Two primes on dual format cameras (24/1.4 and 85/1.4 on D700 and D300)

[john_morris4]

I love my 35/2 on Fx. When I make boring pictures, it's certainly not the lens' fault.

[hocus_focus]

<blockquote>

<p>I think this thread is blown out of proportion...<br /> FWIW I use both a 24 and 85 but for backpacking (not wedding) but I add the 17-35mm as well. Between the three lenses and a FX/DX combo, I find it great for minimizing the load. <strong>The DOF issue is about one stop more on DX. </strong>The multiple FOV/less lenses is the advantage...I have done this for a long time now while most forumers frown upon this very idea. Folks, it's isn't rocket science...use it to your advantage.</p>

</blockquote>

<p>Thanks Leslie. Finally a sensible post. What bodies do you use?</p>

<p>Equivalence aside, how many of you actually own or use 24/1.4 and 85/1.4 ?</p>

<p>I find myself quite often at both ends of 24-70/2.8 and at 135/2.</p>

<p> </p>

[Andrew Garrard]

<blockquote>Thanks Leslie. Finally a sensible post.</blockquote>

 

<p>You know how to make friend and influence people, don't you? :-)</p>

 

<blockquote>Equivalence aside, how many of you actually own or use 24/1.4 and 85/1.4 ?<br />

<br />

I find myself quite often at both ends of 24-70/2.8 and at 135/2.</blockquote>

 

<p>I recently acquired a Samyang 85mm f/1.4, and I like the field of view and DoF so far on FX (likewise the Tamron 90mm that was my short portrait lens until recently). I like 135mm, and own the DC, but don't get along with it especially well; that's not the focal length, though, and I'm happy with a 135 f/2.8 AI when I have time to focus it, but you have to walk a long way away if someone asks for a group photo with one. I have a 14-24, which I use throughout its range for landscapes and architecture, but rarely for shots of people except in very confined conditions - the distortion when the image is viewed at a normal distance is too pronounced for me. I use the 28mm end of a cheap zoom more with groups, and I have an eye out for the 35mm f/1.4 Samyang for candid groups in low light. For what it's worth, I'm usually either wide or medium/long tele - I rarely use 50mm (I use 200mm more than the 50mm, both to stay off the radar of the subject and to control the background). I can't imagine wanting the 24mm f/1.4, given how much it costs and given that I already have the 14-24, but YMMV. It's a lens that I'm happy exists, like a Noctilux, the 300mm f/2 or the 6mm f/2.8.<br />

<br />

I suspect I'd use the wider lenses with people more if I were to pose groups rather than just take candids, which by their nature tend to involve fewer people and longer distances. I hope that sample, albeit from an amateur with NAS, helps.</p>

[craigd]

<blockquote>

<p>Thanks Leslie. Finally a sensible post.</p>

</blockquote>

<p>You were accused of trolling earlier in this thread. I still think your original question was meant seriously, but now that you're resorting to flamebait, I begin to understand why some people thought otherwise.</p>

<p>There was an observation above about how nice it was to see a disagreement that remained civil. This is because Andrew and I are both civil people. There may be a lesson in that for you.</p>

<blockquote>

<p>Equivalence aside, how many of you actually own or use 24/1.4 and 85/1.4 ?</p>

</blockquote>

<p>I own a number of 24mm and 85mm lenses (also 35mm and 135mm, the "effective" focal lengths you originally mentioned) for different cameras. None of them are actually f/1.4, mostly because I find that better performance is usually obtained from f/2 or f/2.8 lenses, which are generally smaller and less expensive too. And it isn't often that I would really want to shoot at f/1.4 anyway.</p>

[Andrew Garrard]

<blockquote>None of them are actually f/1.4, mostly because I find that better performance is usually obtained from f/2 or f/2.8 lenses, which are generally smaller and less expensive too. And it isn't often that I would really want to shoot at f/1.4 anyway.</blockquote>

 

<p>I'd like to pick up on that. In my shooting, it's relatively rare that I actually want a shallow depth of field as such. I do often want to throw the background a long way out of focus - but if I want to exclude/separate the background, I usually do it with a longer lens. I concur with Craig that all the f/1.4 lenses I've seen are far from perfect at that aperture; I'd far rather use a 135mm or 200mm at f/2 than an 85mm at f/1.4. I ended up with a Samyang 85mm f/1.4 because it was cheap enough that I could live with its compromises when I need a bit more perspective or less working distance than the longer lenses offer me - I'd not pay the going rate for the Nikkors. I raised tentative concerns above about the quality of the DX subset of an 85 f/1.4 used wide open compared with a 135mm f/2 on FX; I'm sure this behaviour factors in to the high quality associated with medium format lenses, and I have to say I'm tempted to use a 150mm f/2.8 on a 5x4 if I want a shallow depth of field portrait (one of several reasons I'm looking at view cameras).<br />

