Go Back   Steve's Digicams Forums > Digital SLR and Interchangeable Lens Cameras > Pentax / Samsung dSLR, K Mount Mirrorless

Reply
 
Thread Tools Search this Thread
Old Sep 23, 2006, 11:39 PM   #1
Senior Member
 
Join Date: Jan 2006
Posts: 344
Default

Since I have received four emails within the last two days on the SR issue, I opt to update my review to elaborate my observations on the Pentax latest SR system:-

http://www.geocities.com/ricehigh/K1...ll_Review.html

Click on the "SR Comments" to see, I have tried to write in (much) more details for what I know and experienced.

p.s. Warning: If you want to see those typical comparison pairs of shots which "prove" how SR works with "On" and "Off" comparisons, which have being seen everywhere on the net and in numerious magazine tests, you will probably be greatly disappointed. So, please don't read on if you really expect for those things or alike.

RiceHigh
http://www.geocities.com/ricehigh
RiceHigh is offline   Reply With Quote
Sponsored Links
Old Sep 24, 2006, 5:37 AM   #2
Senior Member
 
Join Date: Feb 2004
Posts: 105
Default

I'm finding noticeably more consistent SR performance at slow shutter speeds if I hold a shutter half-press for a second or two after the hand symbol shows in the VF, rather than pressing it as soon as it's ready. Might just be me of course, but I can't see me wobbling any different, so I reckon it's the camera's SR needing time to settle.
SteveB is offline   Reply With Quote
Old Sep 24, 2006, 9:47 AM   #3
Senior Member
 
Join Date: Sep 2005
Posts: 166
Default

Doesnt give a good impression on the SR , does it :?
Chesslanka is offline   Reply With Quote
Old Sep 24, 2006, 9:48 AM   #4
Senior Member
 
Join Date: Jul 2006
Location: Tumbleweed, Arizona
Posts: 1,381
Default

Good morning RiceHigh - First thanks for the evaluation you have provided, as it has assisted me in selecting the Pentax unit I am ordering today. On the topic of image stablization, in particular the implementation used by Pentax, I do agree with you that there are limits, and that the system can be defeated. Just as the image stablization system effects can be amplified in other situations.

As in any system design, its implementation has limitations, and is designed with certain constraints in mind. Image stablization as applied to general photography, across a wide population of users, situations, lighting effects and subject matter - is a very difficult problem to solve. Pentax's solution and implementation, by examining their design, is geared towards the "high jitter" environment as opposed to the "induced slow motion" environment. There are probably situations that can be created to defeat - or have the SR system introduce additional noise into the end product. However, I believe that these will be on the fringe of situations - i.e., pushing the limits of the ability of the image stablization design in which it is expected to work.

At best, any image stablization is intended and designed to increase/expand the environmental envelope in which acceptable images are obtained. There are also plenty of situations that just overwealm the implementation, where by - the system is only able to reduce the effects or just has no effect at all or possibly negative effects.

Your demonstrated testing approach is to apply SR to a more general photographic situation, than to a very narrow situation, and look for its results or effects. However, your attempting to take general images with a large depth of field and compare them against each other - using the human eye. That is difficult, especially when trying to quantify the results.

In my opinion, based on the examples you have provided, a better testing method is to take two images, using the same camera, SR on and SR off. Then to use some specialized image exploitation software. In an image, isolate an area - for example using your city scape panaroma, use the top of a building that is highly defined (set against a dark background) or some easily identified windows - identifiable and selectable in each image. Use a tie point function on a specifically identifiable set of corners (at least 4 tie points - 6 preferable) and tie the two images together, then use various change detection tools (wipe, swipe & flicker) to observe the number of pixels the building top/side or window edge are "blured" across. There are any number of tools out there that are capable of this. However this is taking it to an extreme (, but this is what image analysis is all about).

On the other hand - what really matters is if the user - whom ever (young, old, tall, short, frail, healthy, etc.), is able to turn SR on and get a better image than what they would have, in the given lighting situation and photo opportunity (building, party, landscape, concert, vacation, etc.). In my opinion - I think for what Pentax offers, it does expand the envelope in which better images are able to be taken. That being said, the more knowledgeable the user or what SR does and how it does it, will be able to apply it for the greatest benefit.


interested_observer is offline   Reply With Quote
Old Sep 24, 2006, 10:51 AM   #5
Senior Member
 
ennacac's Avatar
 
Join Date: Mar 2005
Posts: 4,198
Default

Here are two examples that I posted before showing that SR does work for what it was designed for, getting a sharper image when shooting at a shutter speed slower than what wound normally be required for a good image. These were both shot at ISO 3200, hand held, both at the same f/stop and shutter speed and are of the same subject and taken just seconds apart.

