Just Posted: Sony DSC-RX100 preview with sample images

Forget the sensor... It's a compact f/1.8 lens!!

It would make absolutely no sense for a manufacturer to limit the aperature artificially just to avoid diffraction. If could make sense in a full auto mode for beginners who don't know what aperature is, but educated users should be able to decide when they need to close down further regardless of the slight decrease in IQ.

I shoot with FF dSLRs for 90% of my work, and rarely stop down below f/11. The effect of out-of-focus is far more pleasing than the mushiness from diffraction.

HOWEVER: sometimes I need to stop down to f/16 or even f/22 or f/32 when I need to drag the shutter in not-so-low-light (think motion blur). Sure, I could use an ND, but don't always carry one.

So there is a use to having smaller f/stops, even in compacts.

I have too many cameras and chargers, so on board USB charging works for me. However if this is going to be your primary camera, an independent charger would be nice.

I think he meant more like a Fuji X10, which is the dilema I'm struggling with. I like the smaller body of the RX100, and the much better video quality (this I can guarantee without having seen any RX100 video), but I prefer the fairly fast X10 zoom, which is f2.8 at the tele end. Hmm...

Love the XZ-1, but I think I have to give the IQ nod to the RX100 when comparing both cameras at base ISO. And only 720 HD on the XZ-1. XZ-2 is coming but it will have to be pretty spectacular.

I am presuming the sensor size makes getting a fast lens more difficult. From my point of view I would prefer a smaller sensor if this means a brighter aperture at long focal lengths for taking indoor pictures of the grandchildren. I had a XZ-1 but did not like the image quality and the Fuji X10 does seem the ideal compromise but not with the rather eccentric EXR sensor. The XR100 seems a brilliant concept but the lens is already down to f3.5 at 53mm which is a pity. Not the strongest camera indoors I would think but useable I am sure.

I really think it might change. Haven't MILCs gained popularity for having bigger sensors?. Most users stick to kit lenses. Notice also how Pentax Q failed. It seems people are starting to know what does sensor size mean.

just look at all the discussions about APS-C vs FF - there's usually at least one person that judges larger sensor as inferior.

I think the size of the sensor in camera has been determined by the megapixel number they wanted to market. To get a good quality at this sort of pixel packing it was 1" which was also a good marketing point. This has however compromised the lens making it pretty slow for a modern top end compact. If you look at a light meter it says nothing about sensor size, just the combination of aperture, shutter speed and ISO for correct exposure. This means indoor without flash you are going to have to push the ISO or slow the shutter at f4.9 and already at 50mm it is down to f3.5. Given that pixel packing pushing the ISO is not going to be as good as if the sensor was less stressed by its pixel size. When I first saw tis camera I said wow but that lens is slow and purely I think because cramming in the pointlessly high megapixels was the main driver of the design. People seem to fuss about tiny sensors but see no problem in tiny pixels.

It not a sensor size issue. The iPhone 4 and Nikon 1 don't do 1080/60p which is a demanding frame rate/resolution.

Anyway, I hope not as this camera will make a fantastic B roll camera to use alongside of my D800 for super slow-mo shots.

I'm very much concerned and surprised by the image quality of image 2008006.jpg (the one with the two birds). If one is looking into the patches of grass it looks to me as if this camera would already need noise reduction at ISO 125. I have no other explanation for the smeared out patches of grass. The same holds for image 2011772. It is really annoying.

TrojMacReady:

ISO 1600 is about what I figured the last good setting would be for the Sony. And given what I've seen from the Sony ISO 3200 raws, I'm correct but await better raw extraction software before claiming proof.

As for the equal crop (really equal dpi) trick, yes that's one data point. A much better thing to look at is the evenness of the noise/grain pattern. The Sony RX100's is irregular at ISO 3200 and the Nikon V1's isn't. Again time and better raw extraction software may improve the now unusable Sony RX100 ISO 3200 raw. No I will not be posting cropped proofs.

Strandlaeufer:

If by "smoother interpolation" you mean more blur, less detail, less sharpness, and weaker colour quality, then yes.

Note the irony.

And I note that above you're commenting on jpegs.

TrojMacReady:

Really a good "color profile" is going to solve the noise and blotchyness of Sony RX100 ISO 3200 raws. (As I already said: better raw extraction software may help--but we have to wait for that to come out.)

Your ignorance is really on display here with claims about a colour profile being the solution.

From the raws I can extract, at high ISOs, the Nikon V1 destroys the Sony RX100, for noise, and evenness of colour gradation.

No great surprise, given the 20MP problem burdening the Sony.

TrojMacReady:

You're looking at jpegs? Are you joking?

Again: Jpgs, really, no one who cares about colour, noise, or detail bothers to even look at jpegs.

The fact is that the ISO 3200 Sony RX100 look like crap, and the Nikon 1s don't. (Perhaps there'll be an extraction software fix for the Sony.)

TrojMacReady:

Quoting you from above:

"Looks like jpeg like smudgyness rather than the usual grain type noise.

