1. Reminder: Please use our affiliate links for holiday shopping!

OIS effects on bokeh - high shutter speed

Discussion in 'Panasonic Cameras' started by shnitz, Oct 14, 2011.

  1. shnitz

    shnitz Mu-43 Top Veteran

    989
    Aug 25, 2011
    Austin, TX
    There have been a few threads about this, so while I was doing a tripod/OIS test, I decided to do a snapshot OIS and high shutter speed test, to illustrate why I leave it off if I'm in good light. You can get worse or better results than this, depending on your background and when you "catch" the image stabilization system doing its job. I see a difference, so I leave OIS off unless I need it. But, I try to turn on a light, or use stabilization first.

    Photos taken handheld. Cropped and "auto tone" adjusted in Lightroom. First photo is with OIS off; 2nd photo has it turned on.

    OIS_Bokeh_OFF. OIS_Bokeh_ON.
     
  2. Christopher

    Christopher Mu-43 Regular

    72
    Oct 8, 2011
    Toronto
    hm interesting... i never was a believer in in camera/lens stabilization.. id rather just bump iso up a bit :)

    what were your aperture and shutter speeds set at in relation to focal length?
     
  3. shnitz

    shnitz Mu-43 Top Veteran

    989
    Aug 25, 2011
    Austin, TX
    EXIF has been preserved in the Lightroom JPEG exports.

    OIS Off, photo 1:
    f/9
    1/100 second
    42mm
    ISO 800
    OIS On, photo 2:
    f/10
    1/100 second
    42mm
    ISO 800
     
  4. yottavirus

    yottavirus Mu-43 Regular

    115
    Aug 13, 2011
    If the aperture is different, you cannot call this a scientific test, especially when aperture is a major influence on bokeh.

    You can't say whether the difference you see is because of the difference in aperture or ois.
     
  5. kevinparis

    kevinparis Cantankerous Scotsman

    Feb 12, 2010
    Gent, Belgium
    I thought the recommendation was that if you are shooting on a tripod you should switch off image stabilisation

    just saying

    K
     
  6. Promit

    Promit Mu-43 All-Pro

    Jun 6, 2011
    Baltimore, MD
    Promit Roy
    Yeah, I'm really not seeing your point here.
     
  7. yottavirus

    yottavirus Mu-43 Regular

    115
    Aug 13, 2011
    The hypothesis was that leaving stabilisation on at adequate shutter speeds would visibly affect the quality of the bokeh. However this test is invalid as a variable that would affect the quality of bokeh was not kept constant. There's nothing to stop anyone from saying that the difference in bokeh is because of the difference in aperture.

    Theoretically there should be no downside to having stabilisation on at all times, especially in this situation.

    If you read, the shots were hand held. It is nothing about tripods.
     
  8. shnitz

    shnitz Mu-43 Top Veteran

    989
    Aug 25, 2011
    Austin, TX
    It's 1/3 of a stop, which shouldn't change the depth of field enough to notice at that distance, especially with the subject less than 2 feet away, and the tree over 15, but I can do it again in manual mode; something more exact.

     
  9. shnitz

    shnitz Mu-43 Top Veteran

    989
    Aug 25, 2011
    Austin, TX
    OK, back with a more scientifically valid test.
    1/250s
    ISO 100
    42mm
    f/5.6
    Set my focus point towards the bottom right, framed so that the cross was on the railing, just above the vertical beam to the right of . Exposed more for the sky, which explains the darkness of my foreground.

    No Lightroom adjustments, save for cropping in the second sets of images, and resizing to 600pixels along the short edge.

    Originals, so that you can judge the integrity of my framing:
    IS OFF
    Bokeh_Test_1-250_Originals_1_of_2_.

    IS mode 1
    Bokeh_Test_1-250_Originals_2_of_2_.


    And now for crops, towards the top left of the frame. I see clear indication that image stabilization is hurting the image.
    IS OFF
    Bokeh_Test_1-250_1_of_2_.
    IS mode 1
    Bokeh_Test_1-250_2_of_2_.


    Last night, I took a few other snapshots; I can show them as well if you really want, but I see a definite loss of acuity with image stabilization on, even at 1/160s. I disagree that there is theoretically no downside to having IS on. Rather, I would assume even without testing that there is degradation because there is a loose element in the lens, which could be thought of as akin to loose tolerances.

    Not long ago, I posted in this thread:
    https://www.mu-43.com/f38/u43-sports-oh-no-never-17280/
    I critiqued that the bokeh overall wasn't amazing, but specifically in numbers 2 and 4 were the worst offenders. I would venture that the busy-ness in the background is due to OIS.
     
  10. Sammyboy

    Sammyboy m43 Pro

    Oct 26, 2010
    Steeler Country
    From these photos, I'm seeing just the opposite. There is less blur in the tree branches with IS mode 1 than with IS off.
     
  11. yottavirus

    yottavirus Mu-43 Regular

    115
    Aug 13, 2011
    Fair enough, I didn't think of it like that. In theory I thought it would match movements perfectly, or close to it.

    I agree with your conclusions. On a lens with good bokeh, I would expect to see the effects magnified. I don't think that less acuity is the word as I think that IS off is more blurred, which is desirable.

    Sorry for saying this again, but I still think that the test can be improved by using a static scene to eliminate all variables in the subject matter, as well as a tripod to remove differences in framing. At these shutter speeds, there should be no difference between a steady hand and a tripod.

    @above: I see that as well. If we call the non IS the control, then it can be said that IS reduces the smoothness in the blurring of out of focus areas. In my definition of bokeh, I call it a negative.

