Stabilization. A term that, before a handful of years ago meant “tripod,” or physical bracing technique, has grown to provide various hardware solutions within our camera system of choice. We as consumers have been lucky to have stabilization options within most all digital camera systems, and while image stabilization isn’t going to remedy all problems, it is certainly a nice feature to have.
I’m awaiting a new Panasonic GX8 to arrive within the next couple weeks which will boast a new, dual IS system incorporating both an on sensor IS and lens based IS solution, but before that time, I wanted to really see how the first full frame, 5 axis on sensor/in body image stabilization (IBIS) system from Sony compares to a very good lens based image stabilization (IS) system in the Canon EF lenses, and a better than often credited 2 axis IBIS system from Panasonic’s first foray into on sensor stabilization, in the GX7.
Come on in to see my three different comparisons between these three different offerings, and see if there is a clear winner.
Thanks for stopping by. This test is merely a personal exploration to see what kind of results I can get with the gear I have at my disposal, and the gear I’ve chosen to use for various jobs or situations. I’m not looking to build a comprehensive catalog to be the end all, be all for reference on image stabilization, but rather to share my experience with readers who may be looking for a little insight into how these systems perform. I’ve received emails that have asked why I haven’t compared the Olympus 5 axis IBIS, or super telephoto lenses. Simply put, it’s because I don’t own them. I don’t work for these companies, and I don’t make any money from them, so these are just my own findings because I’m curious. I feel it is also more accurate to see these systems run through a minimal amount of times as opposed to taking 20 pictures at each setting to see just how well I can get it to work. I don’t take 20 pictures of anything, and really just wanted to run through a controlled setup, as I would perhaps on the street, or shooting anything at a moment’s notice as it were. To clarify and as mentioned later in the article, I took a series of shots at one stop intervals, twice through on all setups (3 camera/lens combos for three different setups, twice through each) to see how these various combinations performed. The reason I did two, and not ten or however many, is because I wanted to keep my findings somewhat genuine to an average shooter. I did them all on a full stomach, didn’t take a coffee break half way through, and have worked on my technique to be consistent over the years. These are my findings, and if someone would like to fund me to do a far more comprehensive test, I’d love to talk, but with time limited, this has helped show me what I have in these systems, and hopefully is useful to you as well. Thanks again, and let me know if I can try to answer any questions specifically.
Cheers – Tyson
Firstly, let’s discuss what image stabilization can and can’t do. With stabilization, optical or on sensor, the aim is to account for movement at the physical point of capture, meaning, photographer/camera movement, handshake, vibration, et al. It does nothing for the movement of the subject. Getting a 3 or 4 stop “IS advantage” means nothing if the shutter speed you’re shooting at isn’t fast enough to freeze the movement of your subject, it only accounts for your (photographer) movement or vibration at the location of the camera.
A 5 axis system traditionally accounts for pitch (think nosedive), roll (spinning on the lens axis, like using your camera as a steering wheel), vertical axis, horizontal axis and pivotal movement (like a record spinning), while most 2 axis or optical based IS systems account for vertical and horizontal movement, with newer OIS/IS/VR or even some 3 and 4 axis IBIS systems adding various other axes, in cases.
Some IS systems provide a stabilized live view, or stabilize the image through the viewfinder by way of the optical stabilization. This is very handy when shooting at longer focal lengths to help keep the image from jumping all over the place enabling an easier task when composing a shot.
I’m going to do three different comparisons with three different systems/cameras here. The Sony a7II full frame camera, the Canon EOS 5DII full frame camera and the Panasonic GX7 micro 4/3 camera. The Sony is the first full frame camera to incorporate on sensor IBIS, the Canon system still relies on lens based IS, and the Panasonic GX7 is the first Panasonic m4/3 camera to incorporate IBIS, in a 2 axis system. The tests will be:
- First, on the Sony a7II using the Canon 70-200mm f/2.8 L IS USM lens with the Sony IBIS engaged, and the lens based IS, turned off, shooting at 200mm. I’ll shoot that same lens on the Canon 5DII with the IS on also shot at 200mm, and the GX7 using the EF 70-200 (shot at 100mm = 200mm E-FOV) with the 2 axis IBIS. I want to see 5 axis and 2 axis IBIS vs lens based IS here at the same effective focal length.
- Second, I will be turning off the Sony IBIS, using the lens IS on the Canon 70-200mm set to 85mm, I will test it against the same lens on the 5DII with the lens IS on, at the same focal length, and the GX7 with Pana-Leica 42.5mm f/1.2 OIS (optical image stabilization) which disables the IBIS on the GX7. This will be lens IS/OIS specific.
