Sony's NEX range of compact system cameras have continued expand with the NEX-C3, reviewed recently, appealing to entry-level users keen to get to grips with Alpha optics and a DSLR-sized APS-C sensor. Adding to those niceties at the top end, if you've a grand to spare, there's the 24Mp NEX-7. Yet in the middle, away from the …
ISO Test Images
I am not a photgrapher ...
My interest in a large sensor camera is primarily the better low light capability so that photos at family occasions aren't grainy, blurred or peculiarly yellow.
So when I donwloaded the ISO sample images I was at first confused that they seemed only to show a gradual decrease in quality. After all isn't the point that the bigger senser gives you a higher ISO rating which therefore allows better quality images under low light?
Further contemplation, reveals that the problem is that the ISO 100 image had an exposure time of 1 second and the ISO 25600 just 1/200th of a second, so of course the first image will be of superior quality!
But since the shots are of a static subject reducing the exposure time makes little sense. Wouldn't this test scenario be better suited to photographing a moving target, say a model train moving at speed in front of the camera.
For a static target wouldn't it be better to keep the exposure time the same (0.1 - 0.05 seconds?) whilst increasing the ISO. The images would then display an improvement in quality rather than a reduction!
You may argue that the Register is an IT website not a photographic website, to which I would reply "Then they shouldn't post such excellent camera reviews!"
Re: ISO Test Images
"I am not a photgrapher ..."
Neither am I, but I've been in the industry for a long time first on the processing side and these days in software, and I'll try to explain.
ISO is a measure of light sensitivity. In the film days you basically achieved higher sensitivity (ie ISO) by using larger grains of silver halide, hence the grainy appearance of high ISO films (this is an oversimplification, but will suffice). In digital you increase the sensitivity by turning up the gain on the photosites (again a simplification) which introduces noise into the image. You can compensate for this with noise reduction, but this tends to introduce "smearing" so has to be applied carefully. The advantage of a larger sensor is that you can have larger photosites for any given pixel count, which means greater sensitivity, which in turn means you will get better results at any given ISO setting than a smaller sensor.
"For a static target wouldn't it be better to keep the exposure time the same (0.1 - 0.05 seconds?) whilst increasing the ISO. The images would then display an improvement in quality rather than a reduction!"
Doing this would simply result in grossly overexposed images - think about it: you turn up the sensitivity of the sensor by, say, a factor of 4 from ISO 100 to 800. If you keep the exposure time the same then the sensor records an image that is 4 stops overexposed.
The bottom line is that a high ISO image will always be less good than a low ISO image in decent lighting. A large sensor will have better high ISO performance than a small sensor (all other things being equal), and will generally be better in low light but to get a properly exposed image you can't just "turn up the ISO" and leave everything else the same, and expect magical low light performance.
This being the Reg I'm sure someone will be along at some point to tell me I'm wrong about everything in an amusingly cynical way, but I think i've got the basics right...
Re: Re: ISO Test Images
No - you are basically right. When all is said an done, for equivalent levels of technology, the basic technical properties of image quality (SNR, dynamic range etc.) are functions of how many photons are detected. High ISO just means the gain has been turned up (either analogue, digitally or some combination) to get the required image density. However, where a sensor has low read noise, the analogue amplification doesn't really gain anything over doing it in post-processing as, fundamentally, your are deriving an image from the same number of detected photons. Larger sensors simply intercept more photons for a given field of view, aperture and shutter speed.
Indeed there is something called "isoless shooting" which just involves using a camera's "base ISO" setting - that is the setting which has optimal dynamic range - and dialling in whatever level of under-exposure is required to get the shutter speed and aperture that you want to freeze the image and get the required depth of field.
Tests on sensors using modern Sony manufactured 16MP sensor in cameras such as the Pentax K5, Nikon D7000 and, indeed, thise Nex-5n show that using base ISO with underexposure at base ISO of down to -6 EV compensating in post-processing produces almost identical results to increasing ISO by the equivalent amount with the same exposure. Note this is not true of sensors with higher read noise.
So it's the photons that count, or rather the number of photons that are counted...
For technical measures of image quality, DXOmark is an excellent site.
You guys seem to have missed the 3D shooting mode of this camera too... the sweep panarama can capture mpo images too, which are quite good, but like the panorama it requires a knack to get seamless images, plus you cant take a 3d photo of a moving subject.
Also, creative modes that enthusiasts would like?? seriously? I cant think of a single person, certainly no enthusiast, that (other than my dad and a few 'challenged' individuals) that have used incamera effects... all this can be done much better on a real computer and as far as I am concerend just a waste of firmware space!
I wonder if the included software installs a rootkit on your PC, like their CDs used to.
Re: Included software
Pointless, boring comment.
The Sony camera group is producing excellent, innovative products.
Sony PMB is included, it's crap, but everyone will use Photoshop Lightroom or Apple Aperture anyway. So it doesn't matter.
