Micro Four Thirds DOF Chart of Doom

backwheelbates

I just posted some research into comparing DOF for a few MFT lenses. So, I thought Id share the info so others can try out their own setups.

"I was doing some thinking about lenses and depth of field and I came up with this comparison chart for a few MFT lenses. What I really wanted to find out was, what are the limitations of shallow depth of field based on aperture, focal length, and most importantly distance from the camera. As well as how can I squeeze my lenses for as much shallow DOF as they can deliver. As I suspected the depth of field become less shallow as you move away from the camera. This is all theoretical, but I found the chart to be really interesting to compare lenses to see where the dramatic DOF fun starts and ends. Id be curious if this actually holds up in the real world."

http://www.eric-bates.com/?p=2024

mft_dofChart

balazer

I take it your circle of confusion size is selected to match the resolution of the GH2 in stills mode. Video resolutions will have greater depth of field.

Depth of field comparisons for different lenses are more intuitive when made in terms of the aperture diameter instead of the f-number.

rapsoe

@balazer and all, talking about CoC, is there a consensus on the CoC to use on a GH2 shooting at 1920x1080? (given that the GH2's sensor width is 19mm, I'm currently using 23 microns, but would be nice to know what others PVers are using).

Psyco

Isn't the DoF the same for every focal length, as long as you keep the subject in focus the same size and use the same F-stop?

So, if you want shallow DoF you need a faster lens or let your subject appear bigger.

Kihlian

Really interesting, also this link:

http://www.dofmaster.com/dofjs.html

where you can calculate the DOF. A 20mm lens stopped at f1.7 (the pancake) focal point at 3 meters gives a total DOF of 1,17m, the same shoot (3 meters f 1,7) with a 25mm lens (the leica summilux) gives a DOF of 0,73 m. I have both lens and can from mi observations rougtly confirm this. Sometime a deep DOF is needed when shooting.

backwheelbates

Thanks for the comments everyone.

By the way, all the dimensions are in meters. Both the DOF ranges (vertical) and distance from the camera (horizontal).

@balazer, If video recording is using the same portion of the image sensor as still image capture. I don't see how the DOF would be different. As far as I know, the circle of confusion is not resolution dependent. I get the idea of perceptual differences of DOF at different print/screen sizes and resolutions. Can you explain what you mean?

@Psyco, not sure about your first comment. But in terms of shallow DOF, thats the conclusion I made. If you want shallow DOF, get your subject as close to the cam as possible and use a fast lens:)

@Kihlian, Yes this is a great site. I mentioned it in my blog post, as this is what I based all my calculations on. For sure, there are times when you need a deep DOF. I was more interested in researching shallow DOF at maximum aperture, so this research is slanted in this direction. Great to hear these calculations hold up in reality.

stoney

@psyco

Nope. The longer focal length, the more z-axis distortion, the shallower DOF.

Psyco

@stonebat

But with a longer focal length, you have to be further away to keep the subject the same size (keeping the subject the same size is important in this case). That longer camera to subject distance will give you also a deeper DoF.

I'm not sure if that comment from my first post is 100% true, but I have read it several times lately on very good video/photography pages.

stoney

@psyco You are thinking about cropping FOV. But the longer focal length distorts the z-axis which brings the background closer which gives shallow DOF. e.g. Bokeh will be bigger.

Psyco

@stonebat That video does not proof oder disproof my statement.

Look at the tower in the background, it has the same amount of blur (as far as one can tell with youtube compression) it just gets bigger when the focal length increases. If you take the tower at the end of the shot and resize it to its size at the beginning, I think there would be no difference. (But the video quality is to bad to be sure.)

DrDave

I would not say any of those lenses is really great for DOF, and the 14mm are hyperfocal. Check out this Leica 75/1.4 video

backwheelbates

Psyco, I think in a technical way, you do have a point in that the distances in real world where the images become blurred/in-focus may not change across different focal lengths. But, I have to agree with Stonebat. In the screen space of the final image (aka, the results), those objects have the appearance of being more blurred as they become larger in frame. I'm all for a semantic debate though:)

@DrDave, speaking of semantics, what do you mean by "the 14mm are hyperfocal"?

balazer

Depth of field has everything to do with resolution. The image projected by the lens is sharper for objects that are closer to the focal plane, but within the focus field, everything looks equally in-focus because the circles of confusion are smaller than the output pixels. It's only when the circles of confusion start to be bigger than the output pixels that we start to see them as circles instead of as points or pixels. Typically for depth of field calculations you use a circle of confusion diameter criterion equal to the output pixel pitch.

