Does APS-C sensor provide better bokeh than FF and why?

ashishgarg · February 8, 2012

I guess I know this but looking more for the technical reason. I believe usually longer focal length provides better bokeh, but why?

craig_meddaugh · February 8, 2012

Bokeh is the way a lens renders or draws out of focus points of light. Being a product of the lens, sensor format does not matter.

Now how much bokeh matters is a product of depth of field, in which case FF has the advantage as it can render a shallower depth of field than APS-C (as the subject magnification is larger for FF than it is for APS-C).

As for long lenses, they look like they provide a shallower depth of field or "better bokeh", but this is a product of the narrower field of view. This means that there is effectively less background captured meaning it is easier to isolate your subject.

richardsperry · February 8, 2012

nm

Matt Laur · February 8, 2012

What you're really asking is, "Does a larger format sensor allow me to more easily shoot with a shallow depth of field?" The answer is, in practical terms, is, "Yes." But as Craig points out, the way that out-of-focus blur is actually rendered, the aesethic qualities of the blur (NOT how blurry it is), depends on the lens you're using. You could use three different 50mm primes at f/2.8, and three different f/2.8 zoom lenses that happen to cover the 50mm focal length, and all six of those lenses used on exactly the same subject matter, at the same distances (between you, your subject, and the background) at the same focal length and aperture, would produce: 1) Essentially the same amount of depth of field, and thus the same amount of blur in the foreground and background. This is selective focus. This is a natural byproduct of any and every lens used in the same way. 2) Possibly very different looking blur. Some of those lenses could produce "buttery" or "creamy" blur, and others a more jaggy, nervous, or edgy-looking blur with visible artifacts. This is what we're talking about when we use the word "bokeh." Not how out of focus is it, but what does the out of focus area look like. So, the sensor format doesn't have much to do with it at all, other than that for a given shot (say, a head-and-shoulders portrait from two meters away), you'd use a different lens (a difference focal length) for each of those two camera formats. And because you might choose a different actual lens (perhaps from different manufacturers, or from different price points in Canon's product line) to suit the different sensor format's field of view issues for a given composition, you may find that the lens you choose for one or the other has different bokeh characteristics. Different because of how each lens is designed, what sort of glass is used, the number of iris blades, etc. Longer focal lengths on a given camera format do narrow the field of view. Maintaining the same framing (say, of a portrait subject) also means backing up farther and farther away as you go longer. Which also means you're changing your perspective on the subject, as you also reduce how much background is visible in the frame. Depending on the image, the elements seen in it, the light, and your creative purpose, issues like perspective and field of view may be far more important than the way the background blur is rendered.

bobatkins · February 8, 2012

Depth of field and background blur are entirely different things goverened by entirely different lens properties. Check out <a href="http://www.bobatkins.com/photography/technical/bokeh.html">http://www.bobatkins.com/photography/technical/bokeh.html</a> for more info.

The answer to your question is that the quality of the background blur does not depend on format. It depends only on the lens.

If you want to determine the quantity of background blur in a print of a given size taken with a given lens used on a camera with a given format size, this calculator will do it - <a href="http://www.bobatkins.com/photography/technical/bokeh_background_blur.html">http://www.bobatkins.com/photography/technical/bokeh_background_blur.html</a>

In general, for any given angle of view using lenses appropriate for the format (i.e. with different focal lengths), a larger format will tend to throw a distant background out of focus more.

jamie_robertson2 · February 8, 2012

To the OP: No.

joseph_wisniewski · February 8, 2012

To the OP: Yes.

What the others say is technically true, but in practical terms, mostly irrelevant. There was a "relevant truth" repeated by several people, that the definition of bokeh as the "quality" of the out of focus parts of the image, the freedom from false "bright edges" and "double lines." Although it is possible to measure the quality of the bokeh, this is seldom done, and it's usually expressed in subjective terms, like "good", "creamy", or "smooth" for the bokeh that the majority like, or "poor", "harsh", or "busy" for the sort that the majority don't like as much. There's always a minority who prefer the sort of bokeh that the majority dislike, too...

Bokeh is a function of the lens design.

