Go Back   Steve's Digicams Forums > Digicam Help > General Discussion

Reply
 
Thread Tools Search this Thread
Old Feb 11, 2014, 7:11 AM   #11
Senior Member
 
ItDontMeanAThing's Avatar
 
Join Date: Oct 2010
Location: Hua Hin, Thailand
Posts: 403
Default

Quote:
Originally Posted by TCav View Post
You can define an "Outlier" anyway you want, but my preference is that, for a very small sample size, an outlier is outside two standard deviations from the mean. If you're only going to test three times, it's unlikely that you'd get any outliers.

And since your objective is ...... any results you get are significant because you got them, and shouldn't be discarded.
Outliers will be identified by the Eyeball Test. Sample size is tbd. For each way of holding the camer, I'll start with n=1 to get a ballpark shutter speed. then I'll test double and half of the ballpark with a bigger sample and narrower spacing of shutter speeds. Results of the second test will determine the need for a 3rd test.

Outliers will be ignored because there's a +/- tolerance for determining the slowest non-blur setting. My solution isn't rocket science. When working hand held in low light I shoot a sequence, then pick the best one. If the shot matters, I switch from hand held while standing to a more stable way of shooting.

Last edited by ItDontMeanAThing; Feb 11, 2014 at 7:13 AM.
ItDontMeanAThing is offline   Reply With Quote
Old Feb 11, 2014, 8:02 AM   #12
Senior Member
 
TCav's Avatar
 
Join Date: Sep 2005
Location: Washington, DC, Metro Area, Maryland
Posts: 13,572
Default

Quote:
Originally Posted by ItDontMeanAThing View Post
Both my education and career in data analysis say ignore the outliers and average the rest, assuming there is a central tendency. ...
Quote:
Originally Posted by ItDontMeanAThing View Post
Outliers will be identified by the Eyeball Test. ...
If that's what your education and career in data analysis say to do, but to me that sounds like an opportunity for Selection Bias, either intentional or not.

If you go by the average, then you've got a 50% chance of getting motion blur due to camera shake anyway. If you go with the average plus one standard deviation, then you've got ~15% chance of getting motion blur due to camera shake. If you toss out the "outliers", those numbers get a lot worse.
__________________
  • The lens is the thing.
  • 'Full Frame' is the new 'Medium Format'.
  • "One good test is worth a thousand expert opinions." - Tex Johnston, Boeing 707 test pilot.

Last edited by TCav; Feb 11, 2014 at 8:08 AM.
TCav is offline   Reply With Quote
Old Feb 11, 2014, 10:04 AM   #13
Senior Member
 
VTphotog's Avatar
 
Join Date: Mar 2005
Location: Extreme Northeastern Vermont, USA
Posts: 4,234
Default

Quote:
Originally Posted by ItDontMeanAThing View Post
You were perfectly clear. My response said your contribution had nothing to do with what I requested in post #1. Was I not clear?
I was just responding to your post title, which asked for a critique of your method. Which is what you asked for and got.
Sorry you didn't like it.

brian
VTphotog is offline   Reply With Quote
Old Feb 11, 2014, 1:07 PM   #14
Senior Member
 
TCav's Avatar
 
Join Date: Sep 2005
Location: Washington, DC, Metro Area, Maryland
Posts: 13,572
Default

The unit of measure you use can adversely affect your results.

For instance, suppose you run three tests, and get results of 1/30, 1/60, and 1/125. You may be inclined to call the 1/125 an outlier. Even if you don't, you may be inclined to calculate an average of 1/72 (or 1/80, to keep the result on the conventional scale for shutter speeds.) That would be incorrect.

You should use the Exposure Value (Ev), or "stop", instead. Presuming that ...
Quote:
Originally Posted by ItDontMeanAThing View Post
... I'll start with n=1 to get a ballpark shutter speed ...
... then 1/30 = 6 stops, 1/60 = 7 stops, and 1/125 = 8 stops. Therefore, there are no outliers, and the average is 7 stops, or 1/60.
__________________
  • The lens is the thing.
  • 'Full Frame' is the new 'Medium Format'.
  • "One good test is worth a thousand expert opinions." - Tex Johnston, Boeing 707 test pilot.

Last edited by TCav; Feb 11, 2014 at 1:11 PM.
TCav is offline   Reply With Quote
Old Feb 11, 2014, 1:08 PM   #15
Senior Member
 
TCav's Avatar
 
Join Date: Sep 2005
Location: Washington, DC, Metro Area, Maryland
Posts: 13,572
Default

Quote:
Originally Posted by ItDontMeanAThing View Post
... Results of the second test will determine the need for a 3rd test. ...
BTW, that's blatant selection bias.