<br />

Wide f/1.4 lenses have their use, but given that they won't lose a background nearly as well as a much less exotic longer lens, I tend to consider a fast wide lens as a low-light option (partly because hand tremor is less an issue) rather than being needed for creative depth of field. Besides, with a wider angle lens, I usually want a large amount of subject in focus. Hence my doubting that I'd ever use a 24 f/1.4; I'll be interested in how useful others find it. I'm more tempted to get a 20mm f/4 as a more portable option than the 14-24.<br />

<br />

Of course, I'm simplifying - a 24mm f/1.4 <i>will</i> separate the background more than an f/2.8 version, and this can be useful creatively - but in general that's not how I think of ultrawide lenses. Maybe that's a failure of my creativity.</p>

[ronb]

<p>Per Craig p3:<br>

"Here is a thought experiment: In a controlled studio environment, with a stationary tripod, shoot a scene using the same 24mm f/1.4 lens at f/1.4 using both FX and DX cameras. Print the resulting images at, say, 20x the size of the sensor. Then crop the FX print to match the narrower framing of the DX print. How do the two resulting prints differ?" (Answer - not much, at least in DOF)</p>

<p>...and not to be argumentative: the depth of field is dependent on focal length, aperture diameter (not recalculated f-number), and distance to the subject only. Cropping part of the picture out (either after the fact with scissors, or at the time of exposure with a smaller sensor) doesn't change any of these 3 things. A finer resolution sensor or slow film could make a difference in the circle of confusion, but this would be small. The DOF preview button stops the aperture down so one can eyeball/estimate it before shooting.</p>

<p>Regarding the original post: <br>

"First, while 135/2.2 is still impressive for portraits and sports, the exotic 24/1.4 becomes a boring 35/2.2 on crop camera." (Answer - it becomes a 24/1.4 which has had the top, bottom, and sides cut off.</p>

 

[Andrew Garrard]

<p>I'll revisit this quickly, because I've got the images...<br /> <br />Ron - apologies for not understanding your previous post (I also have paging turned off, which means "p3" confused me!) When discussing depth of field in the context of its relationship to aperture, I always assume that the definition of circle of confusion that we're using is defined in terms of the smallest detail that you can see which will appear as sharp as your eye can distinguish (possibly under poor viewing conditions). If the sensor resolution or film grain defines depth of field, the aperture is not the limiting condition - and possibly none of the image is truly sharp.<br /> <br />If the ability of the eye to detect that an image detail is out of focus is the limiting factor, image enlargement (or at least, field of view of the final print from the viewing distance) is a factor.<br /> <br />To restate the three cases we might consider: 1) Cut the edges off a print and its DoF doesn't change; use the same lens on cameras with different sensor sizes and produce a print whose size is proportional to the sensor size, and the DoF is unchanged. 2) Increase the print size of the smaller sensor to match that of the larger sensor, and the smaller sensor has a reduced depth of field (because it's enlarged more, so a just-visible circle of confusion in the print corresponds to a smaller circle of confusion at the sensor). The field of view of the two prints obviously differs. 3) Fit a lens with a longer focal length on the larger sensor camera to match the field of view of the smaller sensor, and to get the same depth of field, you need the same entrance aperture, which means the f-stop has to reduce by the crop factor - not keep the same f-stop (as you might suspect if you've just been told that the FX equivalent of a 24mm lens used on a DX camera is a 35mm lens, and no other information). In the context of choosing different sensor sizes (or lens focal lengths) to get different fields of view, and considering the relationship between the field of view and depth of field with a given lens on two formats, we're looking at case 3 only.<br /> <br />Example images, because I've got them, follow. I hope this helps clear up any further confusion.</p><div></div>

[Andrew Garrard]

<p>Now with a 105mm lens (ish) at f/5, from the same location. The reduced scaling means that the background is less blurred, and the depth of field is increased compared with the 210mm f/10 image.</p><div></div>

[Andrew Garrard]

<p>Now the middle cropped from this image (as per a 2:1 crop factor sensor) and enlarged to the size of the original. Blur of the background closely matches the 210mm f/10 image (within experimental error and my ability to actually hit exact focal lengths). Therefore we could say that an image taken at 210mm f/10 matches the image with a 2x smaller sensor at 105mm and f/5 (if exposure is balanced - which in this case I did by quadrupling the exposure time with the lens at 210mm), and therefore these lens/focal length pairs on the two sensors are "equivalent" (if we balance exposure). I believe this is the only type of "equivalence" that we're discussing; I hope that clears up any confusion.</p>

