Resolution of Sensor AFFECTS Lens DOF

E

ectobuilder

Veteran
Holy cow I never knew this type of sensor-lens-dof behaviour even existed.

To give you some background, I own a Batis 25 F2. This lens of course has a digital DOF scale readout. And when I used this lens on the A7S (12MP), A7II (24MP), A7RII (42MP) the digital DOF scale reads differently.

For example when Aperture = F22 and the Lens = Infinite (all in FF mode), the DOF scale reads:
A7S (12MP) = 4.4ft to Infinite
A7II (24MP) = 6.1ft to Infinite
A7RII (42MP) = 7.1ft to Infinite


So I wrote to Zeiss to see what their answer would be. I received a response from one of their Customer Care associates, Bertram:

Thanks for your inquiry.
The resulting depth-of-field depends on the resolution of the camera sensor (when compared e.g. in 100% view afterwards). So with different camera bodies (pixel sizes), the display of the Batis lenses will show different depth-of-field data under the same conditions (aperture setting, distance setting).

By the way, the data shown on the display of the Batis lenses (focus distance, depth of field) come from the camera body an cannot be influenced by the lens itself.

For further details, please do not hesitate to contact us!

With best regards

Bertram Hönlinger
_________________
Carl Zeiss AG
Camera Lens Division
Kundencenter / Customer Care Center / PHO-VC
Kundensupport / Customer Support

Bertram Hönlinger

Carl-Zeiss-Straße 22 - 73447 Oberkochen
Telefon/Phone: +49 7364 20-6175
Fax: +49 7364 20-5216
Email: photoapplication@zeiss.de
http://www.zeiss.de/photo


So the higher the MP, the shallower the DOF. This pretty much has made up my mind to stick with all A7S or A7SII's as I want to keep all of my working DOF's the same.

For applications like steadicam, having a hyperfocal distance of 7.1ft to infinite is not cool! 4.4ft to infinite (A7S) is definitely better! That is a huge difference of almost 3ft of working distance!

Up until recently, this hasn't been an issue since most of the popular DSLR cameras ranged from 18-24MP. But now we have a range of 12MP all the way up to 50MP to choose from. So the issue is way more apparent now.

This is something to be aware of!
 
Last edited:
Interesting...

Interesting...

So can I assume then that when a manufacturer advertises a minimum focus distance for a lens, that will apply to
perhaps only certain camera bodies with specific mega pixel counts? Somehow that just doesn't sound right.
 
sbrenner said:
So can I assume then that when a manufacturer advertises a minimum focus distance for a lens, that will apply to
perhaps only certain camera bodies with specific mega pixel counts? Somehow that just doesn't sound right.
Click to expand...

No different than lux ratings for light fixtures i guess. Digital Sensor MP counts have stayed quite the same over the last half decade so i guess we've taken this aspect for granted.

So lens manufactuers need to state which MP count they are using to calculate DOF. This also shows the importance of having a digital DOF scale as it is more accurate.

I still want to hear from Sony to see if their answer is consistent with Zeiss.

But my brief tests in the field (at f22) seem to support this.

Now that we are dealing with 8K and beyond this appears to be an extra thing to consider.
 
Depth of field is a fuzzy topic.

It's not like everything from 4.4 feet onward is perfectly sharp. It's that everything in that range is acceptably sharp.

At one, laser-thin plane, everything is perfectly sharp. As you step away from that plane, things slowly get fuzzier. Perfect, one-dimensional points of focus become larger and larger circles --- "circles of confusion." If the circle of confusion is larger than your pixel, that's where you might draw the line. As you know, a 24-megapixel full-frame sensor has smaller pixels than a 12-megapixel full-frame sensor. So its tolerance for circles of confusion is less.

However, played back on an HD TV, or even on a 4K TV, it's not like the footage will look any different.
 
ectobuilder said:
So the higher the MP, the shallower the DOF. This pretty much has made up my mind to stick with all A7S or A7SII's as I want to keep all of my working DOF's the same.
Click to expand...
A word of caution. What he is saying is perfectly correct (depth of field does change with resolution) but that's not the same as "number of photosites" UNLESS you're thinking of images with full sensor resolution. This is the point of his comment "when compared e.g. in 100% view afterwards"

If you're using the three cameras you refer to for HD video, what's relevant is the SYSTEM resolution (1920x1080) - so regardless of sensor pixel count, if they are all in HD mode, they'll all have the same depth of field characteristics for same f stop, lens etc.

