DX/FX comparison tool

[jon_savage]

<p>I thought this spreadsheet might be of interest to some. I used this to compare DX to FX as I am considering a system switch to FX and I wanted to understand what I would need to replace my DX kit with and to set my expectations. I’m considering moving from a D7100 (£750 and DxO low ISO rating of 1256) to a D750 (£1750 and DxO low ISO of 2956).<br /> <br />The notes on the side in the next image explain a bit more if you wanted to make your own version. I used a couple of simple approximations where needed to make it easier.<br /> <br />But note that by “the same” I am referring to the final photograph. So the image you would print or the image you would stick on a website. The two systems settings are considered equivalent if they would produce the same looking photographs when taken with the same available light. I guess this is only important if you are taking photos where the depth of field is critical but it might help compare the strengths and weaknesses of systems in areas where you use them most.<br /> <br />It’s also interesting to look at the pixel size vs diffusion interaction for the bigger sensors. Obviously in the real world the lens and sensor quality will be impacted by cost and manufacturability which is why some systems don't have some lenses available and there is more to an image when you include other optical characteristics beyond depth of field and diffusion.</p>

<p>It seams to check out OK and it was an interesting exercise! <br /><br /></p><div></div>

[jon_savage]

<p>Detail behind the table and the calculations used....</p><div></div>

[Robin Smith]

<p>How did this help you, and what are you trying to establish, as a matter of interest? I don't understand the ISO row.</p>

[Matt Laur]

<blockquote>

<p>I don't understand the ISO row.</p>

</blockquote>

<p>Because difference size sensors mean that a given lens aperture will result in different effective DoF, you have stop the lens down more on the larger sensors. But that brings in less light. So if the shutter speed remains the same, the only way to compensate for the closed-down lens is to raise the ISO sensitivity. That isn't too horrible of a concern, because you'll usually find that the larger sensor will tolerate (meaning, produce less noise, and better dynamic range) than will a given smaller sensor used at the same sensitivity.</p>

[Wouter Willemse]

<p>I don't know, for me the main reasons to go from DX to FX camera were a much larger viewfinder (easier with manual focus lenses) and less depth of field (and I am guilty of shooting too wide apertures too often). Nothing that can be captured in figures all that easy, but it makes the system a better match to my desired way of working. All the numbers become meaningless to me the second I hold a camera and start working with it. I understand the digits and what they're supposed to mean - but in real life, shooting real photos, are the figures relevant? The larger viewfinder is, though.<br>

So, I really don't see the point of making a calculation on what it takes a FX camera to look like m4/3rds, or vice versa. It's about what each system does best, worst or not at all - that makes or breaks the choice. The rest - as far as I am concerned, just noise (that's a bad pun indeed, I'm sorry). If you feel the specific advantages of a full frame camera as the D750 suit your style, go to FX. Else, DX is far more cost efficient. Don't make the choice more complicated than it needs to be.</p>

[ant_nio_marques]

@Matt, I still don't get what the ISO row is supposed to convey. It says 'given aperture and shutter'. What's that supposed

to mean if then you 'have to stop down more' to get equivalent DOF?

[jon_savage]

<p>Matt/Antonio,</p>

<p>If you imagine two people side by side taking the same picture, one with DX, the other with FX. I think it’s widely accepted that the 1.5 crop factor makes 50mm on DX “the same” FoV as 75mm on FX. Most even agree that f2.8 in DX is “the same” as f4.2 in FX when also considering the DoF differences the sensors show in the final print.</p>

<p>But what most miss when comparing the two systems is that when taking a picture of the same subject is that once you have changed the aperture you need to change the ISO to match (because the available light is the same). Otherwise you would underexpose FX when trying to get the same image (match the DX FoV and DoF). The ISO row is the ISO to go with the aperture and shutter to match the exposures for the column.</p>

<p>In a way you’ve really just said a photograph taken with the same available light will show better sensor qualities when taken with settings that produce a shallower DoF.</p>

<p>It was really just an excersise to understand the system differences and physical interactions and I thought others may be interested in it. Like Wouter says, it's about what each system does best, worst or not at all - that makes or breaks the choice. The spreadsheet may help some understand how and why each system behaves like it does from a purely theoretical sensor/lens point of view. No considerations for any particular camera ergonomics modelled!</p>

[ant_nio_marques]

@jon all I'm saying is that if that's what you mean then the ISO row should say 'given *DOF* and shutter', not aperture

which makes it wrong.

 

I don't think anyone overlooks the DOF factor. What happens is that folks usually want *shallower* DOF. If they want the

same DOF with the same aperture, then they'll choose slower shutter over higher ISO unless the shutter value is

intrinsically needed.

 

While we're at it, the 645 should be to the right of FF and the crop is 0.77 (at that point, it does make a difference).

