background blur from different focal lengths - same image size

[jmahto]

Advertisement (gone after registration)

Not sure it is the right forum to share this. Move it to appropriate one as needed.

 

Long time ago I had read that the background blur is dependent only on the actual image size on the sensor (for the same aperture, independent of the focal length used). Therefore I did a test.

 

I shot with

1) 35mm FLE at f1.4

2) 50mm Lux at f1.4 : moving further back to get the same image size (of the bottle). Check the distance between bottle and the glass and the bottle size, both kept the same between the shots.

 

Note that background changes due to different FOV.

 

The bicycle blur is exactly the same in both shots. (of course the quality of the bokeh may differ due to different lens construction but the blur is same).

 

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[ebf]

Actually no. The size of the sensor helps, of course (more area == more light "ingested" by the sensor) but whenever you see "f/1.4" it actually means "focal length/1.4", and that's the size of your entrance pupil.

 

So, 35/1.4 = 25

 

and

 

50/1.4 = 35.71.

 

50mm lets you get much more light inside, giving more Bokeh.

 

The bokeh depends on your f/number and the focal length. (context: http://www.uscoles.com/depthoffield.html)

 

that's why if you try my old Elmar 135mm f/4, the bokeh you get from it is really, really good. 

 

Check it out

 

Cheers,

 

Eduardo.

[mjh]

Long time ago I had read that the background blur is dependent only on the actual image size on the sensor (for the same aperture, independent of the focal length used).

There is another factor you forgot to mention: the distance. If the distance between subject and camera is kept constant then the above assertion is wrong. However, if you adjust the distance to account for the difference in magnification of the subject – as you did in your comparison –, this renders it approximately true. The change of distance cancels out the effect of the entrance pupil (see Eduardo’s post), if only approximately.

[ebf]

There is another factor you forgot to mention: the distance. If the distance between subject and camera is kept constant then the above assertion is wrong. However, if you adjust the distance to account for the difference in magnification of the subject – as you did in your comparison –, this renders it approximately true. The change of distance cancels out the effect of the entrance pupil (see Eduardo’s post), if only approximately.

 

Exactly! Forgot about distance

[dpitt]

So, If I understand right. Suppose you want to take a picture of something very small with maximum DOF for Macro. You would be better off with a small sensor, because in general you can not close a lens further than F22 ?

Of course, there must be a point where reducing sensor size will also reduce quality too much to be helpful.

 

What sensor/lens would you prefer for the best macro shots?

[adan]

Bypassing dpitt for the moment, to the original question:

 

Another point is actual background (or foreground) size.

 

Note that in the original two pictures, the railings are equally blurred as a percentage of the width a "focusing" railing would be (thus confirming what jayant read). But on the 50mm shot, the railing as a whole is larger, and thus the size of the blur is larger, by an equal amount.

 

Meanwhile, the red stool is larger in the 35mm shot than in the 50mm shot (wide-angle quote-perspective-unquote) - and thus while the blur amount is about the same in both shots (as a percentage of the width of the stool), the blur area is larger in the 35mm shot. (by my rough count - 10 pixels wide in the 50mm shot, and 18 pixels in the 35mm shot).

 

Now back to dpitt's question:

 

Are you talking about simply using a smaller sensor with the same lens? Or changing lenses to get the same equivalent view on the smaller sensor?

 

For the sake of simple math, suppose you want to use a 50mm macro on full-frame (diagonal 43mm), and then take a picture on a sensor with a diagonal of 1/4 that (10.75mm). Are you talking about using the same 50mm lens (equivalent of a 200mm on that small sensor), or using a "50mm equivalent" lens of 12.5mm?

[dpitt]

Advertisement (gone after registration)

Andy,

 

I was thinking about taking a particular shot with the same FOV as the OP did. But in case of Macro you sometimes want to increase DOF as much as possible without losing quality. So this will probably mean taking other lenses on other sensors, or at least increasing the distance to the subject.

Although e.g. I have some R lenses that I could use on FF (Canon 5d) , APS-H (DMR) . APS-C (Canon cropped DSLR or other 1,5 or 1,6 cropped sensor), and 4/3 sensor with crop factor 2.

 

If all these sensors would have the same build quality, would it be best to use the smallest?  

