs which is why the system sits on a freaking steady and heavy stone table. But the end result is worth it In microscopy magnifications the DOF is, well, microscopic ( ), so it really is a game changer. You get images that are in focus throughout the specimen: insect, moss, mineral fragment, your pubic hair, whatever Edit: forgot to mention that those Zeiss microscope cams are truly puny in the megapixel department, from 3 to 5 depending on model, if my memory serves. So a stack of even just 5-10 full resolution Pentax FF or even K-3 images would be a real acid test for any in-camera imaging processor, I'd imagine. Last edited by Rorschach; 04-23-2015 at 11:23 AM.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
What's the first lens you would mount on your new FF Pentax DSLR?	6BQ5	Pentax Full Frame	269	01-06-2016 04:20 PM
Imaging Resource Article on Ricoh Pixel-Shifting	mgvh	Pentax DSLR Discussion	26	03-30-2015 05:03 AM
Super resolution by pixel shifting in future Pentax cameras	Mistral75	Pentax News and Rumors	190	03-29-2015 04:31 AM
The Best Picture of the New Pentax FF	LaurenOE	Pentax Full Frame	27	02-09-2015 09:05 PM
Is Pentax still asking for feature input on the new FF?	mikeodial	Pentax News and Rumors	2	02-09-2015 06:42 AM

04-23-2015, 08:15 AM	#2
jatrax Site Supporter Join Date: May 2010 Location: Washington Cascades Photos: Gallery \| Albums Posts: 12,991	Cannot wait to try it, lots of interesting applications come to mind. The big question I have is whether it will be practical outdoors with wind causing leaf motion. The studio applications are really going to be interesting.

04-23-2015, 08:23 AM	#3
mikeodial Veteran Member Join Date: Jan 2014 Photos: Gallery \| Albums Posts: 370 Original Poster	Originally posted by jatrax Cannot wait to try it, lots of interesting applications come to mind. The big question I have is whether it will be practical outdoors with wind causing leaf motion. The studio applications are really going to be interesting. Yes, the leaf motion may be a problem, but if there is better dynamic range and improved picture quality, it may be possible to overcome some of that with higher shutter speeds. My understanding this will work in camera only, but combined with a nodal rail and panorama head, could permit some really high res/high quality shots previously only possible on much higher resolution platforms/film.

04-23-2015, 08:41 AM	#4
vonBaloney Pentaxian Join Date: Oct 2011 Location: Albuquerque, NM Posts: 6,029	My big question is what sort of files will it produce and what software will be able to deal with these files? Are we talking jpg-only after in-camera synthesis? (There was a confusing sentence about this on the Ricoh page about it.) Or a RAW file, but not a Bayer RAW file? Will we able to use the results with the usual programs (Lightroom, etc) that work with RAW files and be able to have all the usual adjustments available to us? I'd love this technology, but at what cost in post-processing?

04-23-2015, 08:53 AM	#5
derekkite Pentaxian Join Date: Sep 2011 Location: Nelson B.C. Posts: 3,782	Originally posted by vonBaloney My big question is what sort of files will it produce and what software will be able to deal with these files? Are we talking jpg-only after in-camera synthesis? (There was a confusing sentence about this on the Ricoh page about it.) Or a RAW file, but not a Bayer RAW file? Will we able to use the results with the usual programs (Lightroom, etc) that work with RAW files and be able to have all the usual adjustments available to us? I'd love this technology, but at what cost in post-processing? Maybe the stacked output shooting modes is an indication. The raw files contain each shot and the finished one. To access them you need to use the pentax software, other raw converters only see the finished version.

04-23-2015, 08:58 AM	#6
JohnBee Veteran Member Join Date: Sep 2008 Location: Newrfoundland Photos: Albums Posts: 4,667	Originally posted by mikeodial Now the K3ii is announced, it is pretty obvious the Pixel Shifting/Super Resolution mode in this camera will be in the FF. What is interesting to note if this feature executes quickly enough, it could provide medium format quality for certain types of landscape or "non moving" targets. While some are complaining about having to use to a tripod, the benefits could be massive from this technology in terms of effective resolution and detail. I would certainly try an use it in outdoor photography of all sorts. I also wonder what the impact of this would be in creating richer color ala a more realistic HDR from a single exposure. Also if you could use this feature with bracketing could open up a whole world of better HDR imagery. Very interesting. Based on what I've seen with the K3II, I'd say I'll likely buy the new Pentax FF purely based on this function if it delivers. Though I'd need RAW in order to make full use of it. Otherwise, I'd be better served using a third party software such as Photo Acute which isn't camera dependant.