<ul>

<li>Fast normals, like the typical 50mm f1.8 or 1.4 are designed to get added sharpness at the expense of bokeh, so to gain a bit more detail in the part of the image that's supposed to be in focus, they give up smoothness in the out of focus background. Even a legendary one, like the 50mm f1.4 Zeiss, is especially ugly.</li>

<li>Macros tend towards ugly, too, but this is a trend that lens makers are addressing, as more people are making macros "do double duty" as "portrait lenses".</li>

<li>Short teles are usually deliberately designed for "good" bokeh, with the Nikon 85mm f1.4, 105mm f2.0 DC, 135mm f2.0 DC, and older 105mm f2.5 Ai, or Canon 85mm f1.2 being classic examples.</li>

</ul>

This is independent of format. Put that 135mm f2.0 DC on a FF camera or a 1.5x crop APS, and it will still deliver good bokeh, because that's how it's designed. It will also deliver lots of background blur, because it's long and fast. So, it's got quality and quantity, you don't have to pick between them.

The problem is "suitability for purpose". A "conventionally trained" photographer, such as myself, will tend to work around 8-10 feet from a portrait subject. At that distance, on FF, the 85mm is the right focal length for "waist up" portraits or couples, the 105 a classic "head and shoulder portraits", and the 135 a favorite for headshots. That's three lenses that Nikon, Canon, and Sony make in versions with good bokeh.

On APS, you need 50-60mm for a waist up, about 75mm for a head and shoulder, and 90-100mm for the headshot. Nikon, Canon, and Sony don't make "good bokeh" lenses in those focal, although they're getting closer. For example, the current 60mm f2.8 Nikon macro has much nicer bokeh than its predecessor. But it's not in the same league as the 85mm f1.4 I'd use on FF. Pentax, on the other hand, makes some great bokeh lenses for APS, a 55mm f1.4 and 77mm f1.8 that you have to use to appreciate how good they are. Especially the 77mm.

So, the "meaningful" answer is that, if you shoot Nikon, Canon, or Sony, you can't get as good bokeh from APS as you can from FF, because you can get "good bokeh" lenses in the lengths you are most likely to want. If you shoot Pentax, you can get a variety of "good bokeh" lenses in APS.

And, for the four thirds, NEX, and Samsung systems, the "good bokeh" lens landscape is famine-scorched and dead...

So, yes, overall, with the exception of Pentax, FF provides much better bokeh than APS.

sherman_peabody · February 8, 2012

Well to follow up on this - the image circle is produced by the lens, and the crop sensor just uses the center part of that, but the image is the same, and the depth of field should be the same with either sensor, all else being equal. But I wonder if the pixel size has any effect on that - not the megapixels per se, but the size of the individual pixel sites. Does anyone know if that makes any difference in the depth of field?

bobatkins · February 8, 2012

and the depth of field should be the same with either sensor

</blockquote>

Bzzzzzz. Sorry, wrong answer - <a href="http://www.bobatkins.com/photography/technical/digitaldof.html">http://www.bobatkins.com/photography/technical/digitaldof.html</a>

Pixel size doesn't affect DOF

Matt Laur · February 8, 2012

Sherman would be right if the photographer didn't move or care about composition differences when comparing that same lens on the two different format bodies. But that's not a real-world situation. In the real world, photographers have favorite working distances (because of perspective, logistics, etc), and that means choosing the right focal length for each job. And as I and J.W. have each pointed out, you may not have the same choices of lenses for use with each format ... and so there is an indirect impact, possibly, on real bokeh-ishness, because different lenses are involve, and all lenses are different in that regard.

bobatkins · February 8, 2012

Actually, he'd be wrong with the same lens. Depth of field takes into account the magnification required to go from the "negative" (sensor) to the "print" (or monitor). Depth of field is NOT an intrinsic property of a lens like focal legnth or aperture. It depends on a set of assumptions about magnification (print size), viewing distance and visual acuity.

jcoffin · February 9, 2012

While I agree with the basic gist of what J.W. seems to be getting at, some of his post seems self contradictory.