- Finally, I’m going to shoot at a standard focal length with the Sony a7II using the Zeiss 55mm f/1.8 utilizing the 5 axis IBIS, the Canon 5DII using an adapted, and in no way stabilized FD 55mm f/1.2 SSC lens, and the GX7 using the Lumix/Leica 25mm (50mm E-FOV) f/1.4 lens and the 2 axis IBIS. This will be 5 axis IBIS vs 2 axis IBIS with the Canon 5DII and non stabilized 55mm lens as my comparison to see how close I can get with just handholding technique alone.
For all the tests below, I took two separate series at identical settings to hopefully eliminate singular irregularities. I took the best shot, at each setting, from each combination shown, but at most all shutter speeds, results were close to identical. While there are certainly going to be differences in an individual’s handholding technique, and subsequent results, this is me, trying to get the absolute best results that I can with the gear I have to truly see if I can see an advantage. I could have done a dozen series, but felt that would have gone against a true test (plus I don’t have that much time to donate unfortunately, kids, work, life and all). That said, I wanted at least a small sample size to see if results were consistent. In most all cases, they were, and in the few frames I was able to see a noticeable difference, I chalked that up to individual, momentary technique. I rarely have the time, nor does the moment afford me taking 10 or 20 shots of a single subject where IS or IBIS would be useful, so I want to see what I’m able to get while focusing on my shooting technique, and how these systems can benefit that by doing these runs in one fell swoop.
I’m looking at whether I get better results from IBIS, or lens based IS, and just how much better or worse different IBIS systems may be, even though one is brand new and stabilizing a full frame sensor, versus a 3 year old 2 axis system on a 4/3 sensor. All three cameras and lenses, for all three tests were shot at the same exposure settings (shutter speed, aperture and ISO). There were a couple situations, where to match the shutter speed (which is really what I’m looking at here) due to ISO 125 being the lowest ISO on the GX7, I needed to adjust the aperture in an equal number of 1/3 stops to achieve the same shutter speed. Exposure is identical, and again, most importantly here, shutter speed is identical. I’m not worried about resolution, or noise here, but just how steady a result I can get.
Ready? Let’s get into it.
Test 1 – IBIS vs Lens Based IS
For the first test, I’m looking at a telephoto focal length, utilizing an EF 70-200mm f/2.8 IS USM lens, adapted to the a7II and using the Sony’s in body “Steady Shot” image stabilization, as well as shot on the Canon 5DII while using the lens’ internal IS for a native base comparison. To compare this longer focal length on a 2 axis IBIS system, I’m shooting the Olympus 40-150mm f/2.8 lens on the Panasonic GX7 at the same exposure values, and at 100mm to equal the field of view of the 200mm focal length on the two full frame cameras.
The first shot shows the full file as captured at 1/200 sec, for reference, then resized 100% crops are provided below each. I resized them because we’re looking at relative sharpness when handheld, not resolution or s:n here, and to make it easier to see the same relative area in the frame. Click to see this chart at full size (you may need to adjust your browser window to see it at 100%).
To my eye, the Sony a7II, using the in body 5 axis IBIS does really well up to 4 stops, with one slight hiccup at the 3 stop increment where the Lens based IS on the 5DII seems to out stabilize it, but it does well to recover at a stop slower. Again, this was the best of each (of two series) of the shots. The 70-200mm lens’ IS is really good for 3 solid stops in my opinion, and starts to go south after that. The GX7’s 2 axis IBIS holds up well here also. Certainly not quite as well as the 5 axis, nor the EF lens based IS, but I’d still consider it to provide 3 decent stops of handheld compensation, and we may actually be seeing “sharpness” differences, to an extent due to the substantial difference in resolution between a 24mp full frame sensor (a7II), a 21mp full frame sensor (5DII) and the 16mp m4/3 sensor. I’d say that in this test, the 5 axis has a near stop advantage overall on either of the other two at 200mm.
Test 2 – Lens Based IS/OIS
For this second test, I wanted to look solely at lens based image stabilization. I turned off the Sony’s IBIS, and engaged the lens’ IS on the 70-200mm lens for both the a7II and 5DII, both shot at 85mm to equal the Panasonic-Leica 42.5mm f/1.2 lens’ field of view. I wanted to include the Metabones smart adapter’s ability to relay the IS from the lens to the Sony body here, along with seeing a pure lens based IS comparison versus a sensor based, IBIS variable.