04-23-2015, 09:14 AM	#7
quant2325 Senior Member Join Date: Jul 2012 Photos: Gallery Posts: 205	1) The files shouldn't be much bigger, if at all. This is a completely different way of doing this than Olympus. You will get 48MP (or whatever) effective resolution without massive files. 2) The K-3II obviously becomes the sharpest APS-C camera on the market, assuming you are a landscape photographer (use a tripod). Your photos of the next Yosemite vacation (or macro of a still bug) will have enough sharpness to put large posters on your wall. 3) This technology will be massive for the pros once it hits the FF and medium format Pentax cameras. They use a tripod anyway, so using this technology for landscape photography is no big deal.

04-23-2015, 10:09 AM	#8
LeGuillotine Senior Member Join Date: Jun 2009 Location: Finland, Kotka Photos: Gallery \| Albums Posts: 183	Could it be possible to use focus correction system to make DOF deeper, in macro shooting for example?

04-23-2015, 12:58 PM	#10
Gimbal Pentaxian Join Date: Apr 2007 Location: Sweden Posts: 2,106	Originally posted by quant2325 1) The files shouldn't be much bigger, if at all. This is a completely different way of doing this than Olympus. You will get 48MP (or whatever) effective resolution without massive files. It will be three times as big as a normal raw file. (unless of course the result is a compressed jpeg). A normal rawfile is 14bits per pixel of gray info, a full color version has 14 bits of red, green and blue for every pixel. So it's three times as large.

04-23-2015, 01:28 PM	#11
quant2325 Senior Member Join Date: Jul 2012 Photos: Gallery Posts: 205	Originally posted by Gimbal It will be three times as big as a normal raw file. (unless of course the result is a compressed jpeg). A normal rawfile is 14bits per pixel of gray info, a full color version has 14 bits of red, green and blue for every pixel. So it's three times as large. Ricoh is taking four pixels of each Bayer color filter array cell--twice as many green than blue or red--and aligning them (like one on top of the other). The file should have the same number of pixels. The difference is that each pixel now has full G,R and B data. This is completely different than the Olympus approach. Ricoh is not adding more cells.

04-23-2015, 01:54 PM	#12
Barry Pearson Veteran Member Join Date: Jul 2013 Location: Stockport Posts: 964	Originally posted by vonBaloney My big question is what sort of files will it produce and what software will be able to deal with these files? Are we talking jpg-only after in-camera synthesis? (There was a confusing sentence about this on the Ricoh page about it.) Or a RAW file, but not a Bayer RAW file? Will we able to use the results with the usual programs (Lightroom, etc) that work with RAW files and be able to have all the usual adjustments available to us? I'd love this technology, but at what cost in post-processing? I would think they will be DNG files with three samples per pixel. This option has been in DNG for a long time, probably from the start. And hence long supported by Lightroom and ACR. (For example, doing a Pano-merge in Lightroom 6 creates a DNG file with 3 samples per pixel, each one a 16-bit integer). I expect there will be a JPEG option too, but it would be very silly not to have a raw version.

04-23-2015, 07:26 PM	#13
mikeodial Veteran Member Join Date: Jan 2014 Photos: Gallery \| Albums Posts: 370 Original Poster	Just saw these links on dpreview posted as (in Japanese) official samples. Can't read Japanese, but the images are amazing. http://www.ricoh-imaging.co.jp/japan/products/k-3-2/ex/img/bod_mainImg_01.JPG http://www.ricoh-imaging.co.jp/japan/products/k-3-2/ex/img/bod_mainImg_02.JPG Read more at: https://www.pentaxforums.com/forums/6-pentax-dslr-discussion/293823-pixel-shi...#ixzz3YBp7SKZu ---------- Post added 04-23-15 at 10:30 PM ---------- Sample Images?K-3 II \| RICOH IMAGING This time in English. I am very impressed!

04-24-2015, 04:00 AM	#14
Gimbal Pentaxian Join Date: Apr 2007 Location: Sweden Posts: 2,106	Originally posted by quant2325 Ricoh is taking four pixels of each Bayer color filter array cell--twice as many green than blue or red--and aligning them (like one on top of the other). The file should have the same number of pixels. The difference is that each pixel now has full G,R and B data. Yes, the same number of pixels, but three times as much data for every pixel. Thus the file will take three times as much space.

04-24-2015, 02:34 PM	#15
jonby Senior Member Join Date: Nov 2013 Posts: 161	Originally posted by Gimbal Yes, the same number of pixels, but three times as much data for every pixel. Thus the file will take three times as much space. Probably true. However, I wonder if software will be able to work with the files faster than you might expect from an image with this much data in it? If demosaicing is carried out in the camera before creating the raw file then that should mean that the software doesn't have to do it. In theory this could mean that Lightroom/ACR may open files and generate previews pretty quickly. Or maybe I'm completely wrong - any thoughts? Of course we don't know whether the files will be supported yet either.