Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
03-13-2008, 09:32 PM   #1
Inactive Account




Join Date: Mar 2008
Location: Vancouver, BC
Posts: 2
K20D a sugestion to Pentax, for use of the SR system in Super Sampling

Hello all, I am new here, first post. I am contemplating the K20D to add to my film based Pentax cameras. An idea that I passed on to Pentax US through their feedback, which struck me as potentially feasible was to use the SR system in conjunction with still-life tripod mounted photos, for Super-Sampling.

Back in the late 1980s to early 1990s, I came across a video camera that was modified for high resolution capture, through using a piezo-electric substrate below the sensor chip. The piezo allowed the sensor to be moved, in a fashion similar to the Pentax Shake Reduction concept, but for the specific purpose of moving the sensor by a fraction of a pixel. The chip was then resampled, shifted again etc. The design permitted a total of 9 (nine) samples to be captured per frame.

Consider, capture move right 1/3 pixel, sample, move right 1/3 pixel, sample, drop 1/3 pixel, sample, etc until the chip had been moved for a total of 9 locations, 3 horizontally and 3 vertical. 3x3=9 samples.

I realize this is asking the SR to do something it was not directly designed around, but I think that for those instances of still life phootos that warrant the very highest quality, this could be somewhat useful. Within the 9 samples, exposure could also be trimmed fractionally, to increase the effective accumulated density range.

There clearly would be licensing issues to contend with, and it really could mean somewhat longish exposures of exclusively still life subjects. I do however feel, if there were expressed interest, amongst Pentax owners, that Pentax might give it a look.

Before you ask me what camera it was, I no longer have the article that had announced it. I do hope there may be some record of it out on the net. If you feel such may be a worthy idea, doubling or tripling the resolution, merely with a tripod, and a specially software coded custom function, I would ask that we join our voices and ask the fine people at Pentax to give it a look.

Thanks for the chance present an idea.

May your images turn out as you visualized.

03-15-2008, 04:40 AM   #2
Veteran Member
falconeye's Avatar

Join Date: Jan 2008
Location: Munich, Alps, Germany
Photos: Gallery
Posts: 6,863
Carbrofan,

you make an interesting proposal. Yes, it can be done. I don't think Pentax will jump onto your suggestion in the near future, though

Some additional remarks:
With the RGGB Bayer pattern, a shift by 1 pixel to get exact colors is another, more immediate, application of your idea.

And with subpixel shifts, one wouldn't actually supersample to get the best result. One would mathematically derive an image at subpixel resolutions (larger than sensor). You see this in image processing as done for solid state manufacturing (where pixels cannot be smaller than the light's wavelength in the first place).

Only very a few lenses would have the required reolution. I also don't know whether the SR motor is accurate enough. And you need a more solid tripod than usual. All of this is why Pentax won't consider it to be on their short list of things to do
03-15-2008, 05:18 AM   #3
Veteran Member
distudio's Avatar

Join Date: Jul 2007
Location: Sydney
Photos: Albums
Posts: 448
QuoteOriginally posted by Carbrofan Quote
Hello all, I am new here, first post. I am contemplating the K20D to add to my film based Pentax cameras. An idea that I passed on to Pentax US through their feedback, which struck me as potentially feasible was to use the SR system in conjunction with still-life tripod mounted photos, for Super-Sampling.
The software is readily available, whether Pentax would come to the party though is another matter.

PhotoAcute Studio. Leading superresolution technology for better photos.
03-15-2008, 02:47 PM   #4
Veteran Member




Join Date: Aug 2007
Location: Brazil
Posts: 377
This idea has been around for some time. I think the main issue is that even on a tripod the camera will move more than a pixel between shots, and very few DSLRs have the speed necessary for doing that in a single shot (D3/1Ds any others?).

They could come up with that and the SR working at the same time to avoid vibration and make sure the sensor is at the right place for each capture, but I think this precision is way beyond the cost and feasibility of our cameras. And a waste of time, because we are already hitting the limits of useful sensor resolution, it's the lenses that hold us back.

