GX7 IBIS better than EM5 IBIS!

Discussion in 'Panasonic Cameras' started by orfeo, Oct 30, 2013.

  1. orfeo

    orfeo Mu-43 Top Veteran

    673
    Sep 27, 2013
    FR
  2. wjiang

    wjiang Mu-43 Hall of Famer

    I think this may come down to things other than the actual IBIS performance, but nevertheless due to the camera human interface design. The GX7 grip seems a bit better, the shutter button action is different, the EVF position is different (important when it's used as a support point), and possibly differences in shutter shock behaviour. It's great news though, if Panasonic are persevering down the IBIS line as well - we're no longer locked to Olympus if we want awesome stabilisation on our compact lenses!
     
  3. orfeo

    orfeo Mu-43 Top Veteran

    673
    Sep 27, 2013
    FR
    THe GX7 is a great camera with it's built in IBIS and I hope too Panasonic is going to invest further into this in the futur, like in the coming GH4 and next GX cameras... If they could just update the GX7 firmware to have IBIS in video that would be great too...
     
  4. moccaman

    moccaman Mu-43 Veteran

    283
    Jan 4, 2012
    Australia
    Wish I could read his website! :dash2: Had to cut n paste it all, but its certainly very interesting, is the EM5 and EP5 going to perform the same in the IBIS category?
     
  5. jamespetts

    jamespetts Mu-43 Top Veteran

    803
    May 21, 2011
    London, England
    I should be interested to see this test performed with an Olympus camera with the 2 axis stabilisation.
     
  6. wjiang

    wjiang Mu-43 Hall of Famer

    Not sure what would be gained - Olympus are touting their 5-axis system to be the best after all.
     
  7. jamespetts

    jamespetts Mu-43 Top Veteran

    803
    May 21, 2011
    London, England
    It would be useful as a way of calibrating and testing the test itself.
     
    • Like Like x 1
  8. BeyondTheLines

    BeyondTheLines Mu-43 Veteran

    262
    Sep 23, 2012
    Spain/USA
    Patrick
    I'm impressed with the GX7's IBIS, it's an excellent addition to Panasonic's m43 cameras.

    The first results for the EM5 were disappointing so I had to have a go with my EM5 and the 40-150 set to 74mm (I was trying for 75 but was a little off). I took a bunch of shots and was able to get fairly sharp results at 1/2.5 sec but once I went to 1/2 I was getting more noticeable blur. Even so, some are pretty close to sharp, better than what's in the linked blog post for 1/3 sec at 75mm so I guess I don't need a new camera :tongue:.
     
  9. Dramaturg

    Dramaturg Mu-43 Top Veteran

    614
    Jun 7, 2013
    Ukraine
    Yevgen
    I think wjiang is right. I noted many times that when I shoot with long shutter speed and long lenses (let's say with 45mm at 1/20 sec) I get mixed results. Out of four shots of the same steady object two will be razor sharp and two will come out a bit blurry. That's why I always do several takes. I think that has to do with the way I press the shutter release button, with the way I hold the camera or something else on the side of human-camera interface. I am sure if Tyson from the review would take more shots he would get some very clean images out of E-M5 that would outperform GX7. I am simply amazed at the quality of some my handheld shots with E-M5 at 1/3 sec or even slower. Though there are plenty of blury shots as well. That being said, Panasonic seem to get the most of their 2 axis stabilization in combination with physics of the camera. Also the forum members low light sample shots prove that GX7 stabilization is pretty good.
     
  10. meyerweb

    meyerweb Mu-43 Hall of Famer

    Sep 5, 2011
    Tyson did take multiple shots with each camera, but didn't publish them all. He said there was little difference between the shots from each camera.

    That said, IS is a very difficult technology to test objectively. The movement patterns for each photographer and each camera/lens combination are going to be different, and vary depending on environmental conditions (Is it windy? Is the photographer cold and shaking, breathing hard, tired?) No test of 10 or 12 shots is likely to prove anything of a more general manner.

