How to properly compare RX-100 and NEX-6 kit high ISO noise?

Started Apr 29, 2013 | Discussions thread
twald
Regular MemberPosts: 498Gear list
Like?
Re: How to properly compare RX-100 and NEX-6 kit high ISO noise?
In reply to Mel Snyder, Apr 30, 2013

Mel Snyder wrote:

Vlad4D wrote:

Mel Snyder wrote:

"Reciprocity" - which those of us from the film era knew well - is key to dynamic range in sensors as well as film. The sensitivity of sensors and film is not linear. The NEX "jams" the reciprocity curve on only one factor, because it cannot open to an aperture wider than f3.5. That forces the worst case for the reciprocity of the NEX sensor. The RX-100 faces no such single-factor jam.

Yes, it forces worse scenario for NEX. But this was one of the points of the test: will the NEX manage to overcome limited aperture problem with a help of bigger sensor? Seems like no - it only managed to be about the same as RX or little worse (on wide end)

It's not just the "limited sensor/bigger sensor" - it's the way each measures the scene - as I pointed out in the part of my message you didn't mention. That's why the left edge of the RX-100 is overexposed.

This is especially prejudicial because you allowed the cameras to "choose" ISO 3200, which has a horrible dynamic range - anyone who has actually shot much at ISO 3200 understands that it's horrific on subjects with considerable peak-to-trough lighting challenges. Again, you jam the test against the NEX, but favor the RX-100 because it can admit more light.

It is possible that ISO800 on RX-100 is doing internally exactly same signal amplification as ISO3200 on NEX. Similar noise levels partially proves that. So most likely dynamic range should be also similar. It is just my assumption ...

These tests are absurd on the face of it, anyway. One device is an optimized tiny camera that is best of class. The other is an interchangeable lens mirrorless that is, in the opinion of all but the MFT crowd, best of class. Comparing the NEX and RX-100 is as absurd as if someone with a digital Hasselblad came came onto the forum and said, "let's compare." To which I'd say, "sure, I'll put my RX-100 in my pocket, you put your Hasselblad in yours, and let's go test cameras."

I think comparing always makes sense - it helps us to find right tool for our hobby or work.

Perhaps - but this kind of test doesn't prove anything. A camera optimized for a single lens versus one that can handle thousands of lenses? And a scene specifically picked to handicap one and favor the other? That's the way Florida Republicans set polling locations and hours.

Anyone for actually taking photos instead of optical arm-wrestling?

This is brand specific and actually camera line specific forum, and of course people discuss technical aspects here. For pure art we have several other forums on dpreview.com. But by some reason they are no so active as for example this one

Nice picture

Thanks. But this test isn't a technically balanced comparison. If you picked a scene that nullified the differences in metering systems, then - perhaps - the test might have been valid. But as I noted, the scene you selected favored the heavily center-weighted RX100, which is why the left edge blew out, but was properly exposed with the NEX-6. From that point on, the test proves nothing.

NEX-6 with 50mm Summicron, f2.8

I don't think that the cameras' metering is so much different, rather, it is just too dark for the NEX with the  kit lens to make a brighter picture that looks good, and the engineers knew it. At a certain point (ISO 3200, 1/25 for the NEX) cameras algorithms determine that a dark image will look better than one with a balanced histogram. ISO 3200 is as high as auto ISO will go and slower shutter speed begins to blur; it balances exposure with other aspects of image quality.

I recon that if light were to drop even further, the RX100 would make a darker picture and not maintain an ideal exposure either. I do understand your point, though; you are saying that a fair comparison would have been to drop the ISO of the RX100 to 1600 so that the cameras expose equally. The problem is that the NEX image already has more noise than the RX100 does. As I understand it, this noise is not caused by sensor inefficiencies, but rather, it is an accurate representation of the actual image; the photons are noisy because there are too few of them.

The RX100 is not mine; it is a friend's and I borrowed it for a try. I will try to make a better comparison when I have time, but it won't be today or tomorrow.

Don't expect a great difference from a another comparison. The Zeiss is a very nice lens that actually lets in a bit more light than the NEX kit lens, and it is coupled to a superb (if small) sensor. I also did an outdoors test in good lighting and discovered that it outresolves the NEX-5n at base ISO simply by virtue of more megapixels.

Two areas where the NEX--again with a kit lens--wins are dynamic range and color depth. I think the NEX raw files are a bit more malleable; they are easier to tweak in post-processing without completely ruining colors or contrast.

Reply   Reply with quote   Complain
Post (hide subjects)Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow