Brisbane city, taken with an iPhone 5S
Centre
Fuji Velvia 50 transparency film
1/250s, f/8, Nikon FM2, Nikon 17-35mm f/2.8 professional lens
Canon EOS 10D ($2k in 2003)
1/350s, f/8, ISO 100, Canon 16-35 f/2.8L II professional lens, jpeg
iPhone 5s
1/320s, f/2.2, ISO 32, jpeg, Apple camera app
Canon EOS 20D ($2k in 2004)
1/250s, f/8, ISO 100, Canon 16-35 f/2.8L II professional lens, jpeg
Canon EOS 30D ($2k in 2006)
1/320s, f/8, ISO 100, Canon 16-35 f/2.8L II professional lens, jpeg
Canon EOS 40D ($2k in 2007)
1/250s, f/8, ISO 100, Canon 16-35 f/2.8L II professional lens, jpeg
Canon EOS 40D ($2k in 2007)
1/500s, f/8, ISO 100, Canon 18-55mm kit lens, jpeg
Canon EOS 40D ($2k in 2007)
1/250s, f/8, ISO 100, Canon 16-35 f/2.8L II professional lens, raw
Nokia Lumia 1020
1/2000s, f/2.2, ISO 100, jpeg.
Nikon D800 ($3.5k in 2012)
1/250s, f/8, ISO 100, NIkon 17-35mm f/2.8 professional lens, jpeg
How do they look?
I’ll turn to the technical differences in a moment, but first I need to get a few things out of my system: "Just look at that Nokia! Wow!", and "Is that all the difference between four models of Canon camera??", and finally "I’d forgotten film had so much detail and grain!".
When I first saw the images from the Nokia Lumia 1020, I did a double take. Clear and crisp, lots of detail, and super-strong colours that you’ll either love or wince at. I loved them. And did I mention the detail? After years of seeing bigger cameras perform better, I couldn’t believe that a tiny plastic and glass Zeiss lens could resolve so much from the centre to the edge of the image. It was close to the Nikon D800. I was stunned. I’ll list the shortcomings of the Nokia below, but first, some more stand-out results.
The Nikon D800 clearly belongs in its own league, with unmatched sharpness, smoothness and dynamic range, let down at the edges only by this 1999-era (but still current) professional lens which couldn’t keep up with the sensor. Arguably, the Nokia catches up with jpegs from the mighty D800 at these softer edges, albeit with more noise.
The Canon DSLRs steadily increased in detail from 2003 to 2007, but - and I’ll emphasise that this is entirely my subjective opinion - the total improvement across four models seems relatively modest in this sunlit scene. It’s eclipsed by the gulf between the iPhone and the Nokia. It left me wondering “Did we really pay that much just for that improvement?”. Put into its historical perspective now, seeing what the Nokia and the D800 can do, it doesn’t square with the excitement I remember feeling with the release of each new model. Did I really get that excited about such tiny differences? Sure, the cameras got more responsive with each model, particularly the 20D and 40D, but in sunlight these jpegs look pretty similar. Perhaps the improvements in low-light will be more pronounced? See the low-light results below.
Putting the cheap lens on the Canons didn't make much difference at the centre of the picture on these settings, but it softened the edges dramatically (see the pictures above on the 40D). This brought the iPhone into the game. To my eye, the iPhone 5S looked better than the 10D with the professional lens, but not as good as the 20D or later cameras. But with the cheap kit lens, the iPhone looked similar to the 20D and 30D at the edges, while still losing out to both at the centre.
Did the film fall where you expected? I’d forgotten how much detail film could capture - the Velvia 50 was right up there with the Nokia and the D800. I’d also forgotten just how intrusive film grain could be. It used to look much more attractive in my rose-tinted memory. Sharpening in Photoshop has accentuated the grain, but it’s a powerful veil of noise across the image even before sharpening. Unlike digital noise that obscures the details, film grain seems to ‘texture’ the details. But it still looks downright ugly to me now that I’m used to silky-smooth digital. The film’s colours are sensational, though, and I remember the original Velvia from 1990 being like crack cocaine to landscape photographers. It was our ‘saturation’ slider in the days when Photoshop was an obscure Mac-only program.
Turning back to the Nokia, its pictures made such a rosy first impression that I had to look closely to spot their shortcomings. The first was a low dynamic range. Compared to the DSLRs and the iPhone, bright parts easily blanched white while dark parts stayed stubbornly black. It felt like using slide film again - the Fuji Velvia 50 film suffered from this even more strongly. And just like with slide film, it made getting the right brightness a knife-edge proposition. The next problem was that some details were smeared away by noise reduction. Nokia chose a balance between tolerating speckly noise, and smearing the noise. Low-contrast details got caught in the crossfire, and rubbed out. In the centre picture, the fine bricks and leaves go soft, while the higher-contrast bricks keep their detail.
