In Praise of Olympus' in Body IS

[Asmarlak]

I used Canon lenses with IS and Nikon's with VR, I found that the Canon's IS is generally a tad better than the Nikon's VR but both were not as efficient as the Olympus' in body IS when it comes to handheld shooting in dim lighting conditions. These images below were all shot handheld in Shutter Priority mode at shutter speed of 1/5. Something I couldn't do with Canon and Nikon lenses.

[Asmarlak]

And these two were also shot handheld in Shutter Priority mode of 1/10. All images were taken with the e-510.

[armando_m]

really impressive !

i'm able to get motion blurr even with a tripod as I get overly confident and jam the shooter

[Asmarlak]

Thanks Armando. I don't use tripod but If you use it, its a good idea to invest in a remote control because pressing the shutter button could cause camera shake, and always make sure that the lens' VR is turned off.

[Photo-John]

Beautiful photos, Asmarlak! Where were these taken?

I agree about the Olympus built-in image stabilization. I can't say whether it works better than Nikon and Canon's optical IS. But I do appreciate having it in the camera and not having to worry about buying or using image-stabilized lenses. Whether or not it works better, from a practical standpoint, it's better. I think I should point out that Pentax and Sony both use similar sensor-level IS systems. Again, I don't know how the performance compares, but you don't have to pay more or pick and choose your lenses for image stabilization with Pentax and Sony digital SLRs.

[Asmarlak]

Thanks John, Those shots were taken in (1,2,5, and 6) Prague, (7) Spain, (4) Brazil, and (3) Argentina. One of the very important advantages of in-body IS is that it gives the same consistent results all the time while in-lens IS varies from lens to lens, I find that alone as big plus.

[Loupey]

Nice examples. It should be noted that many of these are at super wide angle focal lengths (14mm) and so a slower SS should be possible handheld or otherwise.

But I do agree that in-body IS is a good idea and I wish Canon had a version that had this feature. Though I wonder if in-body IS would work with super telephoto lenses (400mm to 800mm) as well as in-lens IS at these focal lengths.

[Asmarlak]

Thanks Loupey. I don't think Olympus has telephoto lenses of more than 300mm considering that the crop factor for all Olympus DSLR's is x2 which makes it 600mm equiv. I have the 70-300mm and the IS system works beautifully. Here are some examples:
http://gallery.photographyreview.com...&ppuser=257353

http://gallery.photographyreview.com...&ppuser=257353

However, I was unable to obtain the same results by shooting in dim lighting handheld at 1/5 by Canon nor Nikon lenses.

[Loupey]

Those two are beautiful examples! Wish we saw those in the N&W forum.

I don't think Olympus has telephoto lenses of more than 300mm considering that the crop factor for all Olympus DSLR's is x2 which makes it 600mm equiv.

I know about the crop factor. I have a 500mm lens and on my camera it is an 800mm "equivalent". But it still is a 500mm (as is yours still a 300mm). It is important to keep crop factor (a camera function) separate from lens performance. I could always crop every image to be a 2x (or whatever for that matter).

[Asmarlak]

Loupey, I'm surprised to hear that. Because there is a huge difference between the camera sensor's crop factor that applies before the picture is taken and doubling the image size after the picture is taken. The crop factor is equalizing the focal range of the lens to the 35mm in film format because digital camera sensors are smaller than film camera sensors. The camera sees the subject and translates it to the sensor at double the size and the picture is taken at the sensor's original resolution. When you double the image size after the picture is taken is like using the digital zoom mode in Point & Shoot cameras, which is not recommended.

[Loupey]

I'm not talking about doubling the size of the image. I'm talking about cropping an image taken on my APS-C sensor (which has a 1.6x crop factor compared to 35mm film) down to the smaller size of the Olympus sensor (which has a 2x crop factor compared to 35mm film).

In other words, the lens performance remains unchanged. What should be compared is the image quality within the small confines of the imaging sensor (22mm x 15mm on mine, 17mm x 13 on yours) - not the "equivalent" image compared to 36mm x 24mm.

So if you consider any print size (say, 20" x 24" for example) the smaller sensor image must be enlarged significantly more than the larger sensor image and the limits of the sensor become more pronounced. So my 800mm equivalent blown up to 20" x 24" is not going to be the quality of an 800mm shot on a full-frame and blown up to 20" x 24" if all other factors are kept constant.

My apologies for getting this discussion off-track.

[Asmarlak]

Every camera has limited print size that should not go over but that has nothing to do with the focal range of the lens or crop factor of the camera. If a camera has 2x crop factor and use 300mm lens then what your are really getting is the same result in focal range like using 600mm lens in full frame camera with consideration of the difference in image quality between the two cameras.

[OldClicker]

Every camera has limited print size that should not go over but that has nothing to do with the focal range of the lens or crop factor of the camera. If a camera has 2x crop factor and use 300mm lens then what your are really getting is the same result in focal range like using 600mm lens in full frame camera with consideration of the difference in image quality between the two cameras.

Not - All you are doing is cutting off the outer 3/4 of the area. If you use a 300mm on an FF and on a 2x, the only thing that changes is what fits in the space. - Terry

[Asmarlak]

Forget about 3/4. Comparing Canon to Canon and Nikon to Nikon. I was talking about how close the subject becomes in the resulted image when using 300mm in 2x camera compared to 600mm in full frame camera, its the same.
However, we got away from the original subject which is in-body and in-lens IS performance.

[Loupey]

"All lenses of the same focal length give images of the same size at any given subject distance."

