Depth of Field has nothing to do with image sensor size...

Thanks for the video. It's a good tutorial for those not familiar with DoF.

But it's not absurd to think that image sensor size effects the DoF in a film. Notice I did not say cause DoF in a film. It may not be one of the three factors in the formula, but the sensor size does influence the lens choices you make regarding the field of view. Thus these changes in lenses to compensate for FoV reasons changes the focal length variable in the DoF equation.

I know what you are saying, but we'd be insane to think an image sensor does not have an impact on the overall depth of field in the image, despite whether or not it's a direct part of the equation.

Image sensors dictate field of view, field of view dictates lens selection, and focal length is a very important variable in the depth of field equation.
Click to expand...
 
Last edited:
I have to agree with Ryan. At the end of your vid you touched on the practical reason 'why' - when you said you move back to get the same framing
but I have found this seldom happens - in practice you dont move back ... you switch focal lengths.

A better analogy for the end of your vid would have been to show the 5D and the 7D showing the same field of view. 5D with a 50mm and 7D with a 35mm.
Thus to obtain the same shot (field of view) the camera with the larger sensor must use a longer focal length which in turn means it has a shallower depth of field.
 
So, question: if a camcorder, with it's camcorder framing, had a 1.5 lens, would that successfully combat the distance problem to equal a depth of field near what 3.5 is on a DSLR? Just wondering, because as it is, the $800 Panny HDC-TM700K has the lowest aperture of any modern camcorder; 1.5. I'm considering that this would be equal to a 3.5 kit lens in depth of field, once framing was applied. And if this is correct, good-bye DSLR, hello Panny.
 
J Davis said:
I have to agree with Ryan. At the end of your vid you touched on the practical reason 'why' - when you said you move back to get the same framing
but I have found this seldom happens - in practice you dont move back ... you switch focal lengths.

A better analogy for the end of your vid would have been to show the 5D and the 7D showing the same field of view. 5D with a 50mm and 7D with a 35mm.
Thus to obtain the same shot (field of view) the camera with the larger sensor must use a longer focal length which in turn means it has a shallower depth of field.
Click to expand...

I wish i could have done that but then the video would be even longer and many people can't stand to watch a video that's already over 10 min. But yes i agree... like I said in the video. Larger image sensor makes it easier to acheive shallow DOF, cuz of more lens and framing options you have, BUT it doesnt physicaly change it.

To see that sensor size doesn't change the DOF do tests with a a pin hole camera. Or even better an old 35mm film camera you don't care about, remove the back & replace it with paper to view the projected image.. make the paper smaller or larger & the cirlce of confusion stays the same... then change the 3 factors I mention and you will see it change. Don't get me wrong. I'm not saying larger sensors aren't better if you want to achieve shallow DOF, but they are not what actually causes the changes.
 
Ryan Patrick O'Hara said:
I know what you are saying, but we'd be insane to think an image sensor does not have an impact on the overall depth of field in the image, despite whether or not it's a direct part of the equation.
Click to expand...

I got the ASC Toland app for my iPhone. One of the things you can do with it, is fix the FOV, then select lenses. The required distances and resulting DoF for a given Circle of Confusion based on the camera/film used, are 'automagically' calculated... (maybe I should suggest an 'infrared' calculation option from that other thread...).

The app has a little view finder icon, pretty simplistic, but it does give a feel for how things either change when the lens is changed, and FOV is not locked, vs. locking the FOV and changing lenses in the required distances.

One feature related to DoF, is the ability to record 'marks' so you can see if you need to adjust DoF to keep things in 'focus' if they either move, or there are multiple things in the shot that need to be 'sharp', for some value of 'sharp'...


No.. I don't get any kickbacks... just it seems like a good tool to have around...

Website for Chemical Wedding the softwarehaus... (and no, the illuminati won't be knocking on your door anytime soon...)

http://www.chemicalwedding.tv/home.html
 
So basically, does a consumer camcorder has the exact same depth of field - actually more depth of field since some consumer camcorders have much lower apertures than DSLR lenses - as a DSLR, it's just that whatever image you capture from the camcorder, it will be cropped to about half the image - literally meaning, neither full frame cameras or crop sensors are any different, it's just that they cause you to position your shot differently and/or have possibly more-or-less depth of field, which is why productions that want more depth of field choose full frame?

But now that camcorders are built with extremely fast lenses such as 1.5 - lenses that previously were only useable for the people who opted for full frame cameras - I want to ask again; is there now any difference between a camcorder with a 1.5 lens and a DSLR with a 3.5 lens, beyond being smart enough to use a ND filter with the camcorder? Because using my 1.8, I just did a couple shots testing the difference between 1.8 and 3.5 (which would probably be the perception difference between a camcorder frame and a DSLR crop frame), and to me, both apertures look pretty darn cinematic. If a camcorder that had a 1.5 lens translated into having a 3.5 depth of field illusion once the framing distance was accounted for, then I'd gladly use a camcorder for my films.