Both of these limitations should be eased in a few weeks. Nokia have promised to update the firmware of the 1020 in the New Year [UPDATE - they brought it out early! The Nokia "Black" update is now available in USA on certain carriers, and it'll be coming elsewhere soon]. The update lets the camera record raw format pictures. And two virtues of raw are better dynamic range and more control over noise reduction, as long as you’re prepared to play with the photos on a computer. Judging by their public test shots, it will do both, and dramatically improve the results. I just hope that shooting in raw doesn't slow the camera down much further.
A third challenge with the Nokia was its fixed aperture which gives the same depth of field as a full-frame DSLR at f/9. With the Nokia’s wide 27mm lens, f/9 gets a fair amount in focus, but not everything for critical landscapes. I don’t mind the iPhone’s fixed f/18 look - I only point it at scenes that need everything in focus - landscapes, street shooting, and it does the job. But f/9 is more… middling, and it might or might not suit your taste.
Low light scene
Low light is the DSLR’s home ground, and I expected the SLRs to walk all over the phones. Wrong again.
We wanted to see how well each camera could crank up its sensitivity to capture sharp shots of people in the equivalent of candlelight (EV 2.7). So we levelled the playing field by shooting them all as close to 1/15th sec as we could - that’s a practical maximum for shooting living, breathing humans without losing most of the shots to blur. At 1/15th sec, I’d normally expect over half of my shots of people to be blurred, especially if the people are moving. But most smartphone manufacturers have more relaxed standards (or possibly everyone just moves really slowly at their parties), and they set 1/15th sec as the longest shutter speed on their phones. On Auto, the Nokia wanted to show-off its image stabiliser and shoot for even longer, but we wouldn’t let it. Parties at Nokia HQ are clearly not very animated affairs.
Diarama scene at EV2.7 (equivalent to candlelight) in the Queensland Museum,
showing the area chosen for enlargement. The people are wax!
Fuji Superia 1600 print film (rated at 800)
1/8s, f/2.8, Nikon FM2, Nikon 17-35mm f/2.8 lens
iPhone 5s
1/15s, f/2.2, ISO 1250, jpeg, Camera Awesome app
Nokia Lumia 1020
1/15s, f/2.2, ISO 3200, jpeg. +0.66 Exposure in Camera Raw
Canon EOS 10D ($2k in 2003)
1/15s, f/2.8, ISO 3200, professional lens, jpeg
Canon EOS 10D ($2k in 2003)
1/15s, f/2.8, ISO 3200, professional lens, raw
Canon EOS 20D (($2k in 2004)
1/15s, f/2.8, ISO 3200, professional lens, jpeg
Canon EOS 30D ($2k in 2006)
1/15s, f/2.8, ISO 3200, professional lens, jpeg
Canon EOS 40D ($2k in 2007)
1/15s, f/2.8, ISO 3200, professional lens, jpeg
Canon EOS 40D ($2k in 2007)
1/15s, f/2.8, ISO 3200, professional lens, raw
Nikon D800 ($3.5k in 2012)
1/15s, f/2.8, ISO 3200, professional lens, jpeg
In candlelight, the DSLRs could strut their stuff, and pulled further ahead of the phones. The Nikon D800 kept its crown by a country mile, and even with one hand tied behind its back by limiting it to 1/15th second, it could almost challenge what the Nokia could do at its very best settings.
But it wasn't the walkover that I expected. The phones put up a fierce fight. The iPhone lent heavily on noise reduction, giving a smooth but detail-starved picture. The Nokia didn't use as much noise reduction, giving tighter noise and better detail, but its low dynamic range lost lots of detail in the shadows.
Unsure of how to judge the noisy, detailed DSLRs against smoother, less-detailed phones, we asked 15 non-photographers to do it for us. They ranked the iPhone, Nokia, film and Canon EOS 10D from best to worst, without knowing which was which. The Nokia emerged as the popular favourite, but more striking was how much people disagreed. We got every possible ranking, based around peoples' preferences for smoothness versus detail. People who liked the Nokia picture also tended to like the iPhone, while people who liked the 10D also tended to like the film. It seems that we're all either smooth people, or noisy people.
The Fuji Superia film suprised me. It looks ugly up close with this size of englargement, but analogue grain is different to noise... try sitting back from the screen - way back. When I do that, it looks better than many of the others.