03-15-2008, 04:54 PM   #5
Veteran Member
falconeye's Avatar

Join Date: Jan 2008
Location: Munich, Alps, Germany
Photos: Gallery
Posts: 6,863
PhotoAcute Studio

Thanks for this link. I looked up the example section. Looks great! Does the software live up to those examples? Did you use it?
03-15-2008, 06:11 PM   #6
Veteran Member
distudio's Avatar

Join Date: Jul 2007
Location: Sydney
Photos: Albums
Posts: 448
QuoteOriginally posted by falconeye Quote
Thanks for this link. I looked up the example section. Looks great! Does the software live up to those examples? Did you use it?
I tried it about a year back, it definitely does what it claims but unfortunately the demo version is pretty limited and like programs such as DXO and PTLens it doesn't allow the user to construct profiles for unsupported lenses. Which if you don't have the lens/camera combos it supports kind of defeats the advantage.
03-15-2008, 06:17 PM   #7
Veteran Member




Join Date: May 2007
Location: Central NJ
Posts: 469
QuoteOriginally posted by ricardobeat Quote
This idea has been around for some time. I think the main issue is that even on a tripod the camera will move more than a pixel between shots, and very few DSLRs have the speed necessary for doing that in a single shot (D3/1Ds any others?).

They could come up with that and the SR working at the same time to avoid vibration and make sure the sensor is at the right place for each capture, but I think this precision is way beyond the cost and feasibility of our cameras. And a waste of time, because we are already hitting the limits of useful sensor resolution, it's the lenses that hold us back.
I suspect it's abit much to expect with regards to in-camera processing and the precision of the SR mechanism. However, this type of thing is designed explicitly to give more resolution when youa re hitting the limits of the system. Without going to very expensive glass, it might be the only good way to get more resolution because it isn't dependant on putting more receptors in a given area, or making the whole chip bigger, or needing higher resolution lenses.
03-15-2008, 06:33 PM   #8
Veteran Member
falconeye's Avatar

Join Date: Jan 2008
Location: Munich, Alps, Germany
Photos: Gallery
Posts: 6,863
QuoteOriginally posted by distudio Quote
like programs such as DXO and PTLens it doesn't allow the user to construct profiles for unsupported lenses. Which if you don't have the lens/camera combos it supports kind of defeats the advantage.
They write:
Can I use PhotoAcute Studio for the images taken with the camera/lens that are not in the "supported" list?
Yes.
Some of PhotoAcute Studio features are absolutely camera independent. These are: precise image alignment and merging, noise reduction, chromatic aberrations correction, dynamic range expansion and depth of field expansion.
Superresolution processing is camera-dependent, but when there is no profile for specific camera/lens, it is possible to use the profiles designed for other cameras (just select the desired profile in Settings->Camera). Usually it is possible to find a profile that gives good results for "unsupported" camera/lens combination. However, you should remember that using foreign profile will not give the maximum of possible quality improvement, and can lead to artefacts appearance in resultant images.
Click here to see an example of PhotoAcute Studio superresolution feature usage for an "unsupported" camera.
Click here to read the review by Uwe Steinmueller on Digital Outback Photo with the results of superresolution processing of the images taken with "unsupported" camera.
Was it like this when you tried it or may they have improved on it?

03-16-2008, 12:10 AM   #9
Inactive Account




Join Date: Mar 2008
Location: Vancouver, BC
Posts: 2
Original Poster
Thanks for the feedback all. Additionally thanks for the links, gives me something to look into.

While waiting for confirmation my posting had been accepted, I have yesterday picked up a K20D, to add to the arsenal. Just getting a start on learning what it already has to offer.

Much as I will likely be getting another dslr, for a full frame chip, I expect the K20D will regularly be in hand for quite a few years. As I have used large and medium format film cameras for decades, I constantly like to see what notions have been pushed to the boundaries.

By no means do I understand all the implications, but simply enjoy the personal discoveries that being aware of some concepts offer. And if those concepts trigger a fresh idea in someone's mind, then perhaps it may prove fruitful.

I can see that I will be very glad to have joined the forum.

Thanks again to all.
03-16-2008, 01:23 AM   #10
Veteran Member
distudio's Avatar

Join Date: Jul 2007
Location: Sydney
Photos: Albums
Posts: 448
QuoteOriginally posted by falconeye Quote
They write:
Can I use PhotoAcute Studio for the images taken with the camera/lens that are not in the "supported" list?
Yes.
Some of PhotoAcute Studio features are absolutely camera independent. These are: precise image alignment and merging, noise reduction, chromatic aberrations correction, dynamic range expansion and depth of field expansion.
Superresolution processing is camera-dependent, but when there is no profile for specific camera/lens, it is possible to use the profiles designed for other cameras (just select the desired profile in Settings->Camera). Usually it is possible to find a profile that gives good results for "unsupported" camera/lens combination. However, you should remember that using foreign profile will not give the maximum of possible quality improvement, and can lead to artefacts appearance in resultant images.
Click here to see an example of PhotoAcute Studio superresolution feature usage for an "unsupported" camera.
Click here to read the review by Uwe Steinmueller on Digital Outback Photo with the results of superresolution processing of the images taken with "unsupported" camera.
Was it like this when you tried it or may they have improved on it?
I suspect it is a little better than when I first reviewed it however as I mentioned the demo software is pretty limited in features/function and I wasn't prepared to dish out relatively big dollars just to test the program.
03-16-2008, 01:37 AM   #11
Veteran Member