Again, click on the chart to see full sized.
Oddly, at a shorter focal length, the 70-200’s IS seemed to suffer more than when stabilizing at 200mm on both the a7II and 5DII. The Panasonic’s OIS still provided 3 solid stops while we could say about the same for the EF lens’ IS, but not quite as effectively as the Panasonic’s.
Test 3 – 5 axis IBIS vs 2 axis IBIS vs Carbon based/No axis IBIS
Finally, I wanted to look at a native, standard lens shot on all three cameras with the Sony a7II using its 5 axis IBIS, the Canon 5DII using a non stabilized lens for comparison, and the Panasonic GX7 using its 2 axis IBIS. On the Sony, I shot with the stellar Zeiss 55mm f/1.8, on the Canon, a converted FD 55mm f/1.2 SSC lens and on the GX7, the Panasonic-Leica 25mm f/1.4 lens (50mm E-FOV).
Same deal, click to see larger.
The Sony’s 5 axis clearly wins here in my eyes with about a 4 stop compensation, the Panasonic 2 axis pushing to about 3 stops, and my hand holding technique’s limits shown with the non stabilized combo on the 5DII.
I’m impressed by the Sony’s 5 axis IBIS, I’m impressed by the EF lens based IS and I’m impressed by the Panasonic 2 axis IBIS. My takeaway here is that I’d rate the Sony’s 5 axis, overall, the best with the EF lens based IS bettering it in certain circumstances, and the 2 axis Panasonic IBIS nearly matching it, or being within a stop under most circumstances. The final test shows just how much difference in body stabilization can make, regardless of the number of axes.
If you’d like to see more reviews, here are some articles I’ve written on the specific cameras and lenses used, or you can always find a list of gear and software reviews at the top of the page HERE.
Metabones EF>Sony E mount, smart adapter Mk 4
Sony a7II, a game changer? User review
I’d love to hear your thoughts, or your experience with any or all of these systems. Thank you for the read, and hit me up on the socials; Facebook, Twitter, Flickr and Instagram. If you prefer email notifications, please feel free to add your email address at the top right of the page here. You’ll get email alerts as new articles are released.
Pingback: *Sony 5 axis IBIS vs EF Lens Based IS vs Panaso...
Hey Tyson thanks for this review – great as usual! Let us know if you end up doing a “round 2” i.e. OMD-EM5 Mk2 vs GX8! Keen to know what your thoughts are around 5-axis IBIS (EM5 Mk2) vs 2-axis IBIS (GX8)!
Strange, why was Olympus stabilisation excluded?
Because I don’t own an Olympus body anymore 🙂 I can only compare what I have, as it were. The Oly 5 axis (at least the original iteration) did well, and I found it to basically equal the 2 axis from the GX7 in most cases, although, I caused quite a stir when I rated the handheld results of the 2 axis system better at slower shutter speeds versus the original 5 axis Oly IBIS in my reviews pitting the GX7 vs the EM5 a few years back. Still entirely stand by that, but have yet to personally get hold of a EM5II, as I just don’t particularly like interacting with the Oly cams as much myself.
Hey T, It might be interesting to compare hand-held stabilization with a tripod-mounted rig. Maybe use a high end tripod & head versus cheaper combo versus your standard carbon- based system. Enjoy all your postings. Thanks.
It could be interesting, especially for longer focal lengths. Even when on a tripod, tele focal lengths are often affected by vibration. I find using the 2 sec self timer on tripods just about eliminates it most times.
Tyson, this article definitely caught my attention and did not disappoint. Thank you! I’m (currently) shooting with Olympus cameras. So I’d be be greatly interested in your checking out an IBIS Part Deaux between the Olympus E-M5 II with its latest Olympus advances versus the new Panasonic GX8 and its brand new approach to stabilization. Thanks again, and looking forward to more on this topic! Image stabilization seems to be among the top priorities for several camera companies right now.
I’d love to get my hands on the EM5II, and would gladly compare the differences. I’m not planing on buying one though, so I’d have to find a loaner.
Thank you for the read through.