Candlelight also shows off how the Canon EOS cameras have improved throught the years in raw and jpeg. 2007's EOS 40D looks much better than 2003's EOS 10D. Perhaps we didn't waste all that money on upgrades, but it shows how the benefits in that era were mainly in low-light shooting.
iPhone 5
1/120s, f/2.4, ISO 80
The Verdict
Gun to head… time to come up with a number. How many years are smartphones behind the best $2,000 DSLRs? Comparing detail resolved, I'll say the iPhone 5S is equivalent to the DLSRs of 8 to 9 years ago in bright light, while the Nokia trails by less than 6 years - probably nearer 3. This is even when you allow the DSLRs the luxury of a $1,700 lens, and shooting in raw. In bright light, the Nokia came close to competing with the detail from the best DLSR yet made.
Step into candlelight, and the gap between phones and DSLRs widens and becomes more a matter of taste, pivoting around your preferred tradeoff between speckly noise and smeary noise reduction. From our ad-hoc panel of 15 non-photographers, the iPhone trails the DSLRs by about 10 years, and the Nokia about 8.
Splitting the difference between candlelight and daylight, around 6 years of technology has made up for the massive difference in the size of the lenses and sensors between the best phone and the $2,000 DSLRs.
I was stunned.
This isn’t saying that the Nokia is a better camera than a 2007 Canon EOS 40D. It’s not. Detail makes up just a tiny fraction of the goodness of a camera, and none of what makes it a pleasure to use. The Nokia is much slower, can’t focus on moving targets, can’t easily defocus part of the picture, can’t change the perspective and feel of pictures by zooming or changing lenses, and can’t capture the same range of brightness in one shot that the latest SLRs can. Yet.
The curious thing about this list is that everything on it except one - changing lenses - can be fixed with faster processing. The iPhones, Galaxies, and LGs have shown it already. And we know that faster processing is inevitable. The physical design of SLRs gave them a huge headstart over phones for both picture quality and usability, but advances in on-board processing are now quickly eroding that lead.
DSLRs aren’t standing still - they’re improving all the time too. But are they improving fast enough?
Looking forward
The graph below charts the progress of still image quality over time for the Canon EOS 10-series camera models, iPhones, and Samsung Galaxy phones. It uses different scales for the cameras and the phones: DXOMark, and DXOMark Mobile respectively. These are the closest anyone has yet come to condensing the myriad facets of image quality into a single objective number. The graph is misleading at first glance because the phones and the cameras sit on different scales. So it’s not saying that the phones are better than current DSLRs, despite scoring higher. You can only compare phones to phones, and DSLRs to DSLRs.
But it does suggest that improvements in pictures from smartphones compared to other smartphones is going on at a much faster clip than improvements in these DSLRs compared to other dedicated cameras. These numbers don’t directly show that image quality from phones is improving faster than image quality from DSLRs, but they give a pretty strong hint in that direction. I certainly wouldn’t bet against it.
I’d wager that a lot more research and development money is directed at improving phone cameras than improving dedicated cameras. Manufacturers currently ship 13 times as many phones as cameras, and phone sales are going up, while camera sales are going down. Where would you invest?
Looking Back
It’s sobering to look back at the old reviews of the cameras that we included. The earliest, the Canon EOS10D was a marvel of 2003. Phil Askey from DPReview described it as “… the absolute best in its class, with the best image quality, lowest high sensitivity noise, superb build quality and excellent price”. He described the “Excellent resolution”, the “Noise free ‘silky smooth’ images”, with “very low noise levels even at ISO 1600”. The EOS 10D ran rings around the film that we’d been using for 50 years in terms of clarity and freedom from grain.
Yet it’s comprehensively humbled by modern phones. The iPhone out-shoots it, and the Nokia out-resolves it, all by huge margins.
Want to get the most out of your iPhone camera? Check out the World's only jargon-free iPhone Photography course.
The course includes video materials showing how to use the apps we mention. Click here for an example
A huge thank-you to the following people who helped-out with this review:
Mark Eustance for lending the Canon EOS 10D
Eleshia Nielsen for lending the Canon EOS 30D
Lois Robertson for lending the Canon EOS 40D
Mitchell Williams for help with the shooting and editing in Photoshop
Sue for patience above-and-beyond the call of duty while I’d spin the photographer’s eternal lie: “Oh… just one more…”
Apple, iPhone and iPad are registered trademarks of Apple Inc. Take Better Photos does not claim any endorsement by and receives no sponsorship from Apple Inc. We are Apple developers, but the agreement with Apple does not restrict us from being critical of Apple hardware where warranted. We queued and paid normal full price for the iPhone 5S and the Nokia Lumia 1020. Photoshop and Lightroom are registered trademarks of Adobe Systems. Moto is a registrered trademark of Motorola Mobility LLC.