Join Date: Nov 2006
Location: Hong Kong
Posts: 1,934
Nice idea, the same techniques are actually used in film scanners, which is called multi-pass (re-)sampling, which could help to increase DR, decrease noise and improve resolution. The difference is subtle, though.

Having said that, I doubt that if this would be practical for shooting real-time objects by a *camera* which would require instanteously image capture, though.
03-16-2008, 02:12 AM   #12
New Member




Join Date: Jan 2008
Location: Netherlands
Posts: 10
Coincidentally I downloaded the trial version of PhotoAcute studio last week. Firstly as far as I can see the trial version is fully featured, certainly good enough to assess the software. They just print the name and details over the output to spoil it, but this does not impede your ability to assess the results in any way.

So far I have only checked the Image resolution option. I used a K100D with a 35mm f2 (manual focus) and carefully followed the instructions for the RAW option. I can confirm that all 4 tests produced an acceptable image at twice the size of the original, but when these original images were upsized to the same size in Photoshop (bicubic resampling) the difference was marginal. However the Photoacute images were very slightly better. This software package provides several other options and is good value if you need any of them. I will try it with other camera/lens combinations and report any useful results.

The biggest problem for Pentax users is the limited lens support. So many lenses will work on these cameras (my favorite is a screw mount lens made around 1966) that there is no way they could test them all. If they could devise a way we could make our own profiles I would buy the software . The following are the lenses currently supported for the K100D and the K10D.

Pentax K100D
Pentax 18-55mm f/3.5-5.6 (RAW / JPG)
Pentax 35mm f/2 (RAW / JPG)
Pentax 50-200mm f/4-5.6 (RAW / JPG)
Pentax 50mm f/1.7 SMC (RAW)
Tokina 20-35mm f/2.8 (RAW)
Tokina 28-80mm f/2.8 (RAW)

Pentax K10D
Pentax 16-45mm f/4 (RAW / JPG)
Pentax 16-50mm f/2.8 (RAW / JPG)
Pentax 18-55mm f/3.5-5.6 (JPG)
Pentax 35mm f/2 (RAW / JPG)
Sigma 17-70mm f/2.8-4.5 Macro (RAW / JPG)
Sigma 24mm f/2.8 (RAW / JPG)
Tamron 18-250mm f/3.5-6.3 (RAW / JPG)
03-16-2008, 08:04 PM   #13
m8o
Veteran Member
m8o's Avatar

Join Date: Jul 2007
Location: 40-55'-44" N / 73-24'-07" W [on LI]
Photos: Gallery | Albums
Posts: 3,102
carbro, this may be what you are referring to? Pointed out by Buddah Jones a few months back... Helicon Soft: Helicon Focus
QuoteQuote:
HeliconFocus is a program that creates one completely focused image from several partially focused images by combining the focused areas.

The program is designed for macrophotography, microphotography and hyperfocal landscape photography to cope with the shallow depth-of-field problem.

Helicon Focus also aligns images as objects often change their size and position from shot to shot. This function is especially important for macrophotography.
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, chip, dslr, idea, life, look, pentax, photography, pixel, sample, samples, sensor, sr
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
Need your sugestion about the lens DA 70mm or 40mm dreamnet Pentax SLR Lens Discussion 20 10-03-2009 04:30 AM
Pentax photo outing... a sampling creampuff Post Your Photos! 10 04-21-2009 07:53 AM
For Sale - Sold: K20D System Bargain trumpyman Sold Items 18 04-02-2009 06:22 AM
Pentax K20d flash system rwa Pentax DSLR Discussion 23 04-10-2008 07:20 AM
Does a WLAN and/or GPS system exist for the K20D? f8 Pentax DSLR Discussion 9 03-20-2008 08:58 PM



All times are GMT -7. The time now is 02:02 AM. | 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