Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

A Lens is a Lens is a Lens

Stephen Williams said:
Hi Alexander,

Whilst what you say is true, if you look at the images produced by those different lenses the size on the sensor will be different. When you adjust the camera distance so that the size of the image matches at the same F stop the DOF will be identical.

The reason for this is lens thickness, DOF is calculated from the front nodal point of the lens not the film plane. Simple DOF tables assume a 'thin' lens. A Cooke 20-100 is clearly not thin at all!

As I said in the previous post F stop & Magnification are the only 2 components involved.

Best,

Stephen
Click to expand...

We are talking at slightly cross purpose.

I am specifically calculating and describing DoF.

You are talking about the practical use of DoF with Field of View as a primary criteria- which is entirely sensible and a very valid point.

But we aren't talking about the same exact thing.
 
Thus the very point of my video!
 
Alexander Ibrahim said:
I am specifically calculating and describing DoF.
.
Click to expand...

You calculate DOF from the film plane, which is incorrect & the reason for your confusion.
This is the reason that different 50mm lenses will have slightly different DOF's.
Best,

Stephen
 
I believe that Ed Watkins wrote an iphone application - isee4K - to help with DOF. Haven't used it myself yet but read on a number of threads that it works great. Hope this helps.
 
Alexander Ibrahim said:
The only relevant factors when determining depth of field are:

the focal length of a lens
the current f-stop setting
the focal distance.
Click to expand...

Hi Alexander:

Isn't the sensor size an equally important factor in determining DOF?

You also mentioned but seemed to downplay the importance of the Circle of Confusion. The CoC used for determining DOF on 4K footage intended for 4K digital projection on a 20 meter wide theater screen and viewed from a distance of 10 meters with 20/20 vision must be much smaller than for footage intended for 1080i cable broadcast viewed on 27" LCD's at 3 meters.

1. Focal Length
2. F-Stop
3. Focal Distance

4. Sensor Size - Crop Factor
5. Projection or Display Size
6. Viewing Distance
7. Eyesight Quality

Also, shouldn't f5.6 and f2.8 always be written as f/5.6 and f/2.8 as in focal length of lens divided by f-stop?

-michael zaletel
 
michael zaletel said:
Also, shouldn't f5.6 and f2.8 always be written as f/5.6 and f/2.8 as in focal length of lens divided by f-stop?

-michael zaletel
Click to expand...

stops are fractions, thats why the no goes up as the hole gets smaller.

1/2,1/2.8,1/4,1/5.6,1/8 etc
 
michael zaletel said:
Hi Alexander:

Isn't the sensor size an equally important factor in determining DOF?
Click to expand...

Yes and no.

Sensor size has nothing at all to do with DoF on a given lens.

A 25mm lens has the same DoF characteristics no matter what camera system you manage to mount it to.

What happens is that the different sensor sizes "crop" the image differently.

So at about 5 feet focal distance a 25mm lens will render:

A full shot or a master in 65mm photography
A medium shot in S35 3-perf photography
a close shot in 16mm or 2/3" photography.

I think the DoF on a 25mm at 5 feet at f/4 is about two and a half feet.

So you have the same DoF, but three different frame sizes.

Its up to you as the DP to decide what DoF is appropriate to these different shots in order to tell your story.

For what its worth I never approach this task in the way we re talking about.

Usually I start by deciding what type of DoF I want ahead of time and what types of shots I'll be shooting. Then I pick a camera system, lenses and a lighting level that is appropriate.

So, for example on dramatic work, I typically work to achieve a fairly shallow DoF, with lighting set so that I can shoot at f/2.8 or f/4. I usually pick an S35 size sensor- although I am happy with APS-C and the like. I plan on shooting most close shots with a 75mm (if I am using Cooke) or 85mm (Zeiss) lens, 35 for most two shots and something around an 18 for static masters on interiors. 25mm or so for moving masters.

You also mentioned but seemed to downplay the importance of the Circle of Confusion. The CoC used for determining DOF on 4K footage intended for 4K digital projection on a 20 meter wide theater screen and viewed from a distance of 10 meters with 20/20 vision must be much smaller than for footage intended for 1080i cable broadcast viewed on 27" LCD's at 3 meters.
Click to expand...

To all this I say, meh.

CoC is not a hugely important factor in daily work. Oh, it can sure matter in selecting a camera system, but its not something I have to think about as I work.

The way of thinking you are using above is sort of outdated. Instead insternational standards relate CoC to the sensor size. Specifically to the diagonal of the sensor or film frame. The current standard is 1/1000 of the diagonal. For 3-perf S35 film this gives a CoC of 0.028mm or 0.0011 inches.

