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

Wait a minute... GX85 is only 3-axis IBIS??

Discussion in 'Micro 4/3 News and Rumors' started by agentlossing, Apr 11, 2016.

  1. agentlossing

    agentlossing Mu-43 Hall of Famer

    Jun 26, 2013
    Andrew Lossing
    Maybe I misread the initial reviews and tests, but at least according to the pre-order page on Amazon, the IBIS portion of the camera's stabilization is not 5 but 3 axis... Is this true?

    Color me kinda disappointed.

    Edit: looks like this was a false alarm.
     
    Last edited: Apr 11, 2016
  2. tkbslc

    tkbslc Mu-43 Hall of Famer

    GX7 is 2-axis, so it could be worse!

    The marketing says 5-axis is achieved with both lens and IBIS working together.
     
  3. agentlossing

    agentlossing Mu-43 Hall of Famer

    Jun 26, 2013
    Andrew Lossing
    Well, upon further investigation, the specs on Panny's website clearly says 5-axis sensor shift. So I guess that's clear? I wonder why the confusion?
     
  4. Turbofrog

    Turbofrog Mu-43 Hall of Famer

    Mar 21, 2014
    Taking technical information from an Amazon page? Come now, you should know better by now!

    Both Imaging Resource and DPReview clearly state that it is 5-axis IBIS, up from the 4-axis IBIS of the GX8. In fact, DPReview actually corrected their initial article stating 3-axis (+2-axis lens IS for 5-axis) to say 5-axis IBIS. They are much more trustworthy sources than Amazon.
     
    • Agree Agree x 1
  5. tkbslc

    tkbslc Mu-43 Hall of Famer

    That must be where I got the idea it was 3+2 instead of 5 in body.
     
  6. GFFPhoto

    GFFPhoto Mu-43 All-Pro

    Feb 24, 2013
    5 axis IBIS working with the lens IS might put this on my list...
     
  7. SVQuant

    SVQuant Mu-43 Top Veteran

    845
    Sep 20, 2015
    SF Bay Area, California, USA
    Sameer
    Amazon page says:
    • New 5-axis in-body sensor stabilizer works with LUMIX 2-axis optically stabilized lenses in video and still capture modes using class-leading Dual Image Stabilization.
    I would be very interested in a kit with the P15 or a (refreshed?) P20.
     
    • Like Like x 1
  8. agentlossing

    agentlossing Mu-43 Hall of Famer

    Jun 26, 2013
    Andrew Lossing
    Hubbub averted. Amazon should get their stuff together. Hahah.
     
    • Like Like x 1
  9. tkbslc

    tkbslc Mu-43 Hall of Famer

    Well that's good, but I am continually surprised at how much some worry about stabilization. I mean 2-3 stops and you should be good, right?
     
    • Like Like x 1
  10. oldracer

    oldracer Mu-43 All-Pro

    Oct 1, 2010
    USA
    The Japanese "Camera & Imaging Products Association" (CIPA) has developed a camera/lens stabilization testing protocol that attempts to be an objective measure. It's pretty impressive. Shaker table, actual field test data, .. the whole deal. Check out CIPA DC-011 Measurement and Description Method for Image Stabilization Performance of Digital Cameras: Home. Here is what they have to say about vibration:

    ----------------------------------------------------------------------------
    6.1 Vibration Waveform [for camera testing]

    "... Camera shake has components in six directions (Yaw, Pitch, Roll, X, Y, Z) in total. However, when the shooting distance is about 20 times the focal length, Yaw and Pitch are dominant and the other directions can be practically negligible. Accordingly, the waveform only with two axes is adopted."
    ----------------------------------------------------------------------------

    So, based on extensive testing (" ... frequencies and amplitudes of camera shake of many people were acquired by a sensor attached to the camera ... ") their conclusion was that two axis camera stabilization is all that is necessary.

    As importantly, since the only standard test spec. tests only two axes, any manufacturer's claims for improved performance via stabilization of additional axes cannot be objectively confirmed and are potentially bogus.

    Also, since independent camera testers and reviewers do not have access to any standardized testing apparatus at all, their evaluations have no real chance for repeatability or comparability even when evaluating two-axis systems.