    That said, I've seen quite a few "tests" comparing the gX7 to 5-way IBIS and most show the Panasonic to be within about 1 to 1/2 stop of Oly's 5-way. There was one outlier who said it didn't seem to work at all, but I believe he tried to test by using the kit lens and turning OIS off, but we now know IBIS doesn't work in that situation. And then there's Robichaud's that says the GX7 is superior. I don't doubt that in his tests those are the results he got, but I'm not convinced most users, in most situations, would get the same results.

    To really test it out, you'd want a significant sample of photographers (50? 100? My statistics classes were a long time ago), each shooting a large number of samples with multiple lenses and multiple shutter speeds with each lens. But for practical purposes, you'd never be able to analyze the results without writing an image processing program to do it for you. So to try to be practical say you take 20 shots at each shutter speed from 1/FL to 5 stops slower than that, for 4 different FL lenses, for 10 photographers. That's 4000 exposures, which seems like a large enough sample size to be somewhat meaningful. You'd then need to analyze each individual shot blind (not knowing which lens / camera / photographer it represents) at normal magnifications (if you can see blur at 100%, but not in a 16x20 print, does it really matter?) and decide if it's completely sharp, "adequately sharp," or blurred. And you need to decide how to define "adequate."

    Finally, you'd have to apply statistical analysis to the results to determine if one system is consistently better, better at certain shutter speeds or with certain focal lengths, or if the results are too random to be meaningful. An awful lot of work for any one group of people to accomplish. And even then, YOUR results might not reflect the average.

    And don't forget that if you're shooting a subject that moves, whether it be people, pets, or even outdoor scenes with a breeze, you rapidly reach the point of diminishing returns with any IS system. People aren't very good at holding still. So one stop of IS might be useful for shooting people. 3 or 4 stops certainly won't be. Shooting your kids at 1/15th second with the 25mm might be OK, but 1/2 second probably won't be unless they are sleeping. Sometimes you just need to add more light, or shoot something else. For the types of things I shoot, and the lenses I would want IBIS for, I'm pretty confident both systems would be good enough.
     
    • Like Like x 1
  11. b_rubenstein

    b_rubenstein Mu-43 All-Pro

    Mar 20, 2012
    Melbourne, FL
    Do Panasonic camera users ever get tired of reinforcing the image of themselves, and cameras, as Rodney Dangerfields of the photographic world, or just former Pentax users?
     
  12. orfeo

    orfeo Mu-43 Top Veteran

    673
    Sep 27, 2013
    FR
    When roaming this website I would say the same but for OMD users. This review is just counterbalancing the "Do Olympus camera users ever get tired of reinforcing the image of themselves, and cameras, as Rodney Dangerfields of the photographic world, or just former Pentax users?
     
  13. meyerweb

    meyerweb Mu-43 Hall of Famer

    Sep 5, 2011
    About as quickly as Olympus users get tired of claiming the total superiority of everything Olympus and, by extension, Olympus users. :)
     
    • Like Like x 1
  14. jyc860923

    jyc860923 Mu-43 Hall of Famer

    Feb 28, 2012
    Shenyang, China
    贾一川
    I don't have any 5 axis IBIS models but an EPL5, it clearly shows a lot of double image on those taken during a series of continuous shooting, but it's hard to replicate the results, because the shutter shock happens to certain lens\FL\SS combo
     
  15. Dramaturg

    Dramaturg Mu-43 Top Veteran

    614
    Jun 7, 2013
    Ukraine
    Yevgen
    Interesting, but I just performed a simple test. I took one shot with IBIS on and then another one with IBIS off. Both handheld at 1/15 sec. I couldn't believe the results:

    10578951724_df04027f8f_b.
    IBIS TEST by Yevgen Shatalov, on Flickr

    Then I did several retakes and it was very strange. Some IBIS ON images were as clean as IBIS OFF one, others were worse. I believe Tyson got into the same situation. There are so many factors that actually influence the work of IBIS that it is very hard to compare them. Still wonder why handheld shot was cleaner than the IBIS ON one though.
     

    Attached Files:

    • Like Like x 1
  16. jnewell

    jnewell Mu-43 All-Pro

    Jun 23, 2011
    Boston, MA
    This is likely TRVTH. :smile:
     
  17. Amin Sabet

    Amin Sabet Administrator

    Apr 10, 2009
    Boston, MA (USA)
    Please, drop it!
     