Just so you know.
__________________
  • The lens is the thing.
  • 'Full Frame' is the new 'Medium Format'.
  • "One good test is worth a thousand expert opinions." - Tex Johnston, Boeing 707 test pilot.

Last edited by TCav; Feb 11, 2014 at 1:10 PM.
TCav is offline   Reply With Quote
Old Feb 12, 2014, 4:38 AM   #16
Senior Member
 
ItDontMeanAThing's Avatar
 
Join Date: Oct 2010
Location: Hua Hin, Thailand
Posts: 403
Default

Quote:
Originally Posted by TCav View Post
If you go by the average, then you've got a 50% chance of getting motion blur due to camera shake anyway.
Say what? Show your work.
ItDontMeanAThing is offline   Reply With Quote
Old Feb 12, 2014, 7:17 AM   #17
Senior Member
 
TCav's Avatar
 
Join Date: Sep 2005
Location: Washington, DC, Metro Area, Maryland
Posts: 13,572
Default

If you use the average as your dividing line, half the time you'll be above the average (no motion blur) and half the time you'll be below the average (motion blur.) That's what "average" means. That's what happens with a normal distribution.
__________________
  • The lens is the thing.
  • 'Full Frame' is the new 'Medium Format'.
  • "One good test is worth a thousand expert opinions." - Tex Johnston, Boeing 707 test pilot.
TCav is offline   Reply With Quote
Old Feb 12, 2014, 9:24 AM   #18
Senior Member
 
ItDontMeanAThing's Avatar
 
Join Date: Oct 2010
Location: Hua Hin, Thailand
Posts: 403
Default

Quote:
Originally Posted by TCav View Post
If you use the average as your dividing line, half the time you'll be above the average (no motion blur) and half the time you'll be below the average (motion blur.) That's what "average" means. That's what happens with a normal distribution.
Ah ... I see what you're doing, but ...

Quote:
Originally Posted by ItDontMeanAThing View Post
Say what? Show your work.
I wrote that in response to post #12 because I missed post #14. Had I read #14 I would ignored both #12 and #14 because we're not talking about the same thing.
ItDontMeanAThing is offline   Reply With Quote
Old Feb 12, 2014, 10:30 AM   #19
Senior Member
 
TCav's Avatar
 
Join Date: Sep 2005
Location: Washington, DC, Metro Area, Maryland
Posts: 13,572
Default

Quote:
Originally Posted by ItDontMeanAThing View Post
I wrote that in response to post #12 because I missed post #14. Had I read #14 I would ignored both #12 and #14 ...
Posts #12 and #14 are not interdependent. Each stands on its own.

Quote:
Originally Posted by ItDontMeanAThing View Post
... because we're not talking about the same thing.
I'm talking about "average", as in the Arithmetic Mean (the sum of the values divided by the number of values.)

Since you say ...
Quote:
Originally Posted by ItDontMeanAThing View Post
... Results of the second test will determine the need for a 3rd test. ...
... I was pretty certain that you weren't talking about the Median (the middle value of a group of values) or the Mode (the most frequently occurring value in a group of values), since both would certainly require at least 3 tests.
__________________
  • The lens is the thing.
  • 'Full Frame' is the new 'Medium Format'.
  • "One good test is worth a thousand expert opinions." - Tex Johnston, Boeing 707 test pilot.
TCav is offline   Reply With Quote
Old Feb 12, 2014, 9:18 PM   #20
Senior Member
 
ItDontMeanAThing's Avatar
 
Join Date: Oct 2010
Location: Hua Hin, Thailand
Posts: 403
Default

Quote:
Originally Posted by VTphotog View Post
I was just responding to your post title, which asked for a critique of your method. Which is what you asked for and got.
Sorry you didn't like it.

brian
Imagine how useful it might be if you read the post instead of just the subject. Then you would have learned what I was asking:
Quote:
Originally Posted by ItDontMeanAThing View Post
Two questions. I identified 2 shutter speeds as the slowest ones without blur, 1/125 for IS off and 1/50 for IS on. First, did I identify the correct speeds? Between 59 year old eyes and a laptop screen, I need a second opinion. Next, are there any parts of the test procedure can be improved to make judging the results more obvious? Just to be clear, I'm interested in testing methods and interpreting results, not in different ways to hold a camera.
I'm a specific guy in a tl;dr world. So it goes.
ItDontMeanAThing 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 11:42 AM.