Move to 4K, and yes, then the depth of field will get shallower.

Think of it this way. Compared to HD, 4K will show finer detail - but by the same token it will also show up when the focus distance is slightly off more critically. That's another way of saying that (all else equal) you get less dof in 4K than HD - even when the basic photosite count in each case is the same.
sbrenner said:
So can I assume then that when a manufacturer advertises a minimum focus distance for a lens, that will apply to perhaps only certain camera bodies with specific mega pixel counts? Somehow that just doesn't sound right.
Click to expand...
No - because that focus distance applies to where the image will be "perfectly" sharp - the plane of focus. That won't change with any change in depth of field.
 
harddrive said:
If you're using the three cameras you refer to for HD video, what's relevant is the SYSTEM resolution (1920x1080) - so regardless of sensor pixel count, if they are all in HD mode, they'll all have the same depth of field characteristics for same f stop, lens etc.
Click to expand...

That would depend on how the HD is being derived? If derived from a 12MP vs 42MP sensor then the photosite count would matter correct? Hence the Batis lens is giving different DOF readings for each camera body.
 
Yep. What matters is the resolution of your final product. From the Zeiss response, the numbers the Batis lenses get from the Sony body are calibrated for a final product that has as many Mpix as the camera has in stills mode ("in 100% view afterwards"). I have "make a 4K version of my DoF calculator" in my to-do list, but it's very low on that list because, even if I shoot 4K in the near future, the final product will still be 1080p.
 
DOF does not change with sensor resolution.

What does change with that is whether you see something as being acceptably sharp. DOF calculations are all based on circles of confusion which is basically about a theoretical circle of light and the more resolution you can resolve, the smaller that point needs to be to look sharp.

The higher the resolution the more accurate focus needs to be when viewing at 100%.

We must all have had shots that when you chimp on the LCD look fine, but actually when you view them they are off. That's the resolution difference between the LCD and the full image.

The lenses are just being careful. From a point of absolute focus generally 2/3rds behind and 1/3rd in front are acceptable focus but if you have more resolution the actual distances would need to be smaller. There's a nice app called pCam where you can play with DOF calculations for different sensors and CoC.

It's not right to assume the DOF characteristics of a lens *change* with sensor resolution, that's fixed. But what you find acceptable does change.

cheers
Paul
 
(the lens obviously doesn't change, and I agree with everything you said, but my call is that your "what you find acceptable does change" is actually a change in the DoF: at 12 Mpix you consider 4.4ft-to-infinite to be in focus, at 42 Mpix it's only 7.1ft-to-infinite; in any case it's just a terminology discussion)
 
Samuel H said:
Yep. What matters is the resolution of your final product. From the Zeiss response, the numbers the Batis lenses get from the Sony body are calibrated for a final product that has as many Mpix as the camera has in stills mode ("in 100% view afterwards"). I have "make a 4K version of my DoF calculator" in my to-do list, but it's very low on that list because, even if I shoot 4K in the near future, the final product will still be 1080p.
Click to expand...

I'm super confused. So using the A7RII, recording in 1080p (FF) using the Zeiss Batis 25 @ F22 and focused to infinite. What DOF am I actually recording at? Keep in mind the lens does read 7.1ft to infinite.
 
Imagine shooting in SD on a really old camera (sensor A) .. the resolution is limited by the resolution of the camera.. so nearly everything looks 'in focus'

Actually it all looks as wooly as SD video.. really the DOF or sharpness of your lens or even the distance it is focussed at is not the liminting factor

So you could say that a lens presents 'unlimited' DOF (maybe focussed at 10ft at F4) for my SD camera (sensor A)

Now take that same lens and put it on a 40MP stills camera (sensor B) , take a photo, 10ft at F4, blow the photo up exhibition size, walk up to the photo really close and you will see that it is not actually in focus at 3ft, or at 30ft.. only from 8ft to 15ft.