    • Like Like x 12
  18. Reflector

    Reflector Mu-43 Veteran

    406
    Aug 31, 2013
    If anyone wants to send me a GX7 as a "keepe--- I mean "test sample" I will happily go mount my 50 1.2 on a Speedbooster out and give you a few hundred photos between the GX7 and my E-M5 within the 1 second exposure zone and lower. You can sort out all the examples and pick out whichever ones you find more agreeable with your claims. I will happily keep that GX7 with my E-M5 as well, since they're both excellent cameras with excellent sensors.

    Also I recommend realizing the principle of both systems, because the GX7 system is similar to the sensor stabilization found in the PENs, which shifts the sensor up and down in 2 axises. Being able to deflect the sensor off at an angle actually gives a significant advantage given the additional degrees of freedom in the motion.

    For people using _any_ stabilization: Let the system "warm up" for a half a second before you fire off a shot. You are going to feel the VR jerk on a 70-300VR (Definitely one of the heavier VR systems for lens based VR) when it engages and I have no doubt that on my E-M5 if the camera moves, if I wasn't using antishock and fired off a shot immediately, I would probably have the IBIS engage instantly and have the sensor jerk to compensate for small, rapid movement as the exposure starts. I'd also recommend smoothing out the shutter release on the E-M5, the half press click/detent can cause you to twitch the camera.

    For people who need to scream X/Y/Z camera is superior: I've taken several 1/8 second exposure at 213mm effective from my Speedboosted 70-300VR so please crawl back to your hole of specsheet waggling while the rest of us take actual photos.

    And I have to come from Camp Nikon Refugees to post this to the knuckleheads trying to start stupid camera brand/model fights.
     
  19. Rudy

    Rudy Mu-43 Veteran

    449
    Jan 24, 2013
    Oakland, CA
    I have a theory as to why the EM-5 IBIS sometimes does not seem to work as well.

    Basically there is a limited amount of movement the sensor can make to compensate for the user's moving the camera.
    IBIS on the EM-5 is on as soon as you press the shutter half way. This is what gives us the stabilized viewfinder view.
    When you then pan the camera, the sensor movement runs out of room and it jumps back to a more central position. You can actually hear this if you pay attention.

    I can think of two scenarios where this could be problematic:
    If you press the shutter just when such a jump happens, then the sensor would actually introduce the image smear itself by moving. This could be fixed in firmware by delaying the actual shutter release until the sensor is no longer moving. I have never noticed a difference in shutter delay though, so either the sensor can stop very quickly or the firmware is not doing this.

    The other potential problem occurs if you happen to press the shutter while the sensor is near one of its motion extremes as it will have much less adjustment room during the actual exposure. So if you were to continue moving it could run out of space and then the image would be blurred. Of course with IBIS off you get all the motion artifacts, so this should not be any worse unless there is some extra effect with the sensor "banging" into its stops...

    Regarding the GX7 and this is a guess as I don't have one:
    As far as I know the GX7 does not stabilize the image in the viewfinder, but uses IBIS just for the exposure. If that is indeed the case, then it would always start out from the center location of its adjustment range when the shutter is pressed. It would therefore on average be less likely to run out of room, if both mechanisms cover the same range, which again is just an a guess of course. It also would never get into the scenario where it has to center the sensor during framing.
    This could explain why people see mixed results.

    Also of note is that Olympus recommends turning off IBIS when using a tripod. This would seem to be an acknowledgment that IBIS can actually make things worse at least in the case where one is guaranteed to have good support for the camera.
    Rudy
    p.s. Nonetheless, I think we can all agree that being able to handhold fairly long exposure times with either one of these cameras is quite a feat of engineering. At the end of the day it's just another tool and one should know how and when to use it.



     

    Attached Files:

    • Like Like x 3
  20. wjiang

    wjiang Mu-43 Hall of Famer

    You know you can turn off the E-M5 IBIS for half pressed shutter, and only have it engage for final release? I have it configured like this because I don't really need IBIS while composing, so it reduces power consumption.
     
    • Like Like x 1