Do Sony mirroless cameras have superior software capability to DSLRs?

[josh_e]

<p>Yesterday I got my hands on a DSLR for the first time, a Nikon D5300. The first thing I did was to take a few shots and compare them with my favorite fixed lens mirroless Sony. The result surprised me because I had heard that D5300 delivers great image quality. First let me say that I used the auto mode and probably that's not where DSLR shine?<br>

One of the shots was of a tree in overcast weather at 6 pm. Sony's picture surprisingly had much better detail and color balance. Here is the point: In low lights the Sony takes a few different shots in a burst and composes them together. I also took a photo at night with fluorescent light on. Image qualities were similar although Nikon was using a much higher iso. So again it seems that the Sony has tricks up its sleeve to compensate for a smaller sensor. So far the only advantage of the Nikon, apart from the ability to change the lens, seems to be low light photography of moving objects.</p>

<p>So optics and sensors are not all of the equation and software signal processing can do a lot. Something like D5300 seems to be behind in that department. For a price similar to that of (the refurbished) Nikon's I can get a Sony A6000. Do you think I'm being naive comparing the two cameras?<br>

Also what goes into taking great shots with a DSLR? Is postprocessing RAW images a necessity?</p>

[Wouter Willemse]

<p>There is a lot of software behind all digital cameras, so yes: it influences the final result. I would be hesitant, however, to state blankly that Sony has better software capability than Nikon, or fill in any other brand. Reviews all over the internet show the differences between brands are pretty futile; that's when comparing 1 on 1, fairly, which i think your test isn't doing. The Sony applied a HDR-like effect, the Nikon did not: the comparison does go a bit wrong there, it's not apples to apples this way. Second, what is the comparison actually based on?<br>

"Image quality" isn't something you can measure. Many people like rich, saturated images and quite a lot of sharpening - images that pop from the screen. Others like more muted colours, rich tonality and prints that can show deep shiny blacks. What is image quality here? Saturation? Lots of dynamic range, or a high contrast? I don't know, I know what I like and dislike, but that's not a good point of reference by any means. So, what you perceive as better image quality may be inferior for somebody else - not something any of us can draw a hard, unambiguous conclusion on.<br>

Second, how did you compare? On their respective screens on the camera, on a normal notebook screen or TV, or on a well-calibrated computer monitor inside a colour-managed program? The differences can be staggering - screens on cameras aren't as reliable for colour nor brightness as you might want to think. Nor are most normal-use cheap LCD panels.<br>

None of this means your comparison is totally useless, and if this finding makes you decide to not get a D5300, no worries. But be sure to do a fair comparison. The fact that the Nikon was at much higher ISO did put it at a serious disadvantage, and yet it performed equally - I'd personally say the Nikon scored 1-0 here, with much better image processing, in fact. So, the comparison should be equal ISO, equal screen to compare, equal output size and so on - if you want to do this rigorously.</p>

<p>You are certainly not naive comparing the D5300 to a A6000 - both are incredibly good cameras. They have different advantages and disadvantages; a lot will be personal (do you like a small, light camera, or prefer a camera with a larger hand grip?). One thing, though, is that both systems are capable of incredible images; their differences in software shouldn't be overstated. Yes, they matter, but not enough to make the real difference.<br>

<em>Is postprocessing RAW images a necessity?</em><br>

Yes, by definition. A raw file is like a negative, raw material from which you want to make a print. There is always some post processing needed. But it can be as much as opening in a program, and saving out as JPEG - post processing is not necessarily a time intensive task.<br>

<em>Also what goes into taking great shots with a DSLR?</em><br>

The photographer. Really nothing else. Great shots come from people who can translate what they see (or envision) into an image. Doesn't necessarily require a DSLR, can be done with a smartphone, a 8"x10" view camera or a DSLR - whatever works best. Cameras are tools, choose the one that fits your needs best.</p>

 

[john_tran14]

<blockquote>

<p>So optics and sensors are not all of the equation</p>

</blockquote>

<p>Of course not. A lens and a sensor do not make any images. However, ...<br>

<strong>optics, sensors, and software signal processing</strong> are not all of the equation either. You and You are two very important factors also. As a photographer You is an important factor to make a good or bad photo, and as a customer You decides how good or bad the photo is. </p>

 

[SCL]

<p>+1 for Wouter's response.</p>

[josh_e]

<p>I compared them on a high resolution laptop monitor of course. By image quality first and foremost I mean detail. In the Sony RX100 image there were more tree leaves distinguishable. <br>

One thing I like about the Nikon is that you can leave the camera on without loosing much battery power. </p>