Re: Re: Included software
"The Sony camera group is producing excellent, innovative products."
Have you used a HX9V? The pictures look great until you zoom in and you see the awful washy watercolour effect from compression/noise reduction. Awesome video capture though :/ It has put me off buying another Sony camera for quite some time.
I like the ball and dog pic
http://regmedia.co.uk/2012/02/24/sony_nex-5_12.jpg Very nice.
Am I missing the point?
Whilst the body of the camera is certainly compact, by the time you've added a sensible lens you're at pretty much 90% of the size of a normal dSLR aren't you?
Why would I want to buy something that's harder to hold for an insignificant decrease in bulk?
Yes, but I might be biased…
Yes I belive you are missing the point. By the time you've added a sensible lens you are still not even close to the size (or weight) of a DSLR.
Of course the lens size is dictated to a large degree by the sensor size. With the NEX, Sony has tried to cram the APS-C sensor into the smallest bodies possible. Of course you can shrink the sensor size and shrink the lens sizes, but then you lose out on low light performance, depth of field and general image quality. So there is a trade off. Some people won't see it as a trade of worth making - for others DSLR performance in a compact body is definitely appealing.
The other thing (not so important if you don't care about it) is the video performance on a NEX is better than almost any entry level DSLR. And E-Mount lenses are able to continually autofocus (silently) whilst recording.
The big drawback for the moment is the range and choice of lenses. But there is no reason why that won't come over time.
I have the NEX-5
It's a great camera. I picked up the body new with wide lens, and a used 55m lens for under £400 on Ebay.
I purchased it as my friend who is a keen photographer told me he rarely takes his DSLR out anymore because of the bulk of it. This was a perfect compromise.
With the wide angle lens it's the size of a snappy compact. With the 55mm lens on it's about 60-70% size of a DSLR, but still a lot lighter. It also takes great HD video.
re: ISO Test Images @ Felix
Its quite difficult to test.
Firstly, they could have kept the shutter speed at something more realistic. Something like 1/10th , a speed that people can hold so a bit more realistic for realworld use. Then they can change the apperture to control the light. This has some issues as it does effect the final picture due to depth of field differences but ultimately its another way of doing it.
The issue with measuring ISO quality though is you need to test in bright light and low light conditions.
I have a Canon 5D MII and a Canon 450d. As you would expect the 5D is much better picture all round. The difference though with high ISO ratings is marked dependent on the level of light. The 5D at 400 is absolutely perfect in good light conditions. The 450D not so. You really want to go down to 100. However, in low light conditions, even 5D can be a bit grainy at 400 (obviously nothing as bad as the 450D).
Argh, bollocks to touchscreen cameras
They seem like a brilliant idea, to those of us with nice capacitative screen smartphones... Until you realise that it's fine to hold those at a funny angle and shade them on a sunny day.
Flash forward to using a touch screen camera on a summer's day, especially one where the screen is the viewfinder too.. Good luck composing nicely, or driving the camera even slightly manually. You get a nasty case of CSS - Can't See Shit.
Luckily, my GH2 has a touch screen and a variety of physical controls, so I have the option. It would be a reet pain in the ass otherwise. After having played with a touchscreen-only camera, I'm not going there.. Maybe it's fine for the spray and pray crowd though- though would they really want to pay the price of NEX-5N?
A little confused- if it's anything like its predecessors, it will be perfectly nicely made, have a good sensor and all that other stuff.. Usability in all but the most favourable conditions would seem slightly suspect, though.
Touch screen 100% optional - you never need touch it.
The NEX 5N has the same menu system as the cheaper NEX C3 and the more expensive NEX 7. All of its functions are available without ever needing to go near the touch screen.
Use of the the touch screen is entirely optional, thus makes your entire point about usability null and void.
Re: Touch screen 100% optional - you never need touch it.
No, no it doesn't, unless your reading comprehension and understanding of how one uses a camera in modes other than program auto are entirely lacking. Have a nice day.
From the article: "I found I was mostly handling the camera by the lens barrel of the 18-55mm kit lens, despite their being a rubberised front grip. Not exactly the most ergonomic grasp if you want to make swift use of the lens zoom and focusing rings."
So you have to have your hand at the lens barrel, where the zoom ring is located? How is this slowing you down? That's the way you usually handle cameras with protruding lens barrels, especially those with controls on the barrel, isn't it?
This is like making it a negative point that you have to hold on to the steering wheel of a car, when you could also just as well be playing PSP Vita while driving. Yes, what a bummer.
Otherwise a really nice review once again.
The selection of E-mount lenses is still quite small, but with the widely available 3rd party lens adapters I can use my old manual focus lenses with this little gem.
The exceptionally short flange distance of the E-mount allows pretty much any lens to be easily adapted to it, from m39 to middle format mounts, and anything in between. No lenses needed in adapters, so optical quality won't suffer, you only have to take the 1.5 cropping factor into account.