12mp vs 24mp camera shake, rattle and roll?

[lowep]

Have heard here and also verified from my own experience that hand-holding MFDB rigs with higher MP sensors results in higher risk of camera shake than doing same with lower MP DSLRs.

Does this also apply to the newer generation of 24MP full frame DSLRS (eg 5D Mark II) compared to the older generation of 12MP full frame DSLRs (eg Canon 5D)?

Has any Canikonian who has upgraded from "about 12MP" to "about 24MP" noticed any difference?

[torger]

The visibility of shake when pixel peeping increases the smaller the angle each pixel covers. That is higher resolution and smaller field of view (longer focal length). So yes this counts also for 135 DSLRs.

[sbay]

Does this also apply to the newer generation of 24MP full frame DSLRS (eg 5D Mark II) compared to the older generation of 12MP full frame DSLRs (eg Canon 5D)?

Absolutely at a pixel level. On my older 5D I could regularly get pixel sharp images at using speeds as slows as 1 over 1/2 the focal length (with IS). On the 5D II, I don't like shooting handheld unless I'm at 1 over 2x the focal length (with IS as well).

[ejmartin]

You can always get the motion blur of a 12mp camera by resampling your 24mp camera to 12mp.  You cannot however recover the detail of your 24mp camera from a 12mp camera. 

[lowep]

On my older 5D I could regularly get pixel sharp images at using speeds as slows as 1 over 1/2 the focal length (with IS). On the 5D II, I don't like shooting handheld unless I'm at 1 over 2x the focal length (with IS as well).

This is exactly why I looking into a DSLR for shooting hand held that I find difficult with MFDB - though monopod helps. Question is will a newer high MP DSLR make a significant difference or would I be better off with an older generation DSLR like the 5D that I have found is a lot easier.

You can always get the motion blur of a 12mp camera by resampling your 24mp camera to 12mp.

EJ does this also mean: You can always get RID OF the motion blur of a 24mp camera by resampling your 24mp camera to 12mp.[/i] ie if I resample the 24MP file to 12MP size in post the motion blur will be no less noticeable than if I had made the photo with a 12MP DSLR rather than a 24MP DSLR?

[RazorTM]

ie if I resample the 24MP file to 12MP size in post the motion blur will be no less noticeable than if I had made the photo with a 12MP DSLR rather than a 24MP DSLR?

Yes.  If both sensors have the same physical dimensions but different resolutions (i.e. Canon 5D and 5D Mark II), the motion blur will be identical;  however, the 5D Mark II image will simply have more resolution.  Motion blur has everything to do with the field of view and nothing to do with resolution.

[lowep]

Hope this is not tooo dumb a question but ask anyway> in real life how do I do this: ie resample a 24MP file to a 12MP file using photoshop that gives me the options to adjust pixel dimensions, width x height & resolution?

Also curious to learn more about how field of view>focal length plus sensor size> effects motion blur, does it mean the bigger the physical size of the sensor the bigger the risk of motion blur regardless of the number of pixels or... maybe I am the only dumb kook who doesn't understand this and someone can link me to a page that even I can understand?  

[bjanes]

You can always get the motion blur of a 12mp camera by resampling your 24mp camera to 12mp.  You cannot however recover the detail of your 24mp camera from a 12mp camera. 

The degradation if image quality at the plane of the sensor due to motion blur is independent of pixel pith as Emil states. The same applies to diffraction. If the Airy disc produced by diffraction at a given aperture is appreciably larger than the pixel pitch, the full resolution of the sensor can not be achieved. However, if one stops a D3x down to f/22, the effect on resolution in the plane of the sensor will be no greater than with the lower resolution D3. However, if you want to retain the full resolution of the D3x, you can't stop down too far. Also, alignment issues become more critical as resolution increases.

The take home point is that the full resolution of the higher resolution camera will be more difficult to achieve under field conditions, but the resolution will be no less than with the lower resolution camera.

Regards,

Bill

[RazorTM]

Hope this is not tooo dumb a question but ask anyway> in real life how do I do this: ie resample a 24MP file to a 12MP file using photoshop that gives me the options to adjust pixel dimensions, width x height & resolution?

Go to Edit-> Image size... on the menu.  To get half the megapixels, select Percent instead of Pixels and type 70.71% in the box (the square root of 0.5).  Then you can adjust the physical dimensions of the print at the bottom before clicking on OK.

[TheApprentice]

Also curious to learn more about how field of view>focal length plus sensor size> effects motion blur, does it mean the bigger the physical size of the sensor the bigger the risk of motion blur regardless of the number of pixels or... maybe I am the only dumb kook who doesn't understand this and someone can link me to a page that even I can understand?  