<div></div>

[josh_e]

<p>Here is a section of the Nikon shot.</p><div></div>

[Spearhead]

<p>That shot is severely underexposed. Using poor quality exposure is a sure way to determine nothing about a camera.</p>

[josh_e]

<p>Both cameras were on auto. It overcast AND dusk.</p>

[Spearhead]

<p>That's irrelevant. If you want to test a camera, take it off auto and expose correctly.</p>

[NetR]

<p>Nikon has consistently got slightly better images from Sony sensors than Sony can (eg the Nikon D810 vs the Sony A7R - se <a href="http://www.dxomark.com/Cameras/Compare/Side-by-side/Nikon-D810-versus-Sony-A7R-versus-Canon-EOS-5D-Mark-III___963_917_795">http://www.dxomark.com/Cameras/Compare/Side-by-side/Nikon-D810-versus-Sony-A7R-versus-Canon-EOS-5D-Mark-III___963_917_795</a> . A proper comparison needs both cameras to have the same settings in terms of ISO, shutter speed and aperture. Auto cam vary from second to second with light conditions and may be implemented differently between camera makers.</p>

[q.g._de_bakker]

If you want to compare cameras, including their software, you must not take part of the task trusted to them away from them. A proper comparison of the cameras' capabilities does require you to interfere as little as possible. So yes, put them on auto and see how well or badly they handle difficult situations. Not on manual, using your own 'software' to do part of the work. Do not decide they should be on the same setting in terms of anything, except auto-mode.<br>Of course that <i>"may be implemented differently between camera makers"</i>. That's exactly what the question is about. So let the cameras show how things are implemented and what they are capable of, without messing the comparison up by intervening and interfering.

[Jochen_S]

<p>On my not calibrated screen I like the Nikon colors more. looking at the EXIFs: Did the Nikon use a VR lens? If not 1/30sec might mean camera shake. On what did the Nikon have a chance to focus? The bush in the foreground? And hile other issues are most likely more relevant: Kit zooms aren't the bestest lenses to be shot wide open.<br>

I don't really know contemporary Sonys or Nikons. I think the days when DSLRs weren't capable to produce bearable JPGs in camera ended when the last wave of 6MP ones hit the market. <br>

I wouldn't underate low light AF capability again. - Right now I have 2 systems lacking in that field and another compensates with great manual focus only. So some Nikon is on my personal shopping list. - Not sure what to say about the RAW tweaking: I shoot RAW whenever I can since my first camera sucked with JPGs and memory cards and diskspace are comparably inexpensive. But as usual: YMMV and I can't stress enough that great bulky cameras won't take great pictures if they rest on a shelf at home due to their bulk.</p>

[Wouter Willemse]

<p>Probably not clear from my initial answer, but I agree with what Q.G. said: it depends on what you want to measure. If you want to know how well the embedded software does, then by all means: shoot auto and create JPEGs. The point at the same time is that you're not having a well-defined yardstick to measure against, except for <em>your own preference</em>. Which is fine, <em>if the conclusion is only for you personally</em> - but not if you extrapolate that conclusion to a generic statement on brand A versus brand B. For generic statements, the way Jeff suggests to test will be more approriate.<br /> If you use raw files, the comparison does become more complicated as there is another variable brought in - not all raw converters have equally good profiles for all cameras (and also here, taste matters and influences a lot of the tool of choice).</p>

<p>Looking at the two samples, on my screen the Sony looks quite cyan-blue; the colours of the Nikon seem more pleasant to me. In terms of details and contrast, the two samples do not seem miles apart, a bit more for the Sony but not much in it for me. Given that it is shot in bad light, and in general looks murky and lacking in contrast, I would not draw any conclusions from these photos. Except maybe, just maybe, that for ISO 2800, that Nikon shot looks actually pretty decent - I still believe the comparison is skewed there, but again, it depends on what you want to measure.<br>

But given the test shots, I'd do one thing. Shoot in decent light, and have another look - and it can still be you prefer the Sony; it is a good camera, and the Nikon not necessarily better. As said, there is a big element of personal preference involved. But at least you'll have a better idea of what both cameras are capable off in terms of details, colour, contrast and so on.</p>

[q.g._de_bakker]

I disagree, Wouter. It's not your own preference you will see, but what each camera does when left to do what it wants. Your preferences come in later, when deciding what you like best. They are not part of the test.<br><br>When you start tempering with what the camera does, for instance by deciding that it is not the camera but you who will set exposure, you are no longer testing how well a camera does at deciding what exposure to set. Then you will indeed become part of the test. What <i>"expose correctly"</i> means is part of <i>"your own preference"</i>.<br>Yes, when you eliminate that possible difference between cameras, you perhaps get a clearer view of what else (!) the cameras do. But then the test result will be partly due to the camera, partly to you. And you will not know how that will skew the test, which camera benefits most from the help you offer.<br><i>"Expose correctly"</i>, and especially <i>"shoot in decent light"</i> also has another drawback: presenting ideal test conditions is not the best way of finding out how good something does. Ideal, i.e. non-challenging conditions should be part of a test (you never know, maybe there are cameras that can only handle difficult situations ;-) ). But a more challenging set of conditions will reveal more.