[rodeo_joe1]

<p>Jon, there appears to be an error in the crop factor that you've used as a comparator between 645 and Full-frame. It's nowhere near to 0.8.<br /> The "6 x 4.5 cm" format is actually 56mm by 42mm, and full-frame is 36mm by 24mm. There can be no fixed direct comparator because of the different aspect ratios of the two formats, but we can compare horizontal or vertical angle-of-view. (Using the frame diagonal as a comparator is pointless from a practical viewpoint IMO.)</p>

<p>Whatever way you look at it, the crop factor is closer to 0.6 than it is to 0.8.<br /> 36/56 = 0.64<br /> 24/42 = 0.57<br /> And if you <em>must</em> use the frame diagonals: 43/70 = 0.61</p>

<p>That would make the equivalent 645 lens focal length about 120mm, and obviously, this throws all the other calculations for 645 out as well. The C-o-C needs to be increased to ~ 48 microns, which means the F-number needs to be f/4.5 for a similar DOF at 3 metres.</p>

<p>If you're using one particular digital "Medium Format" (not really) sensor as a comparator, that's different. Though I don't know why anyone would bother with such an expensive and trivial upsizing from a full-frame DSLR.</p>

<p>And why the apparently arbitrary ISO figures? If full-frame is your baseline, as it should be with a crop-factor of 1, then why not put 100 or 200 ISO into its ISO box as a starting point?</p>

[jon_savage]

<p>Rodeo Joe, I don’t think there is an error, I think you have assumed a different sensor size than I used for the column headed “Pentax 645”. This was for the Pentax 645 digital cameras (645D and 645Z with a sensor size of 44x33). So maybe I needed a clearer heading so it’s not read as “Medium Format”. For example this Pentax 645:</p>

<p>http://www.techradar.com/news/photography-video-capture/cameras/does-anyone-really-need-a-digital-medium-format-camera--1246587<br>

(in a techradar article on "digital medium format")</p>

<p>I don’t think you must use the diagonal but it’s a good one to use and if you believe Wikipedia the most common one:</p>

<p>“In digital photography, a crop factor is related to the ratio of the dimensions of a camera's imaging area compared to a reference format; most often, this term is applied to digital cameras, relative to 35 mm film format as a reference. In the case of digital cameras, the imaging device would be a digital sensor. The most commonly used definition of crop factor is the ratio of a 35 mm frame's diagonal (43.3 mm) to the diagonal of the image sensor in question; that is, CF=diag35mm / diagsensor”</p>

<p>I think we agree but we are talking at cross purposes on the assumed Pentax 645 sensor size.</p>

<p>For the arbitrary ISO, well the point of it being a spreadsheet is you can put in any ISO you want to try. The screenshot happened to have the DxO low ISO figure in it as that was the last one I happened to type in.</p>

<p>I was just sharing the mathematical relationships between the systems for those that might be interested. I’m not commenting on why anyone would bother switching between systems.</p>

<p>Having said that I think I might agree with you about the Pentax 645 (44x33 sensor) vs full frame looking expensive for the difference you get. But then it’s not aimed at my type of photography (or my budget!).</p>

<p>Antonio, How do you get 0.77? How does it make a difference or are you just suggesting I should show more decimal places?</p><div></div>

[ant_nio_marques]

You were comparing digital formats, so Pentax 645 did have to be the 645z/645d rather than the film system called

'Pentax 645'.

33x44 / 24x36 = 0.77..... Below 1, the decimal digits become more relevant, hence why I said 0.8 was a bit misleading.

Whoever's worked with a 645z can explain why it is anything except a 'trivial' improvement over FF. it's a different system

with different dynamics. By comparison, the difference between APS-C and FF _is_ trivial. They work much the same.

[jon_savage]

<p>Antonio, thanks for the clarification. It looks like you used the sensor areas for the crop factor calculation and I used the diagonals.</p>

<p>Also just to clarfiy I'm not saying the 645z to FF difference is "trivial", but I did say it was too exspensive for me :-)</p>

 

[rodeo_joe1]

<p>Antonio, I remain to be convinced by the current crop of so-called medium format digital cameras. Let's see a side-by-side comparison of a 645Z image up against a D810 with the exact same subject/lighting/etc., which so far I've not seen.</p>

<p>Previous examples from CCD 'Blad digital backs haven't looked that impressive, given the limitations of ISO and flexibility of use. And since the 645Z has an almost identical pixel density to a D810, and presumably uses very similar CMOS sensor technology, I can't see the IQ of a D810 being "blown away" by the Pentax offering.</p>

<p>I've mooted before that once the resolution constraints of film are removed, the format size becomes a choice of convenience, system flexibility, lens choice, frame rate, ISO speed versus noise, and controllability of DOF. Image Quality is an almost insignificant factor.</p>

[ant_nio_marques]

<p>I don't recall saying anything about 'blown away', Joe. I said 'it's a different system with different dynamics.'.</p>

[jon_savage]

<p>Joe, it’s interesting you say that. I’d not paid much attention to “medium format” having never used it hence my confusing label of Pentax 645 not realising it meant 56x42 to most people. But based on this thread I Googled a bit. It was surprising what kind of information the top hits brought up.<br>