[mjh]

In theory you could just as well use a larger sensor and stop down to achieve the desired DOF, increasing ISO if necessary. A smaller sensor would achieve the same result at a lower ISO setting but its smaller pixels would receive less light rendering this apparent advantage void. But then you cannot stop down indefinitely and in practice you may be better off using a camera with a smaller sensor. In fact I have done many product shots with a small (but high quality) point-and-shoot and the results were good enough to be printed in magazines. Of course you have to consider the effects of diffraction, but in that regard there is no real difference between large and small sensors: the same DOF at the same angle of viewing = the same amount of diffraction blur.

[jmahto]

.....

If all these sensors would have the same build quality, would it be best to use the smallest?  

 

It seems so. I came across this macro pic in flickr taken by nikon coolpix. Assuming EXIF data is correct this looks good from a small P&S. I don't think I can get this result from my Leica setup.

https://flic.kr/p/rpdXiM

[hiepphotog]

It seems so. I came across this macro pic in flickr taken by nikon coolpix. Assuming EXIF data is correct this looks good from a small P&S. I don't think I can get this result from my Leica setup.

https://flic.kr/p/rpdXiM

 

I don't think the whole Leica thing is for macro. I have seen more impressive macro shots using FF cameras. For macro, high speed flash or very long focal length to snipe is more important. The problem is, small sensor can't never have the same pixel quality and overall quality as a FF due to much denser pixel pitch (to get a decent MP count).

[Jean-Michel]

Actually no. The size of the sensor helps, of course (more area == more light "ingested" by the sensor) but whenever you see "f/1.4" it actually means "focal length/1.4", and that's the size of your entrance pupil.

 

So, 35/1.4 = 25

 

and

 

50/1.4 = 35.71.

 

50mm lets you get much more light inside, giving more Bokeh.

 

The bokeh depends on your f/number and the focal length. (context: http://www.uscoles.com/depthoffield.html)

 

that's why if you try my old Elmar 135mm f/4, the bokeh you get from it is really, really good. 

 

Check it out

 

Cheers,

 

Eduardo.

Hi,

The calculation is correct but the statement that the 50 lets in more light than the 35 at f/1.4, or any aperture for that mater is incorrect. Both  allow exactly the same amount of light to get in for the same time, same exposure. To fill a container with the same amount of water in the same amount of time from containers with  25 and 35.71 mm diameter holes you would need to put them up at different heights, such as 35 and 50 mm or cm. remember the inverse square law for light intensity. 

Jean-Michel

[jmahto]

Hi,

The calculation is correct but the statement that the 50 lets in more light than the 35 at f/1.4, or any aperture for that mater is incorrect. Both  allow exactly the same amount of light to get in for the same time, same exposure. To fill a container with the same amount of water in the same amount of time from containers with  25 and 35.71 mm diameter holes you would need to put them up at different heights, such as 35 and 50 mm or cm. remember the inverse square law for light intensity. 

Jean-Michel

Yes, it is like shape of a cone. The longer one has wider opening (for the same cone angle) but the same light enters in both.

[ebf]

Thanks guys. I'm still learning this

[mjh]

To put it differently, the larger entrance pupil lets in more light, but then with the longer focal length the sensor is gathering light from a smaller area. The net result is that the total amount of light hitting the sensor stays the same.

 

Having said that, this thread isn’t about light but about background blur, and the latter does depend on the entrance pupil.

[ebf]

Awesome. That's what I was trying to explain. 35mm has larger DOF than 50mm, with the same aperture.

 

And *that* affects bokeh. Thanks @mjh

[CheshireCat]

Long time ago I had read that the background blur is dependent only on the actual image size on the sensor (for the same aperture, independent of the focal length used).

 

This is a wrong generalization.

I recommend reading this:

http://toothwalker.org/optics/dof.html

[jmahto]

This is a wrong generalization.

I recommend reading this:

http://toothwalker.org/optics/dof.html

 

Excellent article. Thanks for sharing. However, if you look at figure 3 and 4 in the article, it does show that for the same image magnification from 28mm and 100mm, both at f4, the relative blur is the same (not the absolute blur). I didn't make the distinction between absolute and relative blur. 

 

The article also goes on to say "The equality of the relative blur of a distant background always holds (on the usual condition of identical subject framing and F-numbers), but the equality of the DOF breaks down when the object distances are no longer small with respect to the hyperfocal distance". I guess this is what Michael referred to in his reply above when he said that "The change of distance cancels out the effect of the entrance pupil (see Eduardo’s post), if only approximately."

 

Therefore it seems that this equality may be approximate. In any case, this result is important to me for a simple application. When I am trying to shoot people with 35 or 50mm, I try to keep their image size same by adjusting my distance from them. In that case I do not expect to get more blur from 50mm @1.4 than 35mm @1.4.