ON


Off


Tom
ennacac is offline   Reply With Quote
Old Sep 24, 2006, 12:14 PM   #6
Senior Member
 
Join Date: Feb 2004
Posts: 105
Default

Chesslanka wrote:
Quote:
Doesnt give a good impression on the SR , does it
I find it very useful from 1/8sec and faster, but slower than that it doesn't give guaranteed results, so I'm trying out different things to see if it helps. I took a load of shots a while ago at 1/8sec on my Olympus C750 at wideangle and NONE of them came out anywhere near sharp, so I absolutely know that SR is a great benefit (to me).
SteveB is offline   Reply With Quote
Old Sep 24, 2006, 3:45 PM   #7
Senior Member
 
mtngal's Avatar
 
Join Date: Nov 2005
Location: Frazier Park, CA
Posts: 16,056
Default

I go along with Steve - I know I can't hold a camera steady at 1/8 sec. butI got a bunch of sharp pictures when in Vegas a while ago at that shutter speed.

On the other hand, the SR didn't help enough when I tried to take pictures of a brush fire several miles away with a 300mm lens and I was in a car traveling at 70 mph so I'm more than willing to admit that there is a limit to how much it can do. It's well worth it for me.
mtngal is offline   Reply With Quote
Old Sep 24, 2006, 4:41 PM   #8
Senior Member
 
Join Date: Apr 2004
Posts: 172
Default

What's amazing to me, after reading all of your prose, is that you continue to buy Pentax cameras.
dequardo is offline   Reply With Quote
Old Sep 24, 2006, 5:35 PM   #9
Senior Member
 
danielchtong's Avatar
 
Join Date: Dec 2005
Posts: 2,890
Default

Come on. Nothing can manage in this type of drive-by shooting.

Daniel

mtngal wrote:
Quote:
On the other hand, the SR didn't help enough when I tried to take pictures of a brush fire several miles away with a 300mm lens and I was in a car traveling at 70 mph so I'm more than willing to admit that there is a limit to how much it can do. It's well worth it for me.
danielchtong is offline   Reply With Quote
Old Sep 25, 2006, 7:43 AM   #10
Senior Member
 
Join Date: Jan 2006
Posts: 344
Default

I appreciate your various insights in your reply. Actually, you have mentioned something which I think exactly.

Regarding your tie-point experiment, I think it should be carried out with extreme great care, i.e., nothing physically should be changed and moved between the two SR on an off test shots.

The only disagreement I would have is your mentioning about the DoF which I think is irrelevant for the blur that caused by unstable sensor (movement).

RiceHigh
http://www.geocities.com/ricehigh


interested_observer wrote:
Quote:
Good morning RiceHigh - First thanks for the evaluation you have provided, as it has assisted me in selecting the Pentax unit I am ordering today. On the topic of image stablization, in particular the implementation used by Pentax, I do agree with you that there are limits, and that the system can be defeated. Just as the image stablization system effects can be amplified in other situations.

As in any system design, its implementation has limitations, and is designed with certain constraints in mind. Image stablization as applied to general photography, across a wide population of users, situations, lighting effects and subject matter - is a very difficult problem to solve. Pentax's solution and implementation, by examining their design, is geared towards the "high jitter" environment as opposed to the "induced slow motion" environment. There are probably situations that can be created to defeat - or have the SR system introduce additional noise into the end product. However, I believe that these will be on the fringe of situations - i.e., pushing the limits of the ability of the image stablization design in which it is expected to work.

At best, any image stablization is intended and designed to increase/expand the environmental envelope in which acceptable images are obtained. There are also plenty of situations that just overwealm the implementation, where by - the system is only able to reduce the effects or just has no effect at all or possibly negative effects.

Your demonstrated testing approach is to apply SR to a more general photographic situation, than to a very narrow situation, and look for its results or effects. However, your attempting to take general images with a large depth of field and compare them against each other - using the human eye. That is difficult, especially when trying to quantify the results.

In my opinion, based on the examples you have provided, a better testing method is to take two images, using the same camera, SR on and SR off. Then to use some specialized image exploitation software. In an image, isolate an area - for example using your city scape panaroma, use the top of a building that is highly defined (set against a dark background) or some easily identified windows - identifiable and selectable in each image. Use a tie point function on a specifically identifiable set of corners (at least 4 tie points - 6 preferable) and tie the two images together, then use various change detection tools (wipe, swipe & flicker) to observe the number of pixels the building top/side or window edge are "blured" across. There are any number of tools out there that are capable of this. However this is taking it to an extreme (, but this is what image analysis is all about).

On the other hand - what really matters is if the user - whom ever (young, old, tall, short, frail, healthy, etc.), is able to turn SR on and get a better image than what they would have, in the given lighting situation and photo opportunity (building, party, landscape, concert, vacation, etc.). In my opinion - I think for what Pentax offers, it does expand the envelope in which better images are able to be taken. That being said, the more knowledgeable the user or what SR does and how it does it, will be able to apply it for the greatest benefit.

RiceHigh is offline   Reply With Quote
 
Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off



All times are GMT -5. The time now is 12:46 PM.