When talking about full-frame, he lists 105 as one of the lengths that Nikon, Canon and Sony all make in versions with good bokeh. Since neither Canon or Sony makes a 105 at all, I'd assume that in their cases he's talking about their 100mm lenses.

On the other hand, when talking about APS-C, he mentions 90-100 as the head and shoulders lengths, and says they don't make "good bokeh" lenses in this length. At least for Sony and Canon, however, this seems to include the very same 100mm lenses he previously said do produce good bokeh.

I'd also mention one other factor: in quite a few (most?) cases, the bokeh produced by a lens varies with the aperture. Obviously changing the aperture changes the DoF, but at the same time, it varies the bokeh. Although a wide open aperture obviously gives the shallowest DoF, it does not necessarily give the smoothest bokeh (in fact, often doesn't in my experience). A few lenses (especially 85mm) seem to produce nice bokeh wide open (or nearly so) but many others "smooth out" quite a bit with the aperture closed a stop (or two).

That creates a bit of a conundrum though: especially for portraits, you generally want pretty tight control over the DoF. I'm usually trying for the widest aperture that still stays reasonably sharp from around halfway down the nose back to about the front of the ears. By whatever coincidence (or possibly conscious design) with the 85, 100, and 135 on full frame, with most peoples' faces I get that DoF at an aperture that also happens to have quite nice bokeh.

On APS-C, however, I have to open the aperture about a stop to maintain about the same DoF -- and especially with my 100mm, that wider aperture gives substantially inferior bokeh.

desmond_kidman · February 9, 2012

1. Bob Atkins and several other posts and sites will all confirm the truth, that the same "equivalent" focal length lens at a given f-stop will give more background blur on a larger sensor camera than a smaller sensor camera.

2. Although amount of background blur and quality of background blur are 2 different things, my experience is that most folks will require a good amount of blur plus nice quality blur to be pleased. If the amount of blur is not enough to make the image "pop" from the background then the effect is not what most folks want. The quality of the blur just does not show up very much unless there is a good amount of blur.

3. Based on my experience, getting enough blur is more of a challenge with 35mm sensor (or film) size than medium format, and more of a challenge yet with APS size relative to 35mm size. The quality of blur starts being relatively meaningless if you can't get enough blur to begin with.

4. My experience is that if you want to play with really effective selective focus effects you want the largest sensor camera you can get. Stretch to full frame (35mm size which equals 24mm x 36mm sensor) if you can. You will less often wish you could get more blur in the out of focus ares. You will have less expensive fast lenses available to you for evaluation for the type of blur you want.

I personally think that the Canon 50mm 1.4 and the 85mm 1.8 lenses, inexpensive relative to their quality and max aperture, give great blur with full frame cameras. I think you'd be hard pressed to get as much and as good blur with an APS sized camera.

I don't think I ever read about anyone saying "I have both FF and APS and I like the bokeh better on my APS".

I have to repeat myself here as I think it is the most salient point: The quality of blur starts being relatively meaningless if you can't get enough blur to begin with.

sherman_peabody · February 9, 2012

So, Bob, if I go to my favorite online DOF calculator, DOFMaster.com, and select the Canon 5DMkii with a 50mm lens at f/2.8 and subject distance of 6 feet, it says the DOF is .73 feet. But if I change the camera to a 7D, with everything else the same, it says the DOF is .46 feet, or much shallower than the 5D. How should I interpret these numbers? Are they just bogus? Is there a way to tell what assumptions are being used? They give the fomulas, but in the end it comes down to the circle of confusion, which they don't define further.

sherman_peabody · February 9, 2012

Bob - nevermind. I understand now. It's 1.6X less DOF with the 7D, which is what your writeup said. But that's because the field of view is smaller. If the field of view with the 7D was the same as the 5D, either by backing up, or switching to a 31mm lens, the DOF of the 7D would be 1.6X DEEPER than the 5D.

