Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
05-19-2008, 03:35 AM   #1
Pentaxian




Join Date: Sep 2006
Location: Kaunas
Posts: 1,452
K10D SR test (by myself)

Hi,

Many times we read that in body SR is not effective, especially with telelenses, stabilisation in lens is much better etc.

So, yesterday I decided to make DPReview style SR test by myself

DPReview have done such tests with Canon and Nikon stabilised lenses on 35mm camera at 200mm therefore I did such testing at 135mm on my 1.5 cropped K10D.

Here are my results:




So, in camera SR gave me exactly 2 extra stops.

Canon 70-200mm f/2,8 for DPReview testers gave 2,5 stops and Nikkor 70-200mm f/2.8 gave 1,5 stops.

So, after all K10D performed a little bit worse than Canon lens and a little bit better than Nikkor lens.

What do you think?

--
Edvinas

05-19-2008, 03:53 AM   #2
Veteran Member
ftpaddict's Avatar

Join Date: Jan 2008
Location: Yurp
Photos: Gallery
Posts: 4,669
All in all, pretty accurate.

I've found I can hand-hold my K100D & kit lens @ 50mm for about 1/2 sec. and still get a sharp(ish) picture.
05-19-2008, 06:36 AM   #3
New Member




Join Date: Apr 2008
Location: Cambridge, UK
Posts: 9
Yes, a good result is very much about the technique of holding a camera steady. Therefore, it's almost impossible to produce accurate tests as you'd need to put the camera on a tripod, which would, um, kind of defeat the object.

I, too, can get acceptably sharp shots with a handheld 1/2 sec exposure using SR. This, IMO, is one of the great advances of photographic technology.
05-19-2008, 07:20 AM   #4
Veteran Member




Join Date: Mar 2008
Location: CT / NY
Photos: Gallery
Posts: 822
Whats the population of the data?

05-19-2008, 07:30 AM   #5
Pentaxian




Join Date: Sep 2006
Location: Kaunas
Posts: 1,452
Original Poster
QuoteOriginally posted by BBear Quote
Whats the population of the data?
What do you mean? Sorry, English isn't my native language.
05-19-2008, 10:37 AM   #6
Loyal Site Supporter
Canada_Rockies's Avatar

Join Date: Nov 2007
Location: Sparwood, BC, Canada
Photos: Gallery | Albums
Posts: 9,547
QuoteOriginally posted by Edvinas Quote
What do you mean? Sorry, English isn't my native language.
BBear was talking about the number of tests you ran - population of data is a term used in statistical analysis to indicate the probable accuracy of the sample used. More population, less chance of false returns.

Stupid example, coin flip, heads or tails. If the population is 1, the result is absolutely meaningless, statistically. The more times you flip the coin, the more chance there is that your results will converge on 50% heads, 50% tails.

Hope this helps.
05-19-2008, 10:45 AM   #7
Veteran Member




Join Date: Mar 2008
Location: CT / NY
Photos: Gallery
Posts: 822
QuoteOriginally posted by Canada_Rockies Quote
BBear was talking about the number of tests you ran

Yep. I ask this because obviously you are not posting the pictures and the categories [Sharp / Mild Blur / Heavy Blur / Very Heavy blur] are very subjective, so the more data you have the more accurate the results.
05-19-2008, 10:51 AM   #8
JMS
Senior Member




Join Date: Oct 2007
Location: Wisconsin
Posts: 273
Give me the complete data and I'll run an appropriate statistical test.

05-19-2008, 12:56 PM   #9
Pentaxian




Join Date: Sep 2006
Location: Kaunas
Posts: 1,452
Original Poster
I did this simple test the same way as dpreview do: 10 pictures with SR off, 10 pictures SR on. Then one f-stop slower shutter speed and again 10 picures with SR off and 10 pictures with SR on. etc. In total I have made 100 pictures

Picture rating was (all samples are 100% crops):

sharp - picture is absolutely sharp viewing at 100%




mild blur - picture seems sharp when viewed resized to fit my screen, however viewing at 100% reveals small camera shake




Heavy blur - picture seems not sharp when viewed resized to fit my monitor




Very heavy blur - completely washed out picture.


Last edited by Edvinas; 05-19-2008 at 01:10 PM.
05-19-2008, 01:38 PM   #10
Ash
Community Manager
Loyal Site Supporter
Ash's Avatar

Join Date: Mar 2007
Location: Toowoomba, Queensland
Photos: Gallery | Albums
Posts: 22,695
BBear's right.
There'll be an element of subjective interpretation on how blurry a picture is, but is still reasonably reliable (for the same person analysing the data, preferably with blinding of the EXIF data when interpreting to prevent user bias).

Then there'll need to be an adequate number of tests taken to consistently confirm the improvement of sharpness of images with SR on, enough to have the 'statistical power' to prove the difference 'beyond reasonable doubt' (i.e. with a 95% confidence level).

Apologies if this is a whole lot of mumbo-jumbo...
This is a good test by Edvinas, and can be made more scientific with the right conditions and definitions.
05-19-2008, 02:04 PM   #11
Veteran Member




Join Date: Mar 2008
Location: CT / NY
Photos: Gallery
Posts: 822
QuoteOriginally posted by Ash Quote
This is a good test by Edvinas, and can be made more scientific with the right conditions and definitions.

Absolutely. It is very easy to bias the results in this kinda of experimentation, because is very subjective and even the test procedures are hard to reproduce (one time you may feel tired and shake more than others and so on).


But luckily, these results are valid for the person that matters the most: you.

You definitely have a great grasp of the capabilities of the SR in your camera and that should be satisfactory for you.


Hecka, anyone interested in the results should actually try the same procedures and come up with their own results!



Great job man!
05-20-2008, 07:35 AM   #12
JMS
Senior Member




Join Date: Oct 2007
Location: Wisconsin
Posts: 273
I'm actually working on an objective method of measuring blur, and I'll post the results when I'm done.
05-20-2008, 03:39 PM   #13
Veteran Member




Join Date: Feb 2008
Location: Chicagoland
Posts: 338
QuoteOriginally posted by Ash Quote
BBear's right.
There'll be an element of subjective interpretation on how blurry a picture is, but is still reasonably reliable (for the same person analysing the data, preferably with blinding of the EXIF data when interpreting to prevent user bias).
Ideally you'd have someone take the pictures who has no idea what the settings are just to prevent trying to hold steady more for the SR or relaxing when its on.
05-20-2008, 06:18 PM   #14
Pentaxian
Arpe's Avatar

Join Date: Sep 2006
Location: New Zealand
Photos: Gallery | Albums
Posts: 4,344
QuoteOriginally posted by JMS Quote
I'm actually working on an objective method of measuring blur, and I'll post the results when I'm done.
That would be relatively easy, no? On the face of it anyhow. Just count the numbe rof pixels a feature has moved on the photo. Sounds like fun
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!
70-200mm, bit, camera, canon, dpreview, dslr, k10d, lens, nikkor, photography, sr, test
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
First test of my new K10D Thumper473 Post Your Photos! 2 02-04-2009 04:27 AM
First test with 1000mm K10D Peter Zack Pentax SLR Lens Discussion 49 07-25-2008 06:32 AM
K10D and Sigma 50-500 test Heinrich Lohmann Post Your Photos! 12 12-08-2007 03:46 PM
test shots/new to K10D jfg69 Post Your Photos! 3 10-11-2007 10:21 PM
K10D first test of SR Duh_Vinci Pentax DSLR Discussion 7 12-01-2006 02:46 PM



All times are GMT -7. The time now is 03:16 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