Shallow depth of field appears to come up a lot with MF users. But they do come out with some confusing comparisons. This guy reviews a 44 x 33 MF sensor:<br>

http://zackarias.com/for-photographers/gear-gadgets/why-i-moved-to-medium-format-phase-one-iq140-review/<br>

Half way down his review he has a man sitting up at a table with back focus on the eyes, “This was shot with the 55mm 2.8 lens at f3.2. The 55mm is “about” a 35mm focal length in terms of full frame DSLRs. So it’s a bit on the wide angle side of things. Notice how I got the far eye in focus but the near eye is going soft. I’m at a decent distance from Dan with a wide lens. A DSLR with this set up at f3.2 would have held both eyes.” I think he’s mixed up his formats and talking about a larger film MF size vs 35mm – see table in image below. I'm not sure the difference is that much (say 17 v 21 cm around where he focused??).</p>

<div></div>

[jon_savage]

<p>Even the link I posted earlier stated “We just couldn't achieve the same sort of dramatically shallow depth-of-field with a 35mm camera” and “that advantage holds true for this new Pentax 645Z. Its draw for the creative photographer isn't so much that it has 54 million pixels with a pitch of 5.3 microns, but that its 43.8x32.8mm sensor can deliver a crisp full length portrait against a creamy background at an aberration-free f/5.6 and a distance of 15 feet. With a standard 55mm lens”</p>

<p>If creamy background refers to this dramatically shallow depth-of-field I think you can get the same Dof as his 55mm/5.6 with a 44mm/4.5 FX or even a 30mm/2.9 on the little DX. While the quality of his £1000 glass and lens design will impact the aberrations I can’t see for example a Sigma 30/1.4 Art prime stopped down being that bad. It would be interesting to see a proper comparison though...</p>

<p>Even Luminous Landscapes refer to “the fact that it can shoot extremely shallow DOF” in a 645z review as an advantage and Amateur Photographer “The shallow depth of field that is afforded by the larger sensor” relative to 35mm/full frame. But with the availability of fast full frame lenses I’m not sure that statement is as unique to this digital medium format (44x33) they test as they think it is.</p>

<p>This guy lost me completely:<br>

gizmodo.com/why-medium-format-is-so-gorgeous-its-about-more-than-r-160193827</p>

<p>“The key point as a result of that difference [MF vs FX equivalent] is that even though the field of view is wider, the geometry, or "look," of the 50mm focal length remains. You don't get the exaggerated perspective that wide angle lenses usually produce on 35mm cameras. In that regard, medium format mimics how your eyes actually see the world, at least more so than the smaller 35mm size.”</p>

<p>I thought if you choose the focal lengths correctly to take into consideration the crop factors the perspectives are the same.</p>

<p>Some good new though, the high megapixel count doesn’t appear to be an issue:<br>

http://guyaubertin.com/review-pentax-645z-from-a-landscape-photographer.htm<br>

“I don't find diffraction a problem and am happy to use F16 if needed”<br>

With a pixel size of ~5.4microns at F16 the diffusion airy disc is ~22 microns. That’s 4 times the size of the pixel. I guess that’s only an issue if you print large enough to see the pixel level detail. But either way the lack of an issue would also apply to the high MP FX sensors as they wouldn’t have to stop down as far as F16 to get the same increased DOF.</p>

<p>If you are interested I found one informative review of 44x33 vs 36x24 format that came with equivalent image samples and lens options comparisons (40MP 645D vs a D36MP 800). Their conclusion:<br>

“Both cameras produce images of outstanding quality and the for all practical purposes the images are identical in resolution, contrast and color rendition in the ISO range 100 – 400”<br>

And that was from a Pentax website so I assume it’s not biased in favour of the Nikon:<br>

http://www.pentaxforums.com/reviews/nikon-d800e-vs-pentax-645d/image-quality-field-tests.html</p>

<p>I don’t know what the IQ different between them really is having never tested them side by side but this aberration-free with dramatically shallow depth-of-field aspect might be exaggerated a bit but then some people will pay a lot for small improvements. Anyway, I’ve digressed form the point of the thread (sharing the maths) and I’m arriving late reading up on MF and it looks like this has all been discussed before:<br>

http://www.photo.net/nikon-camera-forum/00cWTH?start=30 Is the Pentax 645Z a game changer for Nikon?</p>

<p>I’m still interested in FX for the opportunity of less DoF and therefore the low light and artistic opportunities it opens up but maybe not just yet as the price of a body and a couple of lenses that would show a noticeable difference to my current DX set up add up to more than I would want to spend in one go (I'm saving). I think FX is possibly the best compromise for me as a system in the long run (size, cost, lens availability, DoF opportunities, etc.).</p>

<p>As for ever getting medium format I think I’ll just take pictures at f1.8 with a Sigma 18-35 Art on my old DX (that’s only a 34-70/f3.4 on the 44x33), crop them to 3:2 or even square and save a few quid. As long as I don’t print extremely large no one will ever know :-)</p>