And on further reflection, I think I understand that pixel size doesn't affect this unless it's too big. In other words, the 36MP full-frame sensor on the new Nikon won't affect DOF.

porter · February 9, 2012

To the fellow who said something disparaging about 4/3 and their lack of lenses able to produce 'good' bokeh:

from awallphoto (flickr)

Nex might have trouble in this area, but 4/3 has a lot of really good professional (and expensive) glass that can produce amazing results (including bokeh).

porter · February 9, 2012

<a href="

title="Quick Radom Shoot by awallphoto, on Flickr"><img src="http://farm6.staticflickr.com/5184/5801269289_a1d5c501a1_z.jpg" width="640" height="480" alt="Quick Radom Shoot"></a>

mattb1 · February 9, 2012

Sorry, that first 4/3 example looks like bad photoshopping to me.

michael_elenko · February 9, 2012

Yeah, second one too.

ME

zack_zoll · February 9, 2012

The reason that Bob's calculator, and the DOFmaster site, and many others don't jive is because they aren't all functional; some are technical. Technically, a lens will give the exact same DOF with ANY negative or sensor size, assuming the aperture and focus is held constant. Practically, you will often need to move closer or farther to frame the subject the way you want (farther away with a smaller sensor), and the act of focusing closer or farther will change your DOF, even if technically shouldn't.

This means that a 35mm lens on an APS-C camera, focused 10 feet away, will give a similar image to a 50mm lens on an FF sensor, but with greater DOF. A 25mm lens of an m4/3rd will have more DOF still.

Where this matters to you is that wider lenses generally have uglier bokeh to most eyes. This means that using a larger sensor might give you a shallower DOF, but you're also more likely to be using a lens with nice bokeh in the first place. And yes, I realize there are lots of exceptions, and I'm sure many of them will be posted here. That's not important. The OP asked about a general rule, and as a general rule a larger sensor will give you the same bokeh if you're shooting charts, but in practice will usually be better.

Matt Laur · February 9, 2012

That second example, above, is actually an example of bad bokeh. Notice the areas that haven't been deliberately softened in post with a gaussian blue. See those harsh highlights and jittery lines? That's the stuff that many people want to avoid, not achieve.

desmond_kidman · February 9, 2012

Zach, not so. If you capture:

1. The same field of view (meaning same equivalent focal length)

2. From the same viewing distance

3. At the same numerical aperture

4. Enlarge to the same print size

from different format cameras then the smaller format will have greater depth of field / less blur.

This is clearly stated in books from the best experts, and anyone who has shot 8 x 10, 4 x 5, 6 x 7, 6 x 6, 35mm, APS will be VERY aware of this from direct experience.

Just the way it is. My latest move from 6 x 7 to 35mm size frame makes it painfully obvious when trying to get extreme blur and gleefully obvious when going for DOF.

I believe the mathematics of this are laid out in Feininger's (sp?) great work, "The Complete Photographer" as well as numerous other reference sources. But you don't need the experts, try it yourself on different formats.

zack_zoll · February 10, 2012

Desmond, you've just illustrated exactly what I said someone would illustrate. Clearly you've never tried to fill the frame using a 150mm lens on a 4x5, and then tried to shoot the same subject with a smaller camera and a similar lens. If you had, you would know that for practical reasons, the distance between camera and subject often needs to change, which means a direct scientific comparison is not possible - even if you did sort of agree with me towards the end.

And even if it were, it still wouldn't change the fact that the quality of bokeh of wide-angle lenses is often less than stellar.

I'm not going to get into an argument over it though. Feel free to quote all the sources you like; the fact is that producing a scientific test means controlling for factors that are non-issues. Print sizes for instance: why bother to control for print sizes? If you weren't looking to make larger prints, why would you be using larger sensors/film? Numerical apertures ... well, not all lenses are honest about their aperture ... this is why the cinema industry uses t-stops. Do you really want to go through all the self-flagellation of determining if the aperture is correct before you test? Most lenses are off by a few tenths of a stop. And what is an equivalent focal length? Does that even exist? A lens sees what a lens sees, and the perspective and field of view are dependent on where you stand and what size the film/sensor is.

I could go on an on. But the fact is that in this case, the science is worthless. Go with the general rule, realize there are exceptions, and enjoy your craft.

desmond_kidman · February 10, 2012

I'm only responding again in the interest of fairness to readers who wish to understand and in fairness to Zach.