It's not just theoretical. It totally bears out in practice, and there's a huge difference between the resolution of the GH2's 16-megapixel stills and the 1-2 megapixel video.

Don't confuse depth of field with bokeh size. You can get bigger bokeh with larger aperture lenses and/or close objects, independent of the resolution.

Vitaliy_Kiselev

One Ring to rule the DOF, One Ring to find sensor crop factor,
One Ring to bring them all and in the flame wars bind them

stoney

I'd use 75mm 1.8 instead of 20mm 1.7 if I want a long shot with very blurred background in a city street. The longer focal length, the larger circle of confusion, the shallower DOF.

"Long focal length lenses, also called telephotos, have a narrower field of view than normal and wide angle lenses, and compress space along the z axis of the frame. Because of this, telephoto lenses appear to bring the background of an image closer to the foreground, flattening space; movement along the z axis is also distorted." - The Filmmaker's Eye

Here's the z axis distortion.

rapsoe

The fact that a lens can distort along the z axis is only a myth. A myth that is part of the complex mythology surrounding the purported existence of "normal" lenses. A long time ago, the concept of a "normal lens" made sense in certain contexts, but not anymore. That concept, before becoming obsolete, was regrettably transformed into a myth, and myths unfortunately never die.

Anyone interested in knowing more about the misconceptions surrounding the existence of a normal lens can start by giving a look at Anton Wilson's excellent Cinema Workshop (from page 99 to 102):

http://goo.gl/2e9lE

Hope this helps.

B3Guy

Nice gif. to follow Vitaliy,

One dolly zoom to rule them all, one dolly zoom to find them One dolly zoom to bring them all an on their keyboards make them vomit ;)

stoney

@B3Guy LoL

@rapsoe The z axis distortion is a myth? What about x-axis, y-axis distortions? All myths? Yeah keep staring at the animated gif. It will teach you something new today.

DrDave

There's a zillion examples of all lenses at all apertures in stills and videos online. Throw the theory out the window and just browse for the "look" you want. My Vivitar 58 and Russian 58mm have a totally different look, different bokeh, different edges, different shadow detail, contrast, color and so on. The difference is startling, and click the aperture ring one stop and yet another "look" as the iris changes shape. Move the cam forwards and back so the FOV is the same as another lens, and yet again, a different look for the same FOV. Plug the cam into a big, flat screen TV and start playing around with the lenses. Learn the look, pick the lens, adjust the settings. If shallow DOF is what you want, don't get the 14mm :)

stoney

@DrDave That sounds so Wabi-Sabi. I like.

BTW 14mm could give shallow DOF if you wanna use it from minimum focus distance. Otherwise SLRMagic 12mm 1.6.

rapsoe

@stonebat My apologies if you took my rant as a personal attack. It wasn't meant to be. In all seriousness, I've been a victim of that myth for years before finding out what was all about. And no, x and y axis distortion are all another matter, but seriously, I don't want to convince anyone, hopefully Anton Wilson will. I was just trying to be of help by pointing people to a very interesting read.

Peace.

stoney

@rapsoe Your apology accepted. Hehe. Just kidding. Thanks for sharing another perspective. We can agree to disagree :)

DrDave

@stonebat I did get a nice semi-macro, high DOF shot with the Panny 14mm prime at MFD. I bet the close-up adapter would really do the trick, great for my bug movie.

balazer

Z-axis distortion? That's a nonsense phrase. It's just different perspectives.

backwheelbates

In case anyone is interested, I posted this over at dpreview.com as well.

http://forums.dpreview.com/forums/readflat.asp?forum=1041&thread=42027675&page=1

They have taken this idea, much farther than I could have imagined. Its a great discussion on a whole slew of perspectives relating to DOF.

Thanks again everyone for your comments!!

Vitaliy_Kiselev

@backwheelbates

Discussions abot DOF and sensor crops (and FOV) are equivalent of Mac vs PC thing :-) Flame.

Howdy, Stranger!

Categories

Tags in Topic

Top Posters