Most DoF tables, including the DoF tables in the ASC manual and Cooke's tables for the 5/i lenses have a CoC of 0.001 inches or 0.0254mm

In other words they exceed the current standards.

Manufacturers have been pushing changing to change the standard for CoC to 1/1500 of the sensor diagonal. For S35 3-perf that would be 0.019mm or 0.0007 inches.

If you want that level of precision, then instead of looking at your actual stop on the charts, look one stop faster. So, if you are shooting t/2.8, look at the entry for t/2.

As an aside, I recently read that Zeiss supports using 1/3000 of the image diagonal as the standard for CoC. I think that's absurd for motion pictures, but it may have some merit for still imagery that will be examined more carefully.

Its not a huge deal. Personally I find that just using the manufacturer's tables are accurate enough for 1080p or 2K motion pictures projected at up to 100 feet image size.
1. Focal Length
2. F-Stop
3. Focal Distance

4. Sensor Size - Crop Factor
5. Projection or Display Size
6. Viewing Distance
7. Eyesight Quality
Click to expand...

No, a thousand times NO.

Only the first three items I noted are relevant to depth of field.

The rest have nothing to do with depth of field.

Your #4 has some merit as a factor in pre-production, but you sensor has nothing to do with your DoF.

As to the rest, mentioning them is only a nod to the fact that CoC is a subjective determination, and that these are the sorts of things you should consider when determining what CoC you would like to use.

Also, shouldn't f5.6 and f2.8 always be written as f/5.6 and f/2.8 as in focal length of lens divided by f-stop?
Click to expand...

Well yes you should write f numbers that way. While we are at it t numbers should also be written that way.

Your reasoning is wrong.

The f number, or N, is given by the relation:

N = (focal length of lens)/(diameter of entrance pupil)

So for an 85mm lens with an entrance pupil of 60.71:

N = 85mm/60.71mm
N = 85/60.71 * mm/mm
N = 85/60.71 * 1
N = 1.4

Notice that the last number has no unit type? That is because the millimeters cancelled out in the arithmetic. The result of our calculation is a dimensionless number, 1.4.

The mathematically correct way to write and say this is, "The f number is 1.4"

We write f/1.4 by convention. Presumably this is so that communication is less ambiguous. Its a form of shorthand, and should be read aloud as "F number 1.4" Of course you never hear that. You hear people say, "f 1.4" In fact I most often hear people say, "f one four."

All of these are sloppy use of jargon, but we have something to be thankful for. Just imagine the confusion if all you heard on sets was someone yelling, "4?"

When I write about photography I usually write how I speak, "f1.4" kind of all run together. Its lazy and literally unconventional.

I don't care so long as I am understood- hence my usual.

On a side note, your reasoning for the f number notation is wrong because that formula (focal length)/(f number) gives the diameter of the entrance pupil.
 
Stephen Williams said:
You calculate DOF from the film plane, which is incorrect & the reason for your confusion.
This is the reason that different 50mm lenses will have slightly different DOF's.
Best,

Stephen
Click to expand...

Sorry but I have no confusion at all.

The point I am trying to make is that the difference between using the lens front nodal plane and the focal plane for DoF determination is negligible.

It remains negligible until the focal distance gets close to the focal length of the lens.

So essentially, only macro lenses can be affected by this distinction. In fact, many macro lenses can't focus close enough for this distinction to matter.
 
I did some DOF tests once with a receding series of Seimens Stars and found that I had to shift the focus point on the zooms to get the range of what was in focus to line-up with other lenses due to the difference in entrance/exit pupil, nodal point, whatever you call it. However, I did not feel the actual range of usable focus was much different between lenses in any practical sense, just that the physical length of the zoom was a factor in determining where the range began and ended. Now obviously, in theory, if you had to shift the focus a bit closer to compensate, the depth of field would be dropping off.
 
Alexander Ibrahim said:
Sorry but I have no confusion at all.


So essentially, only macro lenses can be affected by this distinction. In fact, many macro lenses can't focus close enough for this distinction to matter.
Click to expand...

With all due respect the difference will be far greater with a zoom due to its 'Thickness'.

If you swap a 100mm Cooke S4 for a Cooke 20-100 set to 100m & shoot a close up of a face, the difference in shot size fairly obvious, move the camera about 5 inches and it lines up again.