    Personally, I think all of this multiaxis stuff is marketing blather and all of the reviews and comments about the subject are highly subject to confirmation bias (Confirmation bias - Wikipedia, the free encyclopedia).
     
    Last edited: Apr 11, 2016
    • Agree Agree x 3
    • Informative Informative x 1
  11. tkbslc

    tkbslc Mu-43 Hall of Famer

    Yeah, I really can't see how twisting or tilting of the camera (the other 3 axes) would be a big issue unless you are trying to take pictures with a heavy lens using only one hand while also roller skating across the deck of a ship sailing through a storm.
     
    • Funny Funny x 1
  12. gr6825

    gr6825 Mu-43 Veteran

    277
    Oct 10, 2012
    GX85 is a budget version of the GX8, right? So I would expect it to have downgraded specs in some areas.
     
  13. agentlossing

    agentlossing Mu-43 Hall of Famer

    Jun 26, 2013
    Andrew Lossing
    Except that in lots of ways, it does the opposite by adding exciting new tech, which is why it's on so many people's radar already.
     
  14. Ross the fiddler

    Ross the fiddler Mu-43 Hall of Famer

    It's only 'blather' to those that don't want to believe those that own, use & appreciate the benefits of the multi axis IS.

    I'll try & explain why I believe that. First off, everyone that holds a camera is different to someone else holding it with the exact hold & body/arm movement & each scenario changes with it. First off with camera movements, the body of the camera may move in three different directions & can be all at once; up/down, side ways &/or rotates (roll) with the pressing of the shutter button. That's just the body movements put simply. The next is how the lens may be held, especially the longer lenses. If it is held further from the body then that may add further complicated movements (in an attempt to stabilise the gear) & the movement sensors can measure that & hence combining all movements there are 5 different axis that need to be compensated for & that is actually done by moving the sensor in three different directions (not 5). When Dual IS (sensor & optical) from Panasonic & Sync IS from Olympus (for the 300mm f4 lens) was added then the resulting compensation became even better.

    I hope that helps.
     
    Last edited: Apr 11, 2016
    • Agree Agree x 1
  15. Turbofrog

    Turbofrog Mu-43 Hall of Famer

    Mar 21, 2014
    Do you believe that it was just a coincidence that Olympus' image-stabilization made a huge leap in performance when it went from 2-axis in the older PENs and 4/3s bodies to the E-M5? Or that Sony's 5-axis IBIS is nearly competitive with Olympus, whereas earlier 2-axis Minolta designs were not nearly as good? Or that Panasonic has demonstrably improved the stabilization in its newer models compared to the GX7 and happened to add additional axes of compensation?

    Newer systems obviously take advantage of more advanced gyroscopes and more sophisticated processors, and you may indeed be correct, that's just a pretty large storm of coincidences that seem to indicate that as manufacturers get better at developing image stabilization techniques, they add more axes to their system.
     
    • Like Like x 1
  16. Turbofrog

    Turbofrog Mu-43 Hall of Famer

    Mar 21, 2014
    Also of note, Pentax has moved to 5-axis with the system in their K1. Who wants to bet it will perform better than their earlier models?
     
    • Like Like x 1
  17. oldracer

    oldracer Mu-43 All-Pro

    Oct 1, 2010
    USA
    @Ross@Ross, @Turbofrog@Turbofrog, essentially you are saying that the CIPA test results are invalid, that movement in axes other than pitch and yaw is not "negligible." Fine. I was trained as a scientist and an engineer and am always interested in seeing new data. Please provide a description or a link to experimental data that would call the CIPA experimental results into question.

    Regarding claimed improvements for multi-axis systems, again I am interested in your data. Please provide a description or a link to shaker table or other objectively repeatable tests, together with experimental data that supports the validity of the waveforms being used for testing.

    Please understand that I am unimpressed with subjective phrases like "demonstrably improved." How demonstrated? By whom? Using what repeatable experimental protocol?

    I get it that you guys are true believers. But belief, no matter how fervent or widely held, cannot make something into a fact.
     