Exactly. That's what I've been saying. And that's why it's not proper to compare the results of in-body performance of a 300mm on a 2x crop factor system to the in-lens performance of a 600mm on a 1x crop factor system.

A proper "test" would be a 300mm lens on both systems and then check the IS performance at the sensor level where the "crop factor" doesn't factor in.

[Asmarlak]

A proper "test" would be a 300mm lens on both systems and then check the IS performance at the sensor level where the "crop factor" doesn't factor in.

That doesn't concern me really when all the cameras I use have a crop factor. I have no intention of using full frame camera at all. As the matter in fact, no one needs to use full frame camera as far as I'm concern unless they do big prints. When it comes to smaller prints I can produce the same quality images if not better as full frame does with a fraction of the cost and weight.

[Anbesol]

Loupey - what of the huge crop factor with super-tele. for example, Canons SX20IS at what is 560mm equivalent, produces a lot more shake than what I find normal for a 100mm lens, or am I over-analyzing something?

But it seems the logic holds true in any DSLR crop factor lens differences, its just super-tele on point and shoots that seems to contradict that logic to me.

[Loupey]

That doesn't concern me really when all the cameras I use have a crop factor. I have no intention of using full frame camera at all...

That wasn't my point. My original question (in post #7) was whether in-body stabilization was as effective as in-lens stabilization when used with super telephotos.

Answer these questions with yes-no ("equivalent" is not an answer )

1) Stabilizing an image coming from a 600mm is much more difficult than those coming from a 300mm, right?

2) If I have a 600mm with in-lens stabilization and test it on camera 1 and then mount it (with lens stabilization OFF) on camera 2 with in-body stabilization, then that would be a fair comparision of stabilization systems, right?

3) Isn't #2 the only way to really test the effectiveness of the two different stabilization systems? This is a stabilization thread right?

In terms of stabilization, I could care less what sensor is actually in the bodies. I would just like to know if in-body stabilization would be as effective (or even possible) with such extreme focal lengths.

[Loupey]

...Canons SX20IS at what is 560mm equivalent, produces a lot more shake than what I find normal for a 100mm lens, or am I over-analyzing something?...

At the long end, the lens is a 100mm focal length lens. But because the imaging sensor is so small, it only looks at a small portion of the image. At this extreme level of "crop", any movement caused by camera shake is going to be much more pronounced.

[Asmarlak]

Loupey,
I think you're going to extreme examples which this thread is not about. My original post is about shooting in shutter mode of 1/5 "handheld" in dim lighting condition. It was not about telephoto shooting. And yet here are few examples of "HANDHELD" telephoto shots, those shots came out at around 600mm equiv., weather I got there by a crop factor or full frame, it shouldn't really matter at least not to me. I never cared for charts or test results that proven to be unreliable and differ from the actual images that cameras produce from day to day shooting under different circumstances.
I think any telephoto shot taken by a lens longer than 300mm would need a tripod regardless of the camera class you use or IS performance, is that satisfying enough?.

[Loupey]

I agree that we'll keep going around this tree

...I think any telephoto shot taken by a lens longer than 300mm would need a tripod regardless of the camera class you use or IS performance, is that satisfying enough?.

And no, this would not be "satisfying enough". It would be a huge step back since I already do a large portion of my shooting at 500mm and 1000mm (real focal length) handheld or on a monopod. But I love to imagine someday when I could use both in-lens AND in-body stabilization in concert.

As for now, I think in-body stabilization is not capable with super-telephotos and that is probably why Canon and Nikon have adopted only in-lens systems at this point in time. And super-telephotos are the realm of sports and nature shooters for reasons I stated earlier.

[Loupey]

FWIW, I can't tell if we're arguing or not. If we are, I don't know why because I think both stabilization systems are great in their current capacities.

[Asmarlak]

Here are more examples of handheld shots I took in Shutter Priority mode at shutter speed of 1/5 (no flash).

[OldClicker]

As for now, I think in-body stabilization is not capable with super-telephotos and that is probably why Canon and Nikon have adopted only in-lens systems at this point in time. And super-telephotos are the realm of sports and nature shooters for reasons I stated earlier.

Loupey, I think that the problem with this logic is that near future IBIS will surpass first generation ILIS so, unless you intend to upgrade lenses as you do bodies now, it will no longer be true. I don't know about Nikon, but I suspect that Canon uses ILIS because they have been doing it for 15-20 years in binoculars.

Asmarlak, Smaller sensors have many advantages to many shooters, but bringing the image closer is not one of them any more than putting masking tape on a print to block off the edges is.

Terry

[Loupey]

Loupey, I think that the problem with this logic is that near future IBIS will surpass first generation ILIS so, unless you intend to upgrade lenses as you do bodies now, it will no longer be true. I don't know about Nikon, but I suspect that Canon uses ILIS because they have been doing it for 15-20 years in binoculars.

Good point, OldClicker - I agree that both systems will continue to evolve and improve. I think Canon's IS is in its (4th?) generation. And no doubt the in-body systems will see dramatic improvement in short order.

From an engineering standpoint I think the hurdle with in-body stabilization with such highly magnified images is that it would take a HUGE amount of compensation (I'm thinking in the range of 3mm~5mm in both X and Y) moving extremely quickly to match the dancing image at the film plane. Whereas by moving the elements within the lens of a lens stabilizated system compensation of that order of magnitude is not required.

There is no question that in-body stabilization helps. All P&S use this technology over the in-lens systems. But I think that is why it is not as effective in P&S that have superzooms (when zoomed out to the long end).