<p>Actually, all these images were taking with a 1.6x crop-sensor Canon. Just to prove that there's nothing magical about DX and FX, and for irony (and not at all because my D700 is off being cleaned).</p><div></div>

[leslie_cheung]

<p>d700 and d200 </p>

<p>I use a 24mm but not the 2k F1.4. I rarely put the 85mm on DX and the 24 on FX. Usually, I have the zoom on one cam while either the 85 or 24 on the other. Depends on the available light as well, of course. I'm street shooter mostly. </p>

<p> </p>

[rodeo_joe1]

<p>"DOF is not, in fact, any different at all" - Ah but it is! D-o-F is affected, among other things, by the amount of magnification that the image has to undergo to produce the final viewing size, either in print or on screen. Six pages of "discussion" and that point still clearly hasn't got across!</p>

<p>Let's try and set out exactly how D-o-F is affected:</p>

<p>Case (1) - same lens, aperture and subject distance; different format size. D-o-F <em>decreases</em> as the format size gets smaller in direct proportion to the circle of confusion chosen. This is simply due to the extra image magnification needed for final viewing and is linear.</p>

<p>Case (2) - same field of view, same aperture and subject distance (in other words we get exactly the same picture, but on different formats with an appropriate change in lens focal length). D-o-F <em>increases</em> as the format size decreases. The F-number needs to be altered in direct proportion to the linear change in format size to maintain the same D-o-F.</p>

<p>Case (3) - same format, same aperture, same subject distance, different focal length lens. D-o-F <em>decreases</em> with increase in focal length. D-o-F is in inverse <em>geometric </em>proportion to increase in focal length. The relative aperture must increase as the square of the increase in focal length to maintain the same D-o-F. For example if you double the focal length then the aperture number needs to increase fourfold.</p>

<p>When we combine these cases, it's easy to see that a change in focal length effectively trumps all other changes, and everything else follows. Which is why a smaller format generally gives greater D-o-F for the same aperture, but <em>only if the field of view and subject distance remain the same.</em> If the focal length and aperture stay the same then D-o-F reduces. Clear?</p>

<p>Just to reinforce the point, here's my depth of field calculator again with a few values inserted. I've also added a total depth column to make it easier to see any differences and similarities.<br>

And yes, depth-of-field is confusing for newbies, but it seems it's also confusing for everyone else!</p>

<p> </p>

[craigd]

<blockquote>

<p>"DOF is not, in fact, any different at all" - Ah but it is! D-o-F is affected, among other things, by the amount of magnification that the image has to undergo to produce the final viewing size, either in print or on screen. Six pages of "discussion" and that point still clearly hasn't got across!</p>

</blockquote>

<p>Wrong. I've referred to greater magnification of prints or other display media as a factor in DOF several times. I've also noted that a smaller CoC value reduces DOF, and the reason you use a smaller CoC value with a smaller sensor is because of the assumption that all images will be displayed at the same size regardless of the sensor used (i.e. that images from smaller sensors will be magnified more when displayed). This is a conventional assumption, not a fact of life, and all it does is obscure the real relationships involved, leading people to think, incorrectly, that the sensor size, by itself, has a direct effect on DOF, which it does not and could not.</p>

<p>DOF changes as a result of a change in focal length, subject distance, or display magnification. The sensor itself is at best indirectly associated with this. You may choose a shorter focal length because you are shooting with a smaller sensor, but that's a choice you make based on your own artistic intent. The sensor is not doing it, you are.</p>

<p>Your examples are cluttered with irrelevant and unnecessary assumptions that only serve to confuse the issues. You want display size to be the same for all images, and you also want field of view and subject distance to be kept constant. These assumptions obscure what's really happening. If the sensor size changes, a narrowing of the field of view is the only direct consequence. If focal length, subject distance, or display magnification change, it's because you decide to compensate for the change in sensor size. Those compensations are actually what changes DOF. The sensor is not doing it, you are.</p>

[Andrew Garrard]

<p>Craig: You could take a photograph with a 100mm f/2.8 lens. Mount it on a 5x4 camera or an APS camera and make a contact print, and I agree that the depth of field is identical - the larger format merely provides a much larger field of view and a much larger final image. In my opinion, it would be odd to consider two such prints as in any way equivalent - they present an entirely different scene to the viewer.<br />

