Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
02-21-2012, 11:11 AM - 1 Like   #16
Pentaxian
KevinR's Avatar

Join Date: Jan 2010
Location: Cape Town, South Africa
Photos: Gallery | Albums
Posts: 646
Being a gradute mechanical engineer with over 20yrs practical and theoretical experience, the practical benefits of this have me a little rattled. With several years of finite element experience and lots of aircraft/marine/naval motion and design problems, normal 3-axis coordinate systems and 6-DOF motion are very familiar to me. As Lowell has pointed out, this should be described as a "5-degree of freedom" (DOF) rather than "5-axis" compensation system. Our 3-dimensional world space has 3-axes with a translation along each and rotation around each making up the 6 unique DOFs to describe any movement (displacement; velocity; and acceleration).

The normal 3-DOF SR (x & y translation plus rotate about the lens assumed z-axis) as used by Pentax makes sense in that these are the normal movements that will affect the image projection onto the sensor.

The part that rattles me, is that I understand the purpose of a lens to be to project the apparent image onto a flat focal plane. The true focus surface will be approximately a hemisphere of equal distance from the lens optical centre. If the SR includes rotation of the sensor about the x & y-axes (yaw and pitch in aircraft or naval terms), then it would also alter the focal distance of the sensor at the edges... This would, to me, in almost all circumstances be bad, and at best with very close subjects might have some relavence, but the camera would also need to know a few other things before being able to know what is required and make a determination. In this case though, the focal distance and hence focus solution would also change.

Sound like a serious marketing con to me, but I remain open to a more edified explanation.

02-21-2012, 11:24 AM   #17
Pentaxian




Join Date: Mar 2010
Location: Frankfurt am Main
Posts: 1,094
(Edited)
@Quicksand:
QuoteQuote:
"It absolutely is true"
You shouldn't write "absolutely" - I wouldn't call compensating vertical and horizontal rotation errors by vertical and horizontal sensor shifting a clean solution (of course it may help to reduce blurring, but never fully compensate).

QuoteQuote:
"unless you want to defocus your image or induce faux-tilt-lens effects"
Of course in principle you are right - however the effect in real world situations must be proved. Tilt effects may indeed occur, but "defocus"?

@KevinR:
I agree with your description. But whether the Olympus solution is an improvement or not, has to show in real world situations. I myself am not sure. I just think it has cost them too much of development efforts to just name it a marketing thing without seeing test results.

Last edited by RKKS08; 02-21-2012 at 11:38 AM.
02-21-2012, 08:41 PM   #18
Site Supporter




Join Date: Apr 2008
Location: Central Ohio (formerly SF Bay Area)
Posts: 1,491
I'll wade in here one last time, then I'm out. After all, we wouldn't want any misunderstandings to enter the public knowledge base, right?



QuoteOriginally posted by RKKS08 Quote
(Edited)
@Quicksand:
You shouldn't write "absolutely" - I wouldn't call compensating vertical and horizontal rotation errors by vertical and horizontal sensor shifting a clean solution (of course it may help to reduce blurring, but never fully compensate).

Of course in principle you are right - however the effect in real world situations must be proved. Tilt effects may indeed occur, but "defocus"?
Compensating vertical and horizontal camera rotation errors with vertical and horizontal in-plane sensor shifting isn't just a "clean" solution -- it's the ONLY solution possible with sensor movements.

It should be apparent that other types of sensor movements that bring part of the sensor out of its original sensor plane, like rotation about one of the X or Y axes in the sensor plane, would cause focus problems. Like I said, "defocus." The portion of the sensor that tilts toward the lens will be back-focused, and the portion of the sensor that tilts away from the lens will be front-focused.

That would not compensate for any kind of camera movement, unless your lens is mounted to the camera with rubber. Now feel free to develop that idea for commercial use -- I won't claim any IP rights.
02-21-2012, 08:53 PM   #19
Veteran Member
Anvh's Avatar

Join Date: Sep 2011
Posts: 4,616
QuoteOriginally posted by RioRico Quote
I'm anxious to see a 5-axis (5-dimension) SR system. I can grok 4 of those -- x,y,z, and t. What's the 5th axis? Luminosity? Tau+teh? Is the SR system formed as a hypercube?
There are 11 dimensions maybe so we still have some way to go