Best Stephen
 
David Mullen ASC said:
I did some DOF tests once with a receding series of Seimens Stars and found that I had to shift the focus point on the zooms to get the range of what was in focus to line-up with other lenses due to the difference in entrance/exit pupil, nodal point, whatever you call it. However, I did not feel the actual range of usable focus was much different between lenses in any practical sense, just that the physical length of the zoom was a factor in determining where the range began and ended. Now obviously, in theory, if you had to shift the focus a bit closer to compensate, the depth of field would be dropping off.
Click to expand...

Hi David,

The DOF will be identical if you move the camera with the zoom lens back a few inches to compensate for the longer physical length of the zoom lens.

Best Stephen
 
Stephen Williams said:
With all due respect the difference will be far greater with a zoom due to its 'Thickness'.

If you swap a 100mm Cooke S4 for a Cooke 20-100 set to 100m & shoot a close up of a face, the difference in shot size fairly obvious, move the camera about 5 inches and it lines up again.

Best Stephen
Click to expand...

Well I've never used the Cooke 20-100, so it may be a thick lens.

I've never had this problem on almost new Angeniuex zooms, which are the only cine zooms I've worked with. (16-42 & 24-290, both on different RED cameras)

Then again, maybe I haven't looked closely enough.

It may also be an artifact of how my ACs work with these lenses. They set focus marks on the handwheel by zooming in using pixel for pixel modes. Then we tape (or laser) the marks out and check the lens barrel for confirmation, mostly for the camera log. The barrel marks are for "sanity" checks, If we tape out 15 feet, and the barrel reads 10 feet there is a huge problem, so we start over. The tape (or laser) is marked more precisely of course, so tape/laser information is what goes in the log.

In other words, we treat the lens barrel markings as the least reliable thing in the system.

Thinking of LDS or /i lenses, how does the information they report line up with what your AC's are taping and what is marked on the barrel. I have no real working experience with those features. Producers and VFX supers have asked me about replacing camera log data with lens metadata.

I am reminded to ask everyone... has Angeniuex started making their lenses with the /i interface yet? I remember reading an announcement, but I have not seen any glass with the feature.
 
Alexander Ibrahim said:
Well I've never used the Cooke 20-100, so it may be a thick lens.

I've never had this problem on almost new Angeniuex zooms, which are the only cine zooms I've worked with. (16-42 & 24-290, both on different RED cameras)

.
Click to expand...

Hi,

The 24-290 is a very thick lens (More than a foot long) , if you swap it for a prime matching the focal length the image size will NOT be the same, you will need to move the 24-290backwards if you want the images sizes to line up. It's a very easy test to do.

It's not a problem just a law of physics, the basis for FOV & DOF calculations

Best Stephen
 
Stephen Williams said:
Hi,

The 24-290 is a very thick lens (More than a foot long) , if you swap it for a prime matching the focal length the image size will NOT be the same, you will need to move the 24-290backwards if you want the images sizes to line up. It's a very easy test to do.

It's not a problem just a law of physics, the basis for FOV & DOF calculations

Best Stephen
Click to expand...

Yeah, you are right. The math supports that interpretation.

That said, I never do that in practice.

In practice I'd set the prime up and not the image size, then I'd make the zoom match what the prime reported from that position- and remark the lens barrel if needed.

In other words, I'm not trusting the barrel markings.

Let me put that differently.

In practice, a given focal length lens is not one where image size matches some standard when distance to subject is measured from the lens front nodal point, but rather where image size matches when measured from the focal plane.

I am going to ask Angeniuex what they calculate for the lens "thickness" of their various lenses. Provided I ask the right person, it could be interesting.
 
Alexander Ibrahim said:
Yeah, you are right. The math supports that interpretation.

That said, I never do that in practice.

In practice I'd set the prime up and not the image size, then I'd make the zoom match what the prime reported from that position- and remark the lens barrel if needed.

In other words, I'm not trusting the barrel markings.

Let me put that differently.

In practice, a given focal length lens is not one where image size matches some standard when distance to subject is measured from the lens front nodal point, but rather where image size matches when measured from the focal plane.

I am going to ask Angeniuex what they calculate for the lens "thickness" of their various lenses. Provided I ask the right person, it could be interesting.
Click to expand...

Hi,

If you are matching with minitures & 3d the angle of view will be wrong.
The marks on the barel are almost certainly correct!

Angeniuex will give you the information, the front nodel point will almost certainly move as the zoom changes focal length. I have known lenses where the nodel point moves more than the length of the lens.

Best,

Stephen
 
You must log in or register to reply here.
Back
Top