So lens A set at 10ft, F4, has infinite DOF with sensor A and maybe 6 foot of DOF with sensor B.

Take that 40mp photo and have a look on your iphone and.. hey it all looks sharp all over.. so sensor B with lens A is sharp from close to infinity after all.. when viewed on a tiny screen.

So the 'acceptable DOF' changes with both input resolution (the capture medium) and output resolution (the screen)

Hey DOF even changes with 'viewing distance' - I do a shot of a head and Im focussed on the ears of the subject - Im peeping on my 23' imac from a foot away. Show me the same shot on my mates telly while I sit on the the sofa at the other end of the room.. and I cant see the focus error.. the whole image looks sharp!

CoC circle of confusion is the thing to understand.

S
 
This is sort of nit picking. I was just clarifying the statement that Resolution *changes* DOF which of course it doesn't.
Nothing *changes* the DOF of a physical lens.

DOF is a sliding scale, you have a focus 'point' and some area in front and behind that also 'looks' sharp. The area of what looks sharp can be expanded by stopping the lens down.

The key is 'looks' sharp, not *is* sharp.

If you could zoom in infinitely you'd find only the actual point of focus would be sharp.

So if you put that lens on a camera that does 640x480 then the size of a single pixel might not be enough to tell if it's sharp or not. The same lens on a 6400x4800 resolution has 100 times as many pixels (10x10 for 1 single pixel) and then you'd be able to see with your eyes if that area is sharp or not.

Does that make sense. That's why Zeiss said at 100% viewing.

If you took that 640x480 and printed it in a 6 x 4 cm rectangle and took that 6400x4800 image and printed it to the same size then they'd basically look the same. They'd 'look' equally sharp to your eyes.

So how you 'view' an image makes a difference too.

Zeiss is just helping out a bit by reminding you if you're pixel peeping then your focus ranges ought to be tighter.

Shooting for web delivery and theatrical release requires a different 'accuracy' for focus too.

cheers
Paul
 
morgan_moore said:
Imagine shooting in SD on a really old camera (sensor A) .. the resolution is limited by the resolution of the camera.. so nearly everything looks 'in focus'
Click to expand...

Just as i was typing pretty much the same thing you were too...

ectobuilder said:
I'm super confused. So using the A7RII, recording in 1080p (FF) using the Zeiss Batis 25 @ F22 and focused to infinite. What DOF am I actually recording at? Keep in mind the lens does read 7.1ft to infinite.
Click to expand...

As you know it doesn't suddenly snap into focus at 7.1 does it? 7.2 might be pretty sharp too. At what point is the level of sharpness acceptable to you?

Some arbitrary calculation is being used to give those numbers but you may be more or less sensitive to what you feel is sharp too.

And naturally where it will be ultimately viewed and what the subject matter is.

cheers
Paul
 
paulcurtis said:
Some arbitrary calculation is being used to give those numbers but you may be more or less sensitive to what you feel is sharp too.
Click to expand...

The calcs are not arbitary they a 'CoC' which is a formal calculation.. the size that a 'point' becomes (ie a full stop out of focus gets bigger) and how we can resolve that for a fixed print size at a fixed viewing distance.

I guess Zeiss are not arbitary!
https://en.wikipedia.org/wiki/Circle_of_confusion

S
 
[h=1]Resolution of Sensor AFFECTS Lens DOF[/h]
No.

DOF is solely affected by

-focal length,
-distance to subject,
-iris opening.

Also sensor size indirectly affects DOF as using a larger sensor forces you to change two of the direct elements that affect depth of field (you use a longer FL and you get closer to subject = decreasing the depth of field)

Resolving power of the sensor or lens? no. A lens has the exact same DOF on an 8 mp camera as it does on a 36mp camera, unless this topic is about a non-technical perception-related subject that I don't understand.
 
I'm with paulcurtis that it is a "personal" thing (not as much as arbitrary, but it does depend on how you watch your content, how good your vision is, and how much you dislike just-slightly-out-of-focus stuff).
 
You must log in or register to reply here.
Back
Top