I hope my summary is correct;

If you hands make the camera shake up and down by 1mm, this will have more of an effect if you use an 800mm lens than if you use a 35mm lens. this is because the narrower field "magnifies" any wobbles...
think about when you look at a an out of focus photograph as a small thumbnail image: it appears acceptably sharp... however as you zoom in, it becomes less sharp... this is because you are making any imperfections bigger.

if you try handholding a long lens like a 200mm on your camera, you see the wobble more than if you use a 24mm lens on the same camera.
your hands still shake by the same amount, but the longer lens makes it more obvious.

it is all about angles and things, so it kind of gets complex, but you can think of it as an un-even see-saw in a playground.
if you hold one end steady, the other end will wobble with your pulse... but if the other end is longer than the end you are holding, it will appear to wobble even more

I hope these ramblings help a little!

[lowep]

thanks

[Anders_HK]

This is exactly why I looking into a DSLR for shooting hand held that I find difficult with MFDB - though monopod helps. Question is will a newer high MP DSLR make a significant difference or would I be better off with an older generation DSLR like the 5D that I have found is a lot easier.

Lets throw in my 2 C of experience for consideration;

I found it difficult to hand hold Mamiya 645 AFD III with (28MP) Leaf Aptus 65, and in particular with a lens weighing around 800g or more.

Rolleiflex Hy6 with (80MP) Leaf AFi-II 12 requires more careful holding due to higher resolution, but when doing so I find it actually easier to hand hold it.

Why?

Hy6 is far more ergonomic than AFDIII (Hy6 has superior balance, the balance of AFD III was poor, same goes for handles and grip, Hy6 superior)

Hy6 is not lifted to eye level, due to waist level finder. Thus indeed waist level finder is important. I use the Hy6 to get sharp photos also at ISO 50 and less than 1/50 sec for 50mm lens, but... it takes holding it really really careful, steady, with deep breath!

Best regards,
Anders

[TheApprentice]

but... it takes holding it really really careful, steady, with deep breath!

Best regards,
Anders

This really reminds me of when I did target shooting whilst at school!
slow, controlled breathing and then pause halfway through breathing out. I could even fire the trigger between heartbeats after a few years of practice!
I tried the same technique with a 150-500 lens on my crop sensored camera (800mm equivalent) and could get handheld shots down to 1/200th with the stabilisation turned off!

[lowep]

i tried this:

Go to Edit-> Image size... on the menu.  To get half the megapixels, select Percent instead of Pixels and type 70.71% in the box (the square root of 0.5).  

no problem

Then you can adjust the physical dimensions of the print at the bottom before clicking on OK.

Erk..   as soon as I do this the number of pixels (in the top box) goes back to where it was again, so the only way I can manage to get the physical dimensions of the adjusted file to be the same size as the original (trying to do this to compare the visual effect of resampling in post on motion blur) is by changing the resolution (dpi), and when I do that I have a hard time seeing any difference in terms of motion blur between the original and resampled image -- so I suspect I am doing something wrong here though not sure what?

[Guillermo Luijk]

A higher resolution sensor (more Mpx) will be more demanding to get the best of it (i.e. sharp images at 100% crops) regarding:

- Trepidation blurring (handheld)
- Motion blurring (non-static subjects)
- Diffraction and DOF blurring
- Lens resolution blurring

but this does not mean that a higher resolution sensor will be any worse than a lower resolution sensor in the final image. The result with the higher pixel count sensor in the final image will always be equal or superior to the result obtained with the lower pixel count sensor.

The bad news for high resolution sensors come from the huge size of the files produced, which need more storage space both in memory cards and HD, and more processing power to run image applications on them.

Regards

[dreed]

If the problem associated with motion blur is a directly related to the field of view, doesn't that imply that cameras with smaller sensors (e.g APS-C) are thus more prone to this than those with larger (35mm)?

[Guillermo Luijk]

If the problem associated with motion blur is a directly related to the field of view, doesn't that imply that cameras with smaller sensors (e.g APS-C) are thus more prone to this than those with larger (35mm)?

An APS-C camera with a 50mm is more prone to blurring (both motion and handheld blur) than a FF camera with a 50mm lens. But... an APS-C camera with a 50mm lens produces a different picture than a FF camera with a 50mm lens  

An APS-C camera with a 50mm lens provides the same field of view as a FF camera with a 85mm lens, and both are equally prone to blur.

[theguywitha645d]

When viewing your images at 100%, then pixel resolution matters for motion blur.

When printing your images, then pixel density does not really change anything.

I handhold my Pentax 645D and I have not found it any more difficult than any other camera, film or digital. Since I end up presenting my images on prints or the web, I really do not worry about pixel resolution of the sensor.

[lowep]

Since I end up presenting my images on prints or the web, I really do not worry about pixel resolution of the sensor.

What if unfortunately it does make a difference - just as a contact printed 4x5 negative looks different to a same-sized print from a 35mm negative? Or something akin to the difference you see when you flick from an image with a broad color gamut to one with a narrow color gamut etc etc? Neither comparison is very adequate but nontheless I imagine that I do see a difference between images made with different sized sensors with different MP even when they are small prints or web images. Maybe this is just the famous "MFDB 3D" effect - ie being confused by the amount of money invested in the gear? The difficulty is to not only to pinpoint the nature of the difference but also to find the right way to describe the difference (which may be the same thing), as in any other research.