Technical question about IS and image stabilization

[inTempus]

Does the crop factor of an ASP-C image sensor have anything to do with how effective image stabilization is? Does it make a difference between lens based IS and sensor based IS?

If you have a 300mm lens, does the image stabilization become any less effective on a 1.6x body vs. a full frame camera?

Just curious.

 

[Garbz]

No. Think of it this way. If you're moving the camera lens a certain distance (say 0.5mm) then the IS needs to compensate for this by counter-moving the 0.5mm. This is the same 0.5mm regardless of sensor size.

With on sensor based IS one could argue that APS-C sensors are smaller, lighter and easier to move. But then the difference is marginal anyway. Ultimately the quality of the IS is dependant on the control system in place.


One thing that is different is the focal length thus a 200mm lens on full frame with a shutter speed of 1/50th should be easily compensated by every IS on the market (assuming the rule of thumb that a 1/(focal length) shutter speed is steady, whereas with APS-C since you now effectively have a 300mm lens, 1/50th would be more than 2 stops, so either the shutter speed would need to be 1/75th, the IS would need to be half a stop better, or the user would have to zoom out to 133mm.

 

[Big Mike]

Yes but....you don't effectively have a 300mm lens on an APS-C sensor. You still have a 200mm lens, it's the camera that is only seeing the centre portion of the image circle.
So if you need 1/(focal length) for a given lens...does that really change because you are only recording part of the projected image?

I've heard an argument that says yes because the image on a crop sensor needs to be 'enlarged' more, thus enlarging the blur from the movement.

I'm not sure if I buy that argument...but I'm not sure where I sit on this issue :scratch:

Either way, a faster shutter speed is always better for avoiding blur.

 

[inTempus]

My assumption is that EFL has little or nothing to do with how effective IS is with either the lens based IS or sensor based IS. Sure, it could be argued that the blur would be more obvious, but in my experience my crop sensor based bodies were no more susceptible to blur than my full frame body was. If there is a difference, it's negligible.

 

[itznfb]

It should not be any different due to the size of the sensor since 300mm on full frame IS technically 300mm on APS-C. You aren't any closer to the subject and neither is the sensor.

 

[Stosh]

I think the answer to this debate is both sides are correct depending on the reference point.

Reference point A - "the pixel level" or "diffraction limit level": At this reference point full frame vs. 1.6 crop frame makes no difference. Of course I'm assuming that the pixel size for the crop frame and pixel size for the full frame are the same. That's why I also mentioned the diffraction limit.

Reference point B - I'm not even sure how to correctly say it, but maybe something like "the best total quality image" or "the largest possible enlargement": From this reference point the crop frame does indeed magnify any problems 1.6 times. Suppose you had a full frame and crop frame of equal megapixels (and assuming your lens is sharp enough and fast enough to not be diffraction limited for either sensors) then it would be even more critical to have a steadier image for the crop frame. In this situation the crop frames pixels are much closer together (they would have to be to have the same resolution). Since the pixels are closer together, a 1 pixel blur on full frame would be a 1.6 pixel blur on the crop frame.

I would think this minor difference would be hard to notice and in fact I have not seen a significant difference between my crop frame and full frame shots with same lens, but it must be there.

EDIT: would the pixel blur be 1.6x or would it be 2.56x (1.6^2)? Sorry, I'm not in math mode right now.

 

[Garbz]

More suseptible to blur is the same principle I was talking about. A 200mm on FX is no more suceptible to blur than a 133 on DX. However if you have a 200mm on FX and a 200mm on DX yeah sure the crop factor would create more blur given the fact that the effective focal lengh is higher and it's harder to hold steady a longer focal length.

This point is moot though when you consider that if you compare two identical focal lengths on FX and DX you end up with two completely different pictures, making the comparison irrelevant.

Stosh you too are comparing apples to oranges. The size of the pixels are an additional variable not properly relevant to sensor size. Consider does the Canon 1Ds MkIII have a higher pixel density than a Nikon D2h?

 

[Stosh]

Stosh you too are comparing apples to oranges. The size of the pixels are an additional variable not properly relevant to sensor size. Consider does the Canon 1Ds MkIII have a higher pixel density than a Nikon D2h?

That's why I used 2 reference points. What you're saying is relative to my reference point B. OK, all the extra talk about pixels was probably unnecessary, but I thought it might make it easier to visualize. But I agree with you that you wouldn't use the same focal length lens when you're switching between the 2 sensor sizes, so yes, it's moot point.

If for some crazy reason you wanted to look at the perspective like Big Mike said - it's still a 200mm lens no matter which sensor it's on. This would be my reference point A. At the diffraction limit of the lens it makes no difference at all what sensor size you have. Once again, this is a moot point because you would have to consider a sensor size so that you can then choose your focal length.

This is basically the same as the recent DOF question so easily explained by Moglex. You can't compare same focal length lenses on different sensors because they wouldn't give you the same FOV.