    However I would say that IS would have a significant positive impact on my shooting experience, and therefore I would prefer to have it. I agree that it should be turned off when not required, as bokeh is quite difficult to improve or correct in post processing.
     
  12. I think that is the crux of the argument.
     
  13. hkpzee

    hkpzee Mu-43 All-Pro

    Sep 5, 2011
    Hong Kong
    Patrick
    My eyes are going dizzy starting at the 2 blurry photos of tree branches, but I agree with you that there seems to be less blur in the tree branches with IS mode 1 than with IS off (Which seems to imply that the IS is working properly)

    If blurriness = good bokeh, then it appears IS off produces better bokeh, but could the blurriness in the "test" shots be caused by handshake, however minute it is? Were these shots taken with a delayed shutter release? :confused:
     
  14. yottavirus

    yottavirus Mu-43 Regular

    115
    Aug 13, 2011
    A shutter speed of 4 milliseconds, combined with a sway distance of about 2-3 millimetres traveled in a few seconds, will not allow any noticeable motion blur. You would see it in the foreground as well.
     
  15. shnitz

    shnitz Mu-43 Top Veteran

    989
    Aug 25, 2011
    Austin, TX
    I'm sorry, I was unclear. Let me re-state my conclusions in a clear, concise way, so that we're all on the same page.

    1. With regards to bokeh, image stabilization hurts the image. It makes the bokeh look "cluttered" and "busy," to use terms that I've seen online. On this, both of us agree.


    2. With regards to the subject, image stabilization hurts the acuity. I mentioned this conclusion, but I did not post any examples, so my earlier post could have been misconstrued. Here are the examples. Newspaper is excellent for testing image quality. Notice in the second shot, with IS on, how the text inside of the depth of field is visibly less sharp.

    IS OFF
    Bokeh_Test_1-160_Subject_Acuity_IS_Off.
    IS ON
    Bokeh_Test_1-160_Subject_Acuity_IS_On.

    I feel I've done a pretty good job with the data. It's not something I would submit to a peer-reviewed journal, but there was no wind, so the scene was relatively still. I think that it's just fine, and it seems to coincide with what I've read online, and my own previous experience with my photos. Anyone of course is free to do any of these tests for themselves. All you need is a scene of your choosing. I wasn't out to prove things to others; I was just doing a test for myself to make sure that I wasn't talking out my ass in other threads, and I decided to share the results with you all, in case you feel like drawing your own conclusions.

    I feel I should reiterate my position however, from those previous threads. Do I think IS is a useless feature? Absolutely not. It is a very useful feature, it has its place, and I'm very glad that my camera has it. I will definitely be using IS when I need to. The misunderstandings came from me talking bad about it. While it's nice, I don't think that it's an end-all, be-all feature. People in the other threads were talking about how they would buy one camera over another, strictly because of the IS available. While I've found it useful, I definitely don't think it's worth discounting cameras like the GH2, for example, if you like everything about the camera except for not having IBIS. If you look at all the specs of a camera, and your make-it-or-break-it decision is due to that, then I think you should read some more photography texts. Heck, I'd trade in all my photography gear today if someone was willing to trade me their Leica M9 with lenses, and gladly give up IS, autofocus, video, etc. I'll even settle for your Pentax 645D!

    So, when will I turn on image stabilization? When I'm taking pictures of unmoving dark scenes, I am unwilling or unable to turn up the ISO any more, and my shutter speed is dropping to unreasonable levels. If I can simply use a tripod or increase the light, I will try to do either or both of those things first. One of the previous threads, for example, was someone trying to take pictures of his friends sitting around a table. At that point, I'd rather throw on a Lumopro LP160, set the flash to 1/2 output, and bounce it off the ceiling.

    No, it is implying that IS is cluttering up the bokeh, because of parallax error.

    There was no visible handshake in my photos. I have been doing photography for a long time, with a lot of different cameras, both lighter and heavier than this one. For this test, I was on my knees, with a spot marked on the ground. I placed a chair to my right, so that I could make sure to line my shoulders up with its back, to decrease shot variance. I took this shot at 1/250s, at 42mm, using the same stable technique I've learned over the years.
     
  16. krugorg

    krugorg Mu-43 Hall of Famer

    Jul 18, 2011
    Minnesota USA
    Okay, I only barely made it through high school physics, but it would seem to me that hand shake (assume very minor amount) would result in some motion blur that would be more pronounced on objects farther away from the camera.
     
  17. shnitz

    shnitz Mu-43 Top Veteran

    989
    Aug 25, 2011
    Austin, TX
    First, that bokeh doesn't look like shake, it looks like a difference in definition. Shake would mean warpage. But, to assuage your fears, I saw some cranes while driving about today, so I took some photos. If my camera technique would cause branches 15 feet away to be blurred at 42mm and 1/250s, then what about this photo?
    1-250_Shake_Test.

    1-250_Shake_Test_closeup1.

    1/250s is well within the limits of handholding with a kit lens. I probably could have been walking briskly and gotten a similarly sharp image! So, I think we can rule out camera shake.
     
  18. zettapixel

    zettapixel Mu-43 Veteran

    470
    Aug 12, 2010
    NY
    Thanks for bringing it to our attention, the results do look pretty conclusive (although I'm sure we're all going to do some tests out of curiosity).

    How exactly does it cause this effect? Does IS move the sensor (while it shouldn't) during this minute (put intended) interval? If so, why wouldn't it look as a minor motion blur?