And despite that smoothing, the RX100 still looks better overall:
http://i46.tinypic.com/bzmle.jpg"

Do you see now that you're writing about jpgs, I don't dispute that you started with raws, in some cases.

Choosing a black point doesn't magically make the noise go away, that's only going to happen with less sensor noise and a camera processor that can identify noise well and keep it out of the raws. By that argument, you could pick a black point anywhere in the processing and get rid of noise--aint how it works.

As I have already written better raw extraction software may save the ISO 3200 raws from the Sony RX100, but it aint looking good, unlike with the Nikon V1 at 3200.

Try finding the "black point selector" in ACR 6. Now if you mean the white balance tool, that doesn't do much about noise.

Enjoy your jpegs.

Menneisyys:

A really good lens will not make up for the fact of the extra noise from the sensor. Also unlikely that the Nokia shoots raw.

Now what could hypothetically make up for the problems of cramming so many pixels in a such a tiny space would be a near magic processor inside the camera; this would have to be a processor able to short out almost all noise and heat errors from the sensor. Somehow I don't think a Nokia phone is going to be the first place this tech shows up for public use.

A simpler solution to improving camera phone photos would be fewer pixels and the ability to shoot raw.

Plastek:

I like the A900 and the Nex 5n. And the SLT's live view is helpful.

No, Sony has made another technology cramming mistake with this RX100. This isn't a new mistake for Sony to make.

Yes, according to some test here at focus numerique, the 20mpx is better, but with 12mpx and less pixels density, it was even better, while beating the nikon in pixels count. I prefer less pixels count and better high noise detail endition than 20mpx only usable at base iso and noise in shadow.
Raw comparisaon here http://www.focus-numerique.com/test-1454/compact-sony-rx100-bruit-electronique-12.html. You find raw for rx100 and nikon.

spidermoon:

Right and the 6400ISO raws from the Nikon V1 sure look usable, unlike the 3200ISO raws from the Sony RX100.

@HowaboutRAW at what size did you print D800/D3/D4 RAW samples?

Couldn't agree moew.

Noise is determined for a given angular resolution and shutter speed by the lens aperture and not by the size of the detector. The aperture of a lense is given by the focal length diveded by the f-number. Attractive apertures are those less than 4 mm and larger than 20 mm. Apertures less than 4mm are close to the aperture of the human eye and are giving your images a similar huge depth of sharpness. Apertures larger than 20 mm lead to a very narrow depth of sharpness so that the object of interest is crisp sharp while the rest is clearly out of focus. The range between is less attractive because large parts of he image are blurred without being clearly out of focus. One has only to look at the example images.

Following your argument there is no advantage of large sensors from noise point of view. I have my doubts.
Larger sensor capture more photons due to larger aperture of an adequate lenses. Alone from this point I would expect less noise.
On the other side the semiconductor processes of the sensors are far from perfect. With the larger lenses and high costs we also buy sensors which are less affected by spot noise and sensor production inperfections.
Did I get something wrong?

Strandlaeufer:

Your summary of what a large aperture lens can purportedly do or not do seems more that a bit contrary to actuality.

pxw:

Yes, your points seem sound.

Strandlaeufer - I have the impression the density of pixels on the sensor is the key determinant of noise, all other things (aperture, light level, sensor type) being equal. As a practical matter, manufacturers have to pack small sensors with a much higher number of pixels per square mm, and smaller detectors generate more noise when respond to lower light levels. Sensor size would be irrelevant if one compared two sensors of different sizes that had the same receptor density. In that case the smaller sensor would necessarily have few pixels. The Sony's lower density vs. a typical point & shoot should result in lower noise.

You would be comparing pixels rather than images. At an image level, the larger sensor would still win. Example? Compare the D3200 with the V1. Almost the same pixel density, but the D3200 is cleaner at ISO 6400 than the V1 at ISO 3200 for a given output size. Why? Because the sensor is more than twice the size, thus resulting in more than a stop better noise performance.

TrojMacReady:

Right and every argument you've applied to the D3200 would make the Nikon 1s less noise prone at high ISOs than the Sony RX100. So for indoor use the Nikon 1 is a much better bet than the Sony RX100. (The Sony would be better big at prints of ISO 200 daylight landscapes.)

Say what? Back up your words with some proof.
What I just explained through a clear example, contradicts your endless theories about pixeldensities.

TrojMacReady:

You so nicely made the point for me directly above, thanks.

Irony, that you do such a good job making my exact point, which elsewhere you just can't accept: Sony has denied itself a market here. That's the market for cameras that can do things other than daylight landscape pictures.

Again: Thanks for the support, now buy Fuji.

You apparently didn't understand what I wrote. The key driver is sensor size, not pixel density. As also shown by the J1 vs RX100 comparison (from RAW) I posted yesterday.

TrojMacReady:

Yep I misread your point about the Nikon D3200. However Nikon does a much better job with noise and Sony sensors than Sony--for example the D3x versus the A900--there was a reason for that extra $5000.

