Nature Photographers Network
This is the preserved archive of the original Nature Photographers Network. The community continues at naturephotographers.network.

Image Stabilization - Sensor vs. Lens

Discussion

6 replies
Posted by David Bostedo on October 1, 2006 at 4:01 PM
I've been thinking about making the jump from EVF to DSLR sometime in the near future, and have been looking at cameras. There are several new models that have image stabilization at the sensor. (Sony a100, Pentax K10D especially) This is in contrast to having the stabilization on the lens, as Canon and Nikon seem to do.

Does anyone know of any resources comparing the two? Anyone have feelings as to which method might be better, and why?

Seems like having it on the sensor would be awesome, since every lens becomes stabilized. Yet Canon and Nikon don't do it - just to make more money perhaps?

Thanks.
Posted by Colin Smale on October 2, 2006 at 3:26 AM
My guess is that Nikon will do it soon, I certainly hope so.
Posted by Basil Greber on October 3, 2006 at 11:15 AM
Well, I think there is a problem with the "Nikon rip-off" theory: The non-IS-ed lenses for Sony, for example, are more expensive than the equivalent VR-Nikkors (at least some, for example the 70-200/2.8). Generally, I think Sony's glass is overpriced.

Additionally, there are rumors that in-body VR is not as efficient as VR in the lens.

Basil
Posted by Jim Huddle on November 1, 2006 at 10:18 AM
I use a Minolta 7D and find the image stabilization to be an excellent feature. In a lot of instances in frees me from the tripod I would have normally used. Already owning a bunch of prime Minolta lense the 7D and now Sony path works well for me. But, Sony has way overpriced the lenses right now and used Minolta glass is also getting premium prices in the used market.
Posted by David Bostedo on November 1, 2006 at 7:29 PM
I've done some more research on this, and found a couple of sites referencing Canon sources. According to Canon (according to these sites), long lenses lose a lot of stabilization if you do it at the sensor. This is because the optics in a longer lens move further (for the same amount of shake) than a shorter lens. This means that the sensor would have to move a lot more to correct it (something like a quarter inch or more - which isn't practical). So while you'll get a lot of stops of correction with short lenses, supposedly long lenses are much better off with the stabilization in the lens. (Note : I don't think I agree that smaller sensors also need to move more, since the image itself will be moving on the sensor in the same way, regardless of sensor size.)

Here's a link :
http://www.photographyjam.com/blog/57/why-canon-and-nikon-use-in-lens-stabilization
Posted by Jim Huddle on November 6, 2006 at 1:01 PM
Hmm, I'm not an expert on this but would it not depend more on the distance of the final rear element to the sensor since that is actually what sends the image to the sensor? In actual use I find the in camera stabilization to work extremely well for me. I regularly my 300mm f/4.0 with a 1.4x t-con at shutter speeds much slower than I would think possible to handheld and still get sharp images.
Posted by David Bostedo on November 6, 2006 at 1:37 PM
Jim - If the final element of the lens in front of the sensor is what moves to correct for shake, then I think you're correct. I'm not sure which element is moving in Canon's or Nikon's stabilized lenses though. Maybe someone else knows (or it might be in the published white papers). If an element further from the sensor were the one moving, it should be able to move less for the same effect.

It's good to hear that you get good results with such a long lens setup though.