This test totally ignores the reason that Canon/Nikon use in Lens Stabilisation, which is that with longer telephoto lenses, on sensor stabilisation becomes progressively less effective. Past 200mm it is bordering on unacceptable. A real world example I have tried is a 500mm lens on a sensor stabilised Pentax – the stabilisation is useless. So you should re-run your test with a telephoto lens, or at the very least the 70-200 set to 200mm. Stabilisation is a fun buzzword, but the reality is that any shutter speed below 200m/s risks motion blur from the natural movement of people – and 200ms is fast enough not to worry about hand held blur on any lens below 100mm. Stabilisation is actually only really becomes a necessity when the lens is over 200mm. Of course tamron and others have started introducing it on wide angle lenses and 24-70 so now the market has to follow – but the truth is that it is really not needed and pros have been shooting 24-70 2.8 zooms and primes without stabilisation or
complaint for many years.
Hello Mr Hope.
Firstly, this test ignores nothing, really. It’s not listed as a fully comprehensive, scientific test to fully explore and deconstruct all aspects of image stabilization, so I apologize if I misled you in any way. I simply wanted to compare the three systems I currently shoot with, with the lenses and cameras I own. Nothing more than that was claimed, nor done.
I feel that Canikon chose lens based IS because the technology for sensor based IS at the time they chose, wasn’t possible, because sensors weren’t around yet seeing and the IS was introduced in ’95 with the 75-300 zoom on the Canon system side. They continued the IS system as it was more advantageous to build it into a lens as it was focal length specific, and arguably better with the technology available at the time, plus they could charge for each lens, not a one for all system in a camera as it were. I feel that as time has passed, sensor based IBIS has done very well to provide a better overall system, and while perhaps there are situations where a lens based IS will be able to more fully support the specific lens it is housed in, that advantage will be measured in fractions of a stop as opposed to full or multiple stops nowadays. Is it better to have a system that provides arguably ~95% of the effectiveness, built in, or to have to pay quite a bit more for lens specific IS with each optic you may want to gain an (assumed) advantage? I don’t have an answer, but I do like that there are choices. Choices are good, and I’m certainly not going to criticize Canikon, Fuji or Sony or Panasonic, et al for lens based IS, but what we’re seeing is that some companies (Sony and Panasonic for instance) offer both options within their system, and that to me seems like forward thinking as opposed to sticking to your 20 year old guns as it were.
Actually, I don’t own, nor plan to own any super tele lenses for the Canon system other than an old film era super tele, but since you did bring up the super tele focal lengths, I actually did test my 400mm f/4.5 adapted FD lens on both the 2 axis GX7 and 5 axis EM5 a couple years ago, and both did very well with the GX7 offering upwards of 3 stops of hand held compensation, which at an 800mm equivalent was, in my eyes, substantial that I was getting really sharp shots at 1/100-1/125 sec exposures. So no, I don’t think that on sensor IBIS is necessarily less effective, wholly speaking. Over the last decade, it has been built upon to do very well if the proper focal length is input into the camera. To clarify, telephoto focal lengths are any focal length longer than the physical length of the lens, enabled by way of optical groups (a telephoto group specifically). “Telephoto” focal lengths actually start, in many cases around 70-80mm in many lens designs with “short” tele lenses falling into this range, the “mid tele” range between 100-150mm or so, and beyond that, we’re starting to get into long or super tele range. 200mm is absolutely a long telephoto focal length, and the focal length for the first test. Beyond that, we’re talking about super telephoto lenses, which are specific tools, and from my experience renting the 300mm f/2.8 IS lens (or using the adapted 400mm lens with IBIS), while it’s IS system is great, it’s not head and shoulders above an IBIS system shot at equivalent focal lengths. Where I do see the potential for very specific and situational wins on the Optical IS side, is in horizontal or vertical panning perhaps, which would be handy for birders or sport shooters for instance. Even that though has been well designed into the IBIS systems, again, I’d guess within a half stop of the lens based IS systems at worst.
Sorry for the lengthy response, and I don’t mean to come off as disrespectful. I just don’t feel my user test here “totally ignores” anything, but rather looks at a few things when adapting lenses in our new, digital reality.
I appreciate the comment.
All the best,
Hmmm. One responder states that stabilization only becomes necessary for lenses over 200 mm, but I think I can show comparisons (please don’t ask me to) of photos taken with a short focal length at the usual recommended minimum shutter speed, or rather, significantly above that, both with and without a tripod (this was pre-image stabilization, and I think, pre-digital, but no matter), and after viewing enlarged versions of both, the tripod wins hands down. It’s actually quite shocking. This is not a defense of the tripod per se, believe it or not, but rather a defense of image stabilized lenses or bodies even when the lens is of a shorter focal length. If necessary, I can provide sources (I asked that I wouldn’t be asked to do so!). So even though I myself initially minimized the importance of image stabilization on shorter focal lengths, I have to concede that in poor lighting, if higher ISO/ASAs are not desired, they definitely have their place.