02-21-2012, 09:01 PM   #20
Veteran Member
Anvh's Avatar

Join Date: Sep 2011
Posts: 4,616
QuoteOriginally posted by RKKS08 Quote
Of course in principle you are right - however the effect in real world situations must be proved. Tilt effects may indeed occur, but "defocus"?
If you've worked with a technical camera you would know the effect.
Tilt effect will have an effect on the focus, if the bottom part move forward then that part is effected the same way like if you move the whole sensor forward, and moving backwards as the same effect off course, you're basic changing the focus.
02-21-2012, 10:48 PM   #21
Veteran Member
Laurentiu Cristofor's Avatar

Join Date: Nov 2009
Location: WA
Photos: Albums
Posts: 3,044
QuoteOriginally posted by Quicksand Quote
Compensating vertical and horizontal camera rotation errors with vertical and horizontal in-plane sensor shifting isn't just a "clean" solution -- it's the ONLY solution possible with sensor movements.

It should be apparent that other types of sensor movements that bring part of the sensor out of its original sensor plane, like rotation about one of the X or Y axes in the sensor plane, would cause focus problems. Like I said, "defocus." The portion of the sensor that tilts toward the lens will be back-focused, and the portion of the sensor that tilts away from the lens will be front-focused.
None of the corrections is perfect. By shifting the sensor around, you can have one side of the image show more vignetting than another. Light can fall more obliquely to a side of the sensor too, creating other issues. Rotation around optical axis is the cleanest of these fixes, but that could also run into problems due to lens and mount designs having rectangular profiles, hence causing other vignetting (though less significant). With the extra movements added by the Olympus system, you will have some slight tilt effects applied besides such vignetting, but such tilting would only matter if you're photographing flat surfaces.

In the end, all these small imperfections will be unnoticeable in real life, so it's a moot point.

From another perspective - having the ability to tilt the sensor could be a great feature on its own if they would allow manual control over it, as Pentax did for framing adjustments, by allowing you to manually rotate the sensor using SR. With such control, you wouldn't need mechanical contraptions - you could just tilt-shift the sensor.
02-22-2012, 05:50 PM   #22
Pentaxian
Lowell Goudge's Avatar

Join Date: Jan 2007
Location: Toronto
Photos: Gallery | Albums
Posts: 15,400
QuoteOriginally posted by Quicksand Quote
I'll wade in here one last time, then I'm out. After all, we wouldn't want any misunderstandings to enter the public knowledge base, right?





Compensating vertical and horizontal camera rotation errors with vertical and horizontal in-plane sensor shifting isn't just a "clean" solution -- it's the ONLY solution possible with sensor movements.

It should be apparent that other types of sensor movements that bring part of the sensor out of its original sensor plane, like rotation about one of the X or Y axes in the sensor plane, would cause focus problems. Like I said, "defocus." The portion of the sensor that tilts toward the lens will be back-focused, and the portion of the sensor that tilts away from the lens will be front-focused.

That would not compensate for any kind of camera movement, unless your lens is mounted to the camera with rubber. Now feel free to develop that idea for commercial use -- I won't claim any IP rights.
I think there are 2 points to separate here, the first is what are the degrees of freedom we have with the sensor. Without a serious modification to the internals, we have 3, lateral vertical and rotational.

Depending on the number of accelerometers we can have either 3 degrees of freedom for sensing, if all the sensors are on the plane of the image sensor, or we can analyze motion for all 6 degrees of freedom, even including lateral movement along the lens axis, but the question you have to ask yourself is what do you plan to do with the data because we don't have sensor shift along the lens axis,

To make al use of the remaing 5 degrees of freedom does not really help unless you really want to try and better correct for yawl and pitch as different than simply x and y motion. Perhaps the motion can be better reduced but this is nothing good technique can't solve.
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!
approach, k-mount, pentax lens, slr lens, sr
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
Pentax K-x mirror slap and/or SR blur problem firefly Pentax DSLR Discussion 100 02-02-2011 07:26 PM
Pentax video samples? (Esp. ~ 300mm and SR) emr Video and Pentax HDSLRs 5 02-28-2010 05:21 AM
Firmware 1.03 K7er Pentax DSLR Discussion 44 02-25-2010 04:07 PM
K-x SR in Cold Weather winglik Pentax DSLR Discussion 7 02-19-2010 12:50 PM
Misc Good old K100d Super and M-28 3.5... sagiboy Post Your Photos! 2 02-19-2010 02:28 AM



All times are GMT -7. The time now is 01:31 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