[Wouter Willemse]

<p>QG, I understand your point, and I think in many ways we may say the same thing:</p>

<blockquote>

<p>Your preferences come in later, when deciding what you like best. They are not part of the test.</p>

</blockquote>

<p>I see the 'conclusion' as part of the test. Technically, you're correct and it is not. But seperating the two, especially in a not very formal test, is also a bit artificial. So I just lumped it together ;-) Also agree that challenging set of conditions isn't a bad thing to do, and I did not imply that should be the only test, but same difference in thinking: it is the conclusion drawn from only that one single challenging test I'd dispute, because I think it just draws a far too partial picture. You'd need multiple scenarios to get the possibility to draw a balanced conclusion.<br>

If I read your post correctly we're much saying the same thing, just the definition of 'test' is slightly differenmt.</p>

[josh_e]

<p><a href="/photodb/user?user_id=282122">Q.G. de Bakker:</a><br>

Yes, that's exactly what my question about.<br>

<a href="/photodb/user?user_id=640949">Jochen Schrey:</a><br>

Yes it was a VR lens. Both cameras were focusing on a tree part of it you can see on the right side of the pictures. I cropped and resized the photos. (The part I chose was for convenience reasons but other areas were not much different.)</p>

[josh_e]

<p>I took some more shots and although the amount of detail in Sony's shots are close to Nikon's, the latter provides much more realistic colors. Now it seems there is something wrong with Sony's white balance (probably a result of the signal processing postproduction that I was touting at first.)<br>

<br /> But the fact that a tiny camera like RX100 can deliver image detail close to that of D5300 is sobering. Now I should make up my mind whether I want to keep the Nikon, exchange it for an A6000 or collect my pennies to get a hardcore beast later. It seems that I expect too much for an inexpensive camera. I like it to be able to take low noise photos of candlelight objects (using a tripod)!</p>

[Peter_in_PA]

<p>Josh, I'm late to the party, but here goes.<br /><br />You compared the image quality of a camera you really know how to use and are used to... with one that you "just got your hands on".<br /><br />The thing you've proven, above everything else, is that <em><strong>you can take a better picture with a camera you know well than one you don't.</strong></em><br /><br />Both of these cameras can take great photos if they are used right, and someone who knows how to use one and someone who knows how to use the other can take the same photo, same quality.<br /><br />And, that said, any camera RAW image processed carefully and thoughtfully through Photoshop will almost certainly always be a better image than an out-of-camera jpeg.</p>

[q.g._de_bakker]

I wouldn't say that he's proven that, Peter. From what do you infer that the difference in the results come from Josh being more familiar with one of the two cameras?<br>What is proven perhaps is that when two cameras are used the same, there is some difference in the result. And i am assuming that both are indeed used the same (set to Auto, point & shoot).<br><br>Your point is of course valid: even when the results obtained straight out of each of these cameras in auto-everything mode differ, that does not have to be the final word on what you can get out of them.

[Supriyo]

<p>Josh,<br>

I think, using your example, the conclusion that Sony has better software capabilities than nikon is misleading. Besides the color balance, the reason why sony's sample seems to have better details than nikon is due to increased sharpening. If you take Nikon's sample and apply sharpening in Photoshop (and adjust color balance), it will look like Sony. If you prefer increased sharpening in your images, you can crank up the sharpening in Nikon's settings. However over sharpening has serious problems in some situations. You can always sharpen an unsharpened image, but cannot do the reverse. Nikon's default settings are therefore prudent.</p>

 

<blockquote>

<p>But the fact that a tiny camera like RX100 can deliver image detail close to that of D5300 is sobering</p>

</blockquote>

<p> <br>

You cannot come to this conclusion by examining just one sample. You need to compare quite a few images from both. DPreview studio comparison tool has both these cameras. If you compare closely, you can see much more noise and less detail in RX100 than D5300 at high ISO 3200.<br>

<br>

Sony A6000 is a good camera, but with much less lens choices than nikon slrs, and also current and upcoming sony lenses are on the expensive side.</p>