Anyway, thanks Tyson for yet another informative, objective test.
I think that my ‘normal’ lens test here can prove that while perhaps there is a decent argument that shorter focal lengths don’t quite require stabilization as we can arguably use the 1/focal length rule to use slower shutter speeds to begin with, my hand held results have shown me that even with a ~50/55mm lens, stabilization can certainly help.
Usually true but for example, the E-M5ii 5-axis provides excellent stabilisation with the Panasonic 7-14, by killing vibration it produces a highly detailed image. I believe this has also been demonstrated (in a magazine) by a Sigma UWA OS lens. Olympus may also be developing dual-IS for their yet to be seen extra long lenses as they use the same VCM as their body stabilisation system in the E-M5ii (which they have stopped calling IBIS, go figure).
There are absolutely situations where stabilization aids wider angle lenses, I totally agree. We can traditionally get away with less of a need for those though, in most cases.
I think that this will become largely irrelevant moving forward because I think we’re going to see stabilization, in some form, cover most every scenario in the near future.
correction – usually true [that stabilisation is not required for short lenses].
Thank you for the good comparison ^^.
my pleasure, plynoi 😉
I did a similar series of tests, about 600 shots, using the E-PL7 (3-axis IBIS) as the main test subject, with some comparison to the GH4. Main finding was Panasonic lens OIS provided the better stabilisation in normal use above 100mm (note this is above equivalent 200mm FoV). Below 100mm, the IBIS. The 100mm focal length (200mm FF equivalent) seems to be the “knee” or cross over point. The IBIS could and did often choose shutter speeds too slow for the job. The Panasonic OIS can be equivalent to 5-axes by combining axial XXZ movement, pitch and yaw can be compensated but not roll, it appears to be better at holding focus by jiggling the lenses than body stabilisation. The main benefit of stabilisation seems to be killing the jizz of vibration rather than gaining extra stops. I’ve repeated this with the E-M5ii with the same conclusion; it may in part be due to the length and mass of long and extra-long lenses; lens tests from 12mm to 450mm (i.e 24mm to 900mm FF equiv).
Pingback: a bit of everything… | Photohangout
the a7 5axis on active with native FE lens with on lens IS. if you adapted and non Native IS lens to it it only 3axis works
It should be interesting to see what the Panasonic GX8 does with Panasonic OIS-equipped lenses. Given that there is also an Olympus patent, I wonder if this will actually work well.
I agree. I have very few Pana OIS lenses, and from memory, they aren’t yet supported by the dual IS yet, but hopefully will be updated soon. In theory, allowing the lens and body to work together should be effective in that the lens OIS is specifically designed for the focal length while the 2 axis IBIS can account for vibration or movement of the camera itself on the axes that aren’t really compensated via the traditional optical formula… It’s only gonna get better from here I’d guess 🙂
Hey, got the GX8 a few days ago, in bundle with the 12-35mm f/2.8 and I’m not convinced yet. I keep wishing that Panasonic had designed for a 67mm filter size instead of 58mm to use a more powerful lens and OIS design. The 35-100mm f/2.8 still stutters during video.
Curious to hear about your feelings on the GX8. I’m up and down. A lot of great features, and a few oddities I’m finding frustrating so far.
I’m pleased for the most part. It’s so much like the GH4. I’ll leave a link at the end of this comment with my blog entry.
The video is fine, and it was already set to the 1080p video quality that I already used on the GH4. Response seems a little slow but firmware is version 1.0.0.
The EVF makes it difficult to find a good angle. I sometimes have a blackout angle when I’m wearing sunglasses, as with the GH3 and E-M1.
Image quality is good so far, JPEG or raw files. Battery life is not good.
Pingback: Sony Tidbits… | sonyalpharumors
In a short time canon will also have IBIS. Let’s see how that turns out to be, comparred to Sony IBIS.
I believe that having a history of lens based IS can provide further benefit for Canon and Nikon (if they choose to pursue on sensor IS) just as Panasonic has started to mature their dual IS which I feel can be the best of both worlds, in theory. Lens based IS is superior for many focal lengths, but the benefit of having a fully stabilized (non jumpy) live view through an EVF or on screen full time, is very handy with sensor based. The sensor based IS, for me, is wonderful in that any adapted lens can benefit from it as well.