<br />

We talk about "equivalent focal length" because most people choose their lens focal lengths in order to control the field of view on a sensor (or film format that they own). If you're used to thinking "I need a field of view that matches a 50mm lens on a 35mm camera" and I have a DX format camera, it's useful to know that a 35mm lens is (roughly) what you need. If you put a 24mm lens on a DX camera, it's nice to be able to think about what lens would have given a similar field of view on a 35mm full frame, if that's what you're used to. Maybe naming lenses in terms of their angle of view on a standard format would have been preferable, but we have precedent for the current scheme. Doing otherwise means that you need to work out the (not clearly stated) sensor size of your compact camera in order to work out the field of view of its 8mm lens. And if you're going to think in terms of "equivalent focal length", it's just as valid to think about the depth of field of the "equivalent" lens that you're imagining.<br />

<br />

Similarly, most people, in my experience, do not use their sensor size as a factor in deciding how large a final image will be printed (for example, I've never met a camera that actually puts the sensor resolution into the EXIF ppi field). That the actual scale factor is different may be an indirect consequence of this, but I've never met someone who thinks "I need to enlarge the negative by 10" rather than "I'd like a 10x14 print".<br />

<br />

So, true, it's not the sensor size that causes any depth of field changes: it's the differing ways in which most people would want to treat the image produced on these sensors. Does this "confuse the issues"? Well, if you don't perform different image scaling, there <i>are</i> no issues to consider; the entire point of talking about "equivalent" focal length and depth of field is when you want to produce the same image from two formats different formats. It's hard to call this an "irrelevant" or "unnecessary" thing to want to do, at least in the context of this discussion, which is explicitly about using two formats to take advantage of this precise difference. If you switch from FX to DX and you know what effect you want, it's simple enough to increase the focal length by a factor of 1.5 and decrease the f-stop by a factor of 1.5 (and either let the camera worry about the exposure, or increase ISO by 2.25x). Even though I could, I don't work out everything from scratch given the sensor measurements and a calculator with transcendental functions. (To bring this vaguely back to the OP's question: I don't find scaling small numbers by 1.5 to be confusing.)<br />

<br />

I was just putting a web page together so I don't have to discuss this again. Then I thought "this is dull, I'm wasting a Sunday, and the tennis is too interesting; I'll just check Photo.net..."</p>

[dan_brown4]

<p>Andrew, you might benefit from reading Section 3 on exposure in that book "The Negative" LOL.</p>

<p>Just kidding ;-)</p>

[Andrew Garrard]

<p>:-) I'm glad someone appreciated my attempt to produce a pretentious sample image - it was this or a sleeping cat. I'll admit that I didn't use the zone system, though. I actually had to increase the exposure by two stops in RAW conversion - I was so concerned with framing that I forgot to bring up the histogram, and the meter was thrown by the black and white areas - but I did the same for all the images. I've read The Print too, and don't think Ansel would mind too much. The effect of lens and format changes on field of view and depth of field is, of course, in The Camera, although (IMHO) some diagrams would have helped, and "depth of field changes with the square of the focal length" could have benefitted from "linearly due to magnification and linearly due to the increase in aperture". But the results are more useful than the working. Oddly enough, I doubt Ansel ever carried a 10x8 and a 5x4 to increase his lens range...</p>

[sunray1]

<p>Well, sidestepping the whole DoF issue, I think shooting a D300/D700 combo with a fast 24 and a fast 85 would make sense. In fact it's what I hope to end up with eventually. I think at ISO 1600 both camera's would yield good IQ and the similar button lay out would work easy. I'd add a 50/1.4 and then have the following (for me most useful) options:<br />-wide + short tele (D700/24+D300/50)<br />-standard +short tele (D300/24+D700/85)<br />-standard + medium tele (D700/50+D300/85)<br />-wide + medium tele (D700/24+D300/85) <br />At arrival at an event I'd assess the situation and chose the most appropriate combo, leaving just one 'redundant' lens in the bag. That way I'd foresee not much lenschanging or confusion...</p>

[rodeo_joe1]

<p>Craig, any attempt to calculate or quantify depth of field <em>must</em> take into account the magnification needed to view the image. Any depth of field calculation needs to be based on a suitable circle-of-confusion and this needs to be chosen to reflect both the format size and an "average" print size and viewing distance.</p>

<p>The standard way of choosing a suitable c-o-c is to divide the format diagonal by some constant - usually a number somewhere between 1500 and 1750. (Now personally I have strong objections to using the image diagonal for any comparative purpose, but that's another issue.) Anyway, whether we use the format diagonal or some other metric, the c-o-c always needs to be linked to the original image size - i.e. the camera format.</p>

<p>If we throw away this linking of c-o-c to format size, then there can be <em>no depth of field calculation</em>. So either you accept that the format size is integral to the calculation of depth of field, or we don't have any depth of field calculation to argue about, and we can all give up and go home. The fact remains though, that such a thing as depth of field exists. So are you saying that we shouldn't even attempt to quantify it?</p>

<p> </p>