    • Agree Agree x 1
  18. oldracer

    oldracer Mu-43 All-Pro

    Oct 1, 2010
    USA
    Sorry, I missed responding to this. The proliferation of multi-axis stabilization can be as easily be understood as a marketing response. If the market is convinced that it wants multi-axis stabilization, fallaciously or not, vendors must respond.
     
    • Like Like x 1
  19. Hypilein

    Hypilein Mu-43 Veteran

    290
    Mar 18, 2015
    I think the problem is, that with the CIPA test results, they have decided that measuring the effectiveness of multiple axis is not worth it. They probably have their reasons, but it unfortunately also leads to a situation, where IS systems are not tested with there full functionality.

    That said, it probably doesn't matter much, because most manufacturers now seem to have added 5 axis IBIS (even if just for marketing reasons). It is also likely that the tech in their IBIS systems improved. We just don't know whether this was due to added stabilisation axis or because other parts of the tech improved. Would be nice if CIPA did use their test apparatus to find that out for us.
     
    • Like Like x 1
  20. Turbofrog

    Turbofrog Mu-43 Hall of Famer

    Mar 21, 2014
    Well, first of all, I don't think that CIPA's data actually supports the position you're espousing.

    "... Camera shake has components in six directions (Yaw, Pitch, Roll, X, Y, Z) in total. However, when the shooting distance is about 20 times the focal length, Yaw and Pitch are dominant and the other directions can be practically negligible. Accordingly, the waveform only with two axes is adopted."

    That line alone demonstrates that there may be benefit to additional axes for a number of different types of photography. Macro is one that immediately comes to mind.

    With a macro lens you might use an FL of 45, 60, or 90mm. So that's a subject distance of 0.9m, 1.2m, or 1.8m respectively. With a 60mm macro, however, a common shooting distance would 0.2m. This is six times smaller than CIPA's threshold of "practical negligibility."

    Or with another scenario, with a 300mm lens, you're now talking about a 6m subject distance. In my use, such a lens is commonly used in situations where the subject is at distances 1/2 or even 1/3 as far as that. On a full frame camera, you'd be using a 600mm lens for the same 2-3m subject distance if you wanted to fill the frame with a bird, so the shooting distance is now 12m before it falls below the threshold of "practical negligibility." Note that I'm taking CIPA's text at face value here, even though it seems suspect that a value such as focal length could be used here independent of angle-of-view, or at the very least pixel density, in determining acceptable tolerances for motion compensation.

    So for telephoto and macro, two scenarios where stabilization is intuitively of the greatest value, CIPA's findings suggest that 5-axis stabilization may indeed hold a benefit because the Roll, X, Y, Z components are non-negligible.

    The practical answer seems to be that from CIPA's point of view, it was far more complicated, difficult, and costly, and for them to try and simulate a mechanism that could reliably and predictably simulate 5 axes of motion compared to 2 axes. Like any engineering problem, they weighed the costs and benefits and produced a "good enough" experiment that covers a broad range of photographic use cases, but certainly not all of them.

    The other thing is that people are not shaker-tables. They are people, and the interactions between the varying frequencies and amplitudes of their body movements are pretty complex. There are just so many degrees of freedom in a human body gripping a camera...

    As far as quantifying the multi-axis benefits in a "demonstrable" way, a variety of websites (Imaging Resource, Optyczne, and several others) test the efficacy of stabilization systems. Other review sites do it in a more impromptu fashion. They do this as you or I would, by shooting repeated series of shots of a given target at a variety of shutter speeds, and measuring how many fall within a range of acceptable sharpness (whether that's based on an MTF calculation or just eye-balling, I don't know). In these situations, there are clearly uncontrolled variables that can confound the experiment. The experimenter themselves, how much coffee they've had, how tired they are, whether they've lifted heavy weights earlier, etc... and yet, despite all the confounding factors, they all seem to show that Olympus' 5-axis system and Sony's 5-axis system are better than the competitors. Again, maybe that's a coincidence, or just a product of more development resources and more advanced technology, and the additional axes were just added for marketing's sake. I don't know. It just seems like a heck of a coincidence to me.
     
    Last edited: Apr 12, 2016