But if the translation is more than that (I wish someone would make a chart!) then I'm not sure. Please, if anyone has access to a Panny TM700K with it's 1.5 lens, 46mm filter mount, and 3MOS sensor, I'd love to see a rough experiment to see at what distance the same depth of field can be accomplished on it as on a crop DSLR that's set, say, at 2.8 apeture, a common cinematic standard. If you try this, again, remember to get your hands on a 46mm ND! And honestly, if I can, I may do this test pretty soon if no one else does.
 
Last edited:
People overcomplicate this so much! For a smaller sensor, you will need a wider angle lens for a similar FOV as a larger sensor with a less-wide lens. Compare depth of field in the wider angle lens to the depth of field in less wide lens and you will see that the wider lens shows greater range of depth of field focus. So if you want to get the same field of view from two cameras right next to each other and one is large sensor and one is smaller sensor, then the smaller will need a wider lens which will result in a greater DOF. This is the functional reality of shooting in the real world. Again, I suggest you watch the video I did for our blog that I mentioned in Post #5. By the end I show three cameras, side by side, with three different sensor sizes but lensed to the same FOV and the same aperture. The difference in Depth of Field is obvious.

There are many ways to obfuscate and confuse this, but in the end that stuff just doesn't matter. The video in Post #1 gives suggestions about workarounds, but it does not address the actual question at hand.
 
Snapper123 said:
If a camcorder that had a 1.5 lens translated into having a 3.5 depth of field illusion once the framing distance was accounted for, then I'd gladly use a camcorder for my films.

But if the translation is more than that (I wish someone would make a chart!) then I'm not sure. Please, if anyone has access to a Panny TM700K with it's 1.5 lens, 46mm filter mount, and 3MOS sensor, I'd love to see a rough experiment to see at what distance the same depth of field can be accomplished on it as on a crop DSLR that's set, say, at 2.8 apeture, a common cinematic standard. If you try this, again, remember to get your hands on a 46mm ND! And honestly, if I can, I may do this test pretty soon if no one else does.
Click to expand...

Since my Red Ryder BB-Gun with a fake wood stock and a Toland Calculator App doesn't have a 1/4" sensor digital camera setup, I give a couple of examples from a 1/3" sensor.

Circle of Confusion for the 1/3 sensor, 0.00043, F-stop 1.4

FoV such that at the plane of sharp focus, the frame is 2' 11"x1' 8" (WxH) (Closeup, head top of shoulder)

For a 5mm lens:

subject(principle focus): 3'
near: 1' 11"
far: 6' 11"

For a 12mm lens:

subject(principle focus): 7' 2"
near: 5' 10"
far: 9' 5"

For my GH-1

Circle of confusion: 0.00059, same approximate FoV, 2' 11"x1' 8" using a 14mm lens.

subject(principle focus): 2' 4"
near: 2' 2"
far: 2' 7"

Lock FoV, use a 50mm lens.

subject(principle focus): 8' 5"
near: 8' 3"
far: 8' 7"

----------------------

Estimate what lense would be required on the 1/3 sensor camera to match the GH-1 @ 50mm, that is a DoF of about 5".

Ok, even at an 135mm lens length with FoV of 3' 2" x 1' 9" (WxH) (approximately the same as above... sort of...)

subject(principle focus): 88' 5"
near: 86' 5"
far: 90' 6"
---------------------------

Notes:

The GH-1 14-140 does not open up wider than 4, so this data is just for theoretical approximations.

I don't have the TM700K, so I don't know if it has a zoom capability to 135mm...

--------------------------

The main point being, the circle of confusion, the required FoV, and the resulting 'magnification' required factor into how DoF is preceived...

And as usual, I may have botched some of the calculations... so sue me... but the idea is that for a small sensor to match the 'big' sensor, one has to use a longer focal length, and get further from the subject, to match DoF, for a given FoV, and a given Circle of Confusion which is related to the sensor characteristics, size being one of them.
 
All of this confusion! Formulas and guesswork and sensors, OH MY! :)

I worked a shoot earlier this year and the director and I had a lot of discussion and he and I could never seem to figure out what he wanted in the way of FOV and stuff. He is a visual guy. It was a lot of trial and error and for the most part, even though he got what he was looking for, it took forever changing out lenses and moving around and such. I worked with the same director a month ago but this time the director bought one of those "director's viewfinders". He simply walked around and yelled out what lens he wanted beforehand. I made a list and as we moved I swapped lenses for the scenes and we kept moving right along.

I suggest people make it easy and just get one of those fancy viewfinders. They really make things go a whole lot smoother and faster.
 