Nope, you brought up jpegs in the J1 versus RX100 claims. NB:I was commenting on the Nikon V1. As evinced by my misreading of your Nikon D3200 claims above, we all make mistakes.

I assume that Imaging Resource has raws from D3200 and that ACR 6.7 will extract said raws.

Typical debunked nonsense about Nikon again. The D3X had an improved version of the A900 sensor (it came out later and you paid for it...), but also one that sacrificed colour accuracy over low light sensitivity at higher ISO's (the A900 had much better colour separation, for which you'll find agreement in the Nikon forum and DXO measurements).

The D7000 was not better than the A580 carrying the same sensor. The D300S is no better at high ISO than the A500, etc. Verifiable facts.

Selective quoting from my messages doesn't make you right. I was talking about the RAW files, mentioned the word RAW twice and posted the converted jpegs, since posting non converted RAW files would be of no help to anyone. Key difference compared to claims with nothing to back them up, as you did.

I guess this sensor will take place in a high end bridge. With a faster and wider zoom range.

"RX100 = R2."

Exactly! So far, the R1 has been the only Sony camera I *loved* - other cameras all had some major problems (e.g., low IQ (16mm) / huge (18-55/18-200) wide lens for the NEX series; worse high ISO for SLT models etc., and the IQ of ALL their P&S cameras have been extremely bad). Now, the RX100 is my new love and will purchase it as soon as it hits the shelves. Hope it delivers in the DPReview test as good as in that of IR.

I think by "this sort of camera" I am talking about something that can output something approaching a digital SLR but making as few compromises as needed on quality and be pocketable.

From a personal perspective along the lines of my 35mm film Yashica T5 / XP2 combo which by your standards would probably be a brick

I agree with Jimmy. There are aspects of the Nikon 1 like AF speed and accuracy via PDAF and fast FPS that a camera like the RX100 may not be able to compete with.

just get a hoodman loupe for 3 inch lcd. i have been using it on my s100 and it is better than a digital viewfinder or tilting screen and stabilizes the camera to my head and hands.

I agree with rpm40, except I would have liked just a little bit of grip. Small enough grip to not compromise the smallness yet be big enough to comfortably hold it safely.

Agree with Miwok. Tilting screen on NEX is great and would have been a great addition to this camera.

Take into account downsampling the Sony to S95 sizes and you will see the quality is much superior to the S95.

What the problem with the samples ? I can see some blue dots in the two girl hair and smudge area in the hair, but only at 100% and close look. It's good skin tone and colors. On the first girl, you can even see the rx100 in her eyes. The "Pont-Des-Arts" photo show good DR, between the white cloud, not burned and detail in the the black suit, again with some blue dots at very close inspection. The "Hotel de Ville" of Paris" night shot is nice too, showing some deep shadow detail while not totaly blowing highlight. For a compact camera, it's a very good result.

@spidermoon

A agree the Pont-Des-Arts image really caught my eye for the DR, as did the portraits of the pretty ladies. Excellent detail. I did notice on the TechRadar samples images of the football match, the ISO 800 JPEG has a bit of watercolor look at 100%, but this won't be seen in the RAWs nor is it visible at normal viewing size.

I wonder if it's made by Zeiss Oberkochen or at Zeiss Jena. Just kidding. But, if you gave people a choice between an average quality lens marked Zeiss or Leitz and a truly spectacular lens made by some company they never heard of, which would they choose?

Well, each KFC outlet "engineers" its own fried chicken, but all supposedly according to the instructions of the deceased Harlan Sanders, who was an honorary colonel of the Kentucky militia. Names, titles, brands, and whatnot have a very diffuse (mongrelized) pedigree.

I wonder why Sony doesn't leverage the G Lens team... because Zeiss is more marketable? I am under the impression G Lens is more premium in Sony's lineup

Zeiss sells better.
In terms of "premium" quality - usually Zeiss lenses are premium lenses with optical designs made in germany by Zeiss engineers, Sony G are premium lenses with optical designs made in Japan by Sony engineers. That's the whole story. Both have great QC and are really stunning pieces of glass.

Awesome!

Or the same people that got the s90/95. Looks like a good upgrade for them

You guys are thinking Canon 580 EX though. We're talking about a small flash and EVF like the one for the NEX-5N.

http://www.imaging-resource.com/PRODS/NEX5/ZFLASHANIM_500.GIF

http://cdn4.mos.techradar.com///Review%20images/PhotoRadar/Sony/NEX%205N/sony-nex-5n-front-380-75.jpg

I don't want to buy into the NEX system though. I just want a camera like this one (RAW, some zoom, good IQ, pocketable for 75%+ of the time I'm not using the accessories) with the flexibility to add some accessories if required/wanted.

Because cameras are getting smaller, we must now forgo creative lighting?

Usually, people who utilize only available light don't yet understand how to use strobes. So I say learn to love the hot shoe so you'll at least have more options.

Better to have slave flash which work in infrared or radio mode. This feature can be easily done (by Sony), very small and hidden inside the camera body.