First, in my last post, by "equivalent" focal length lenses on different formats I mean that each format use a focal length that will give the same field of view from the same distance. In other words, the scene will be the same when shot from the same spot.

Somehow I misread Zach's post. I went back and read it carefully, and noted this line:

This means that a 35mm lens on an APS-C camera, focused 10 feet away, will give a similar image to a 50mm lens on an FF sensor, but with greater DOF. A 25mm lens of an m4/3rd will have more DOF still.

</blockquote>

So, since this is what I was saying, greater DOF for smaller formats, we appear to be on the same page.

I have observed that viewers tend to like limited DOF along with nicer blur, I maintain that the larger format, the greater potential to use bokeh as a tool.

I also agree with Zach on the issue of lenses. Since the 35mm frame (24 x 36) was present for a much longer time than APS and smaller frames, the quantity of lenses optimized for certain uses, such as portraits, is greater, and more effort has been put into the designs to deliver qualities that fit those uses. Hence, the common knowlege about the existence of lenses in the 85mm range that have nice qualities for portraits (in addition to just the focal length) including very nice bokeh.

I use just about all formats, and I'll stand firm on the opinion that with DSLRs, if you want the best bokeh, and the greatest amount of potential for out of focus blur, go with the full frame sensors.

philip_wilson · February 10, 2012

I will just add that I agree with Dennis (and others like Bob and JW) and find that in the real world sensor size does matter. The reason as others have stated is that you have a better choice of lenses for the working distances that you are shooting at. This gives you more lenses with a shallower DOF and good bokeh. If we take a real example we can see that at 10 feet with an 85mm equivalent lens at a subject distance of 3 meters (10 feet) using my cameras I can have

Fuji GX680 (MF film) with 180 F3.2 shooting at F3.2 the DOF is 12cm

Canon 5DII with 85 F1.2 the DOF is shallower at 9cm

Canon 7D with 50 f1.2 the DOF is 16cm

So the DOF is similar in all three cases (I actually do not have the 50 F1.2 I use the F1.4 version which does not have great bokeh) and should not make a major difference to the OOF background. However, the real issue is the quality of the bokeh. While all 3 lenses are designed for portrait use the fact is that the Fuji lens creates the creamiest Bokeh and gives the most three dimensional image. This is obviously a subjective assessment but is something most people looking at the prints can see. I also wonder if the Fuji is helped by the fact that it shoots film where unlike a sensor the film emulsion has a finite depth. I have seen that my old FD 85 F1.2 produces (what I consider) to be slightly more pleasing bokeh than the new 85 F1.2 and again in may be due to the use of film.

All of the lenses and bodies listed above will produce fine portraits and good bokeh but the Fuji 180 does a (slightly) better job than the Canon 85 F1.2. Of course this should be expected.

The Fuji was built with studio portraits in mind - the camera allows full front movement with all lenses (which allows the image plane to be manipulated), and has a range of ten prime lenses in the portrait FL range 125mm to 300mm (35mm equivalents of having 58mm, 63mm, 69mm, 83mm, 89mm, 97mm, 115mm, 138mm). Each of these lenses sold for between $1000 and $2000 in the mid 1990s with the better portrait lenses being at the expensive end. The 180 F3.2 takes 95mm filters and the camera with a lens weighs 10lbs so there are other compromises.

Bokeh quality is difficult to measure but the longer Canon lenses (85 F1.2, 85 F1.8, 135 f2) all do a good job. I am sure the 100 f2 is good but I have never used it although I find my 100 F2.8 L IS is a pleasant portrait lens. The faster 50mm lenses do not (in my opinion) produce the bokeh quality as the longer lenses so with an APS-C Canon you are left with a longer shooting distance and the 85mm lenses if you want great bokeh.

As an aside the 300 F3.8 has always done a great job of producing great bokeh but it is generally impractical for portrait use due to the shooting distances. To get the same shot as a FF 85mm lens at 3 meters you need to shoot at 9.5 meters although at F2.8 you do get an 18cm DOF.

Sign In

Does APS-C sensor provide better bokeh than FF and why?

Recommended Posts

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Create an account or sign in to comment

Create an account

Sign in