With the Panny camera, the specifications describe it as 3.45-41.4mm, but then describes it as 35.0-420mm as the 35mm equivalent, I'm guessing that the second description is what the focal length is once the crop is applied, and the first is what the lens actually is (like how a 50mm is like 85mm on the DSLR's.) - and it describes the aperture as being 1.5 - 2.8 : at least on the DSLR, the difference between 1.5 and 5.6 is extreme; 5.6 being what 50mm is on a kit lens, just as an example.

So I'm guessing with this information, that the 35mm focal length on the Panny will have roughly the same depth of field as a 3mm 1.5 lens ??? And that 420mm would have roughly the same depth of field as a 40mm 2.8? Surely that's not right...
 
*sigh* I hate to break it to you all, but the sensor size DOES effect the DoF IF the resolution is the same and the pixels are different. It's not just a simple crop of the lens information, but a magnification due to the pixel size differences. Sure, if you took a pinhole camera and cropped a portion of the sensor and looked at the same thing, but that is not what you are doing with these cameras. Because of the pixel differences and the magnification caused by that, the Circle of Confusion is different and therefore, you do get different DoF.

So when someone has a FF35 that shoots 1080p and a APS-C that shoots 1080p, the DoF is NOT the same for the same lens, distance, F-stop. Any calculator will show you this because the CoC is different.
 
polcan99 said:
So many times I hear people argue and talk about what affects depth of field and so many times I hear completely absurd things. Bottom line depth of field has nothing to do with image sensor size... here is a simple video that explains it:

http://www.youtube.com/watch?v=L0n4pyhx7Dk
Click to expand...

I recently came across your introductory filmmaking tutorials and I am very impressed by how good they are. You have a talent for explaining things audiovisually, putting images and narration in a very coordinated way, so when the viewers are looking at something on the video, the narration is precisely talking about that very thing they are looking at. In your tutorials you also anticipate the questions that the viewers are asking themselves, which is very important in eLearning material where you don't have interactivity and the possibility to interrupt and ask.

IMHO, instead of doing commercials and that kind of stuff, you should be doing eLearning audiovisual material for the web (provided you can make a living from that). Anybody can do commercials, or indie films, but it is not easy to do good eLearning tutorials. They require the skills of a good professor combined with the skills of a good audiovisual narrator, and you seems to be good at combining both (Barry Green, Alister Chapman, Philip Bloom are names that come to my mind as good examples of talented people in this respect).

Mitch Gross' video (post #5) is very helpful too, but I think it only works with people who already know about this stuff, while your video on DoF is more for novices and people not familiar with this notions.

I specially like your recent tutorial on framing:


so thanks for the tutorials and keep that excellent work going.
 
Last edited:
5450392846_073253d217_b.jpg



A shallow DOF sample from my 1/3" Canon GL2. The vase is about the size of a human head. I was trying to see just how shallow I could get with this camera. F=60mm @ f2 from about 8' away. Only the middle two or three inches of the vase are actually in full focus. The hardest part was judging where image focal plane was using the camera EVF. Not enough resolution for critical judgement. If I wanted to shoot a two shot around the table though, a nice shallow dof fall off would not have been possible. With head and shoulders closeups it can be done.
 
Last edited:
Thank you very much for that comment. Im happy to know that you think I have talent for making these tutorials. Makes me want to do more of this free work to help others out. Thanks!
 
Indeed FOV, Avaialable Fstop whatever, they all work hand in hand - its a circular argument

Like wattage is a function of current and voltage

What ever way you look at it you cant get a 645 at 2.8 look with a 2/3 or 4/3 chip because you cant buy F.05 lenses for those systems - which is what you would need

everyone knows that it is hard to get thin DOF with a small chip camera, and deep dof with a large chip camera using currently avaialbe lenses -surely it is that simple

everyone surely knows too that a at f8 5.6 etc often a similar look can be created across chip sizes

eg a 7d/F3/Red will look similar to a 5d if you open up a couple of stops compared to your 5d (and choose a lens of similar FOV ie shorter)

Look at this.. (same FOV)

confusion of an OOF area with similar FOV

intro.jpg


You cant get the narrow look with the D80 (S35) that you get with the blad (double vistavision)- because you cant open up enough, you can however create a deeper look with both systms

S
 
Last edited:
There is an aspect of DOF that comes into play with smaller formats and higher pixel densities. That is critical COC, or how small the smallest sharp focus point is. A 1080p 1/3" chip has a much smaller critical COC than say an AF100 would at 1080p. Effectively at the same focal lengths you will get a more rapid fall off in DOF with the smaller sensor, so this can compensate to some extent for the shorter focal lengths required for equal FOV. Where the difference is hard to overcome is with wide shots against close backgrounds. A larger sensor gives a more gradual and easier to control fall off.